Sunteți pe pagina 1din 103

A STUDY ON THE PERFORMANCE OF THOMSON REUTER MANILAS CASES-GENERALIST TEAM

Table of Contents
Abbreviations and Acronyms ..................................................................................................................... i Executive Summary................................................................................................................................... ii INTRODUCTION ......................................................................................................................................... 1 Nature and Scope of the Study ............................................................................................................. 1 PROBLEM DEFINITION............................................................................................................................... 3 Purpose of the Study............................................................................................................................. 3 Significance of the Study ....................................................................................................................... 3 Current Knowledge about the Problem ................................................................................................ 4 Preliminary Hypotheses ........................................................................................................................ 4 METHODOLOGY ........................................................................................................................................ 5 Analytical Procedure ............................................................................................................................. 8 DATA PRESENTATION.............................................................................................................................. 12 Descriptive Statistics of Key Variables ................................................................................................ 12 STATISTICAL ANALYSIS ............................................................................................................................ 19 Analysis of Performance based on Output ......................................................................................... 19 Analysis of Performance based on Errors ........................................................................................... 26 Analysis of Performance based on Output vs. Benchmark ................................................................. 40 Analysis of Performance based on Independence between Error Types ........................................... 43 Analysis of Performance based on Output of Multiple Sample Means .............................................. 46 Analysis of Factors Affecting the Performance based on Output and Error....................................... 47 Conclusion ........................................................................................................................................... 49 RECOMMENDATIONS ............................................................................................................................. 51 APPENDICES ............................................................................................................................................ 52 Audit Criteria ....................................................................................................................................... 52 PH-Stat Results .................................................................................................................................... 53 BIBLIOGRAPHY ........................................................................................................................................ 96

Tables and Figures


List of Tables Table 1 Team Performance by Output........................................................................................................ 16 Table 2 Number of Errors Committed by Team .......................................................................................... 17 Table 3 Number of Errors Committed by Gender ....................................................................................... 17 Table 4 Number of Errors Committed by PS Level...................................................................................... 17 Table 5 Number of Errors Committed by Age............................................................................................. 18 Table 6 Number of Errors Committed by Tenure ....................................................................................... 18 Table 7 Number of Sub-Type Errors Committed by Team .......................................................................... 18 Table 8 Number of Sub-Type Errors Committed by Gender ....................................................................... 18 Table 9 Number of Sub-Type Errors Committed by PS Level...................................................................... 18 Table 10 Number of Sub-Type Errors Committed by Age........................................................................... 19 Table 11 Number of Sub-Type Errors Committed by Tenure ..................................................................... 19

List of Figures Figure 1 Frequency Distribution by Age...................................................................................................... 12 Figure 2 Frequency Distribution by Gender ................................................................................................ 13 Figure 3 Frequency Distribution by Income ................................................................................................ 13 Figure 4 Frequency Distribution by Tenure ................................................................................................ 14 Figure 5 Error Categories and Sub-Categories ............................................................................................ 17

Abbreviations, Acronyms, and Definitions


Workflow: A progression of functions that comprise a work process, involve a group of people, and create or add value to the organizations activities. Team: A group of people working on a common workflow. In the study, it is composed of Teams H and M. Name: A personal identification of the member of the team. Wave: A type of hiring process where a company hires in groups. In this case, wave is indicated by numbers which identifies when the member was hired. Wave 1 is the very first wave hired and wave 8 is the last. All the next hires after wave 8 do not come in groups anymore. Function: A step or task in order to complete a work process. The team being observed in this study has three functions: a) Statute Verification (SV): The process of verifying and styling laws cited in a case. b) Opinion Verification (OV): The process of verifying case laws cited in a case. c) Copy Preparation (CP): The process of copy editing and preparing the case for publication. PS Level: It stands for Publishing Specialist Level, which refers to the position label or hierarchy of team members in this study. Total cases: Total number of cases processed by a team member. Total number of pages: Sum of the pages of cases processed. Production hours: Number of hours spent in doing the function. Average Pages per Hour: Computed as total number of pages divided by production hours. Transition Phase Benchmark: Daily performance metrics; the higher the position, the higher the benchmark is set.

Executive Summary
This research study report is submitted by Group 2 in partial fulfillment of the course Management Statistics 501M under Ms. Maria Angeli Reyes. The subject of the study is the performance of the Manila Cases-Generalist Team of Thomson Reuters, a leading and worldwide intelligent information provider. The purpose of the study was to evaluate the performance of two sub-teams vis--vis performance benchmarks and other variables of interest. The specific objectives of the study were: 1) to assess and compare the performance of two sub-teams of the Cases-Generalist workflow on statute verification, opinion verification, and copy preparation functions; 2) to determine the process area where higher numbers of errors have been committed to recommend ways for improvement of the workflow process; 3) to identify factors which significantly affect the number of errors committed by team members in each of the workflow process; and 4) to provide feedback to the Team and Management on the result of the performance assessment for the purpose of further improving effectiveness and efficiency. The research study is observational and retrospective and employed descriptive statistics to summarize data, Chi-Square statistics to test the independence of variables, means hypotheses to compare performance against benchmarks, between teams, gender, age ranges, seniority levels, analyses of variance to compare performance among wave levels, and linear regression to determine linear relationships between production and age, income, and tenure. Data was provided by Thomson Reuter Manilas Cases-Generalist team managers and was processed using Excel and its add-in PHStat. Two-tailed hypothesis tests indicated that there is no difference in the means of total production hours and average production per hour between sub-teams, members with age ranges 24 and >24, and members with tenure of <2.5 and 2.5 years. There is a difference in the means of total production hours between PS levels 1 and 2. There is also a difference in the means of all errors between Team H and M, with Team M having committed more errors on average. There is no difference in the means of errors between Team H and M with regards to copy preparing, statute verification, opinion verification, as well as in the error sub-types of stylistic, content and major content. With regards to gender, the group found no differences in the means of all errors by females and males, and no differences in the means of copy preparing, statute verification, opinion verification, stylistic, content, and major content errors. ii

With regards to PS level, we found no differences in the means of all errors by PS 1 and 2, and no differences in the means of copy preparing, statute verification, opinion verification, content, and major content errors. There is a significant difference in the means of stylistic errors, with PS 2 having committed more error on average. With regards to age, we found no differences in the means of all errors by members with ages 24 and >24, and no differences in the means of copy preparing, statute verification, opinion verification, stylistic, content, and major content errors. With regards to tenure, we found no differences in the means of all errors by members with tenure of <2.5 years and 2.5 years, and no differences in the means of copy preparing, statute verification, opinion verification, content and major content errors. There is a significant difference in the means of stylistic errors, with members with tenure of 2.5 years having committed more stylistic errors on average.

Upper-tailed hypothesis tests revealed that all members assigned to statute verification and copy preparing were able to meet production benchmarks. While all other members assigned to opinion verification were able to meet production benchmarks, new hires assigned to this function were not able to meet theirs. Meanwhile, Chi-Square statistics uncovered that type of error is related to team membership with Team H being prone to commit statute verification errors and Team M to committing copy preparing and opinion verification errors. Type of error is also related to gender. Females tend to commit copy preparing errors while males tend to commit statute verification and opinion verification errors.

As well, PS level is related to type of error. PS level 2 members tend to commit copy preparing errors while PS level 1 members tend to commit opinion verification errors. Further, PS level 2 members tend to commit stylistic errors while PS level 1 members tend to commit content errors. Type of error and age was also found to be dependent. Members who are 24 years old and below are prone to committing opinion verification errors while members who are more than 24 years old are prone to committing copy preparation and statute verification errors. Type of error and tenure was also found to be related or dependent. Members with tenure of below 2.5 years tend to commit opinion verification errors while members who are with the company for 2.5 years or more tend to commit copy preparing errors.

iii

Analysis of variance showed that there is no significant difference in the performance of among the six Waves that comprised the Teams in terms of total cases and average page per hour production. On the other hand, regression analysis found no linear relationship between average per hour production and members age, income, and tenure. It also found no linear relationship between members total errors and their age, income, and tenure.

Overall, members of the Cases-Generalist Team seem to be performing well. However, based on research results, it is recommended that review should be conducted for Team H on statute verification and for Team M on opinion verification and copy preparing. It is also suggested that review on stylistics should be conducted for PS level 2 and for members with tenure of more than 2.5 years. Finally, more intensive training should be conducted for new hires on opinion verification as this seems to be a weakness that should be addressed.

iv

INTRODUCTION
Nature and Scope of the Study Thomson Reuters is an intelligent information provider. It publishes credible information both in printed and online publications covering financial, governance, risk, and compliance, intellectual property, legal, media, science, and tax and accounting products and services. These products and services have their respective departments that work closely on the information to be provided to its customers. Its values Business is Global, Customers are the Heart of Everything, People Make the Difference, and Performance Matters all aim to uphold quality work from its employees in order to satisfy its customers.

The Legal Department of Thomson Reuters has strategically moved some of its operations in Manila last June 2009. With a huge amount of work being transitioned from employees in the US who have been working for several years to Manila employees who have just learned how to process information, the company aims to keep its work at its best quality.

This case study focuses on one of the workflow in the Legal Department in the Manila branch, which is the Cases-Generalist workflow. Due to a big number of members, the CasesGeneralist workflow is divided into two (2) sub-teams with a total of 35 members. The team works on live Federal and state cases in all US jurisdictions. The case being reviewed goes through three (3) processes before it is submitted and published online at Westlaw.com:

1) Statute verification: Team members check the correct citation of state and Federal laws; 2) Opinion verification: Team members check the correct citations to previous court cases; and 3) Copy Preparation: Team members check the spacing, indention, spelling errors, more on how the document looks like.

Each member of the team is tasked to achieve a certain metrics or quota within a day. The metrics depend on the workflow he or she is assigned to. Since statute verification is the 1

fastest to do as not all cases cite a lot of laws, statute verifiers are tasked with a larger metrics to follow. The longest process is the copy preparation as copy preparers do not look for available citations to verify, but rather verify the whole case for any typographical errors. Metrics also depend on the position of the team members in the company. The higher the hierarchy, the higher metrics he or she must achieve within the day. To ensure the quality of work against the quantity of work, a quarterly audit is held by their mentors in the US. These mentors follow a certain audit criteria (See Appendices) and send feedback on the errors of the team members. Team members are always expected to have very minimal errors to no errors at all. Both quantity and quality of work are set in each of the team members annual objectives affecting their annual bonus and appraisal. The study seeks to evaluate the performance of team

members against benchmarks and other variables such as age and tenure, and to assess factors that may affect how errors occur in the workflow spectrum.

This study is observational in nature as no treatments will be applied to the elements of the study, i.e. members of the cases-generalists teams. It is also a retrospective study as data have been collected about the elements (or sample) with regards to performance outcomes that have already taken place. With regards to scope, the sample comprises of 28 cases-generalist team members (14 members each from the two sub-teams) whose performances for two quarters of the year will be assessed and compared. It should be noted that complete audit data is available only for the 2nd and 3rd quarter of 2012 as no audit was conducted during the 1st quarter and the 4th quarter is still ongoing.

PROBLEM DEFINITION
Purpose of the Study

The general objective of this study was to evaluate the performance of the two subteams of the Cases Generalist Team vis--vis performance benchmarks and other independent variables such as gender and tenure. Specifically, the study aimed to: 1) To assess and compare the performance of two sub-teams of the Cases-Generalist Team on statute verification, opinion verification, and copy preparation workflows; 2) To determine the process area where higher numbers of errors have been committed to recommend ways for improvement of the workflow process; 3) To identify factors which significantly affect the number of errors committed by team members in each of the workflow process; and 4) To provide feedback to the Team and Management on the result of the performance assessment for the purpose of further improving effectiveness and efficiency;

Significance of the Study While Thomson Reuters use regular metrics to assess individual performance against performance benchmarks, there have been no existing studies or written analysis on the performance of the Cases Generalist Team of Thomson Reuters that compares the performance of sub-teams, and that examines the relationship between performance, workflow errors and variables such as gender and tenure. The primary beneficiary of this study is the Legal Department of the company who will learn from the implications of the results of this study. These will serve as input on what they can do to maintain or improve, if needed, the workflow performance of the Cases Generalist Team. Each team member provides raw data to their managers on how many cases and how many pages they processed and how long they processed them. The managers are tasked to summarize all of the members production statements and present it to higher management to

see whether the team is hitting targets or having backlogs into the cases being sent from the US. No other study with regards as to why they perform as such had been conducted in the past. This study will also give team members a statistical analysis of their production outputs and workflow errors. With the results that will be analyzed in this study, team members will gain insights on how they perform as a team. This study will also provide the readers a general idea of the workflow of the Cases Generalist Team. This will make them appreciate what it is like to be in a team of young, competitive and enthusiastic professionals, who work on tedious U.S. cases. US cases or cases in general are very delicate products as the information they contain serve as basis for the credibility of attorneys, basis of court decisions, and ultimately, basis of justice. Studying performance on a workflow that involves this delicate information is relevant as implications are significant.

Current Knowledge about the Problem

Basically, each member knows about his or her accomplishment, since at the end of the week they submit a summary performance to their manager. It also known to them how the other team is performing based on performance metrics. But by connecting one variable to another, the study will give them a sense of how their performance can be further evaluated. Because the study is observational in nature, it will answer questions that provide statistical inferences about performance and relationships among variables of interest.

Preliminary Hypotheses

A number of assumptions have been considered in formulating the preliminary hypotheses. First, the training provided by Thomson Reuters to members of the cases generalist term was deemed to be standardized and to be effective in providing requisite skills and knowledge. Second, weekly performance metrics provided to all team members is succeeding in driving peak performance and compliance with benchmarks among all team members.

Therefore, the following are the preliminary hypotheses of the study: 1) There is no difference in the performance of two sub-teams of the Cases-Generalist Team on statute verification, opinion verification, and copy preparation workflows with regards to transition phase benchmarks; 2) There is no difference in the workflow errors of the two Cases-Generalist sub-teams; and 3) There is no difference in the performance and workflow errors vis--vis variables such as age, gender, income and tenure.

METHODOLOGY
This study made use of secondary data in determining the performance of the CasesGeneralist workflow. Weekly, each team member sends his or her production statements composed of the number of cases and pages made across all the jurisdictions he or she handles and how long it took him or her to finish everything. From these raw data, the managers of each team consolidate all information per team into one Microsoft Excel spreadsheet. The researchers were able to ask the secondary data from the managers. There was a limitation on getting the raw data from each of the team members as not all of them were able to keep their data from the time period that the researchers wanted to explore. Also, it was very time consuming to gather data from each of the 35 members of the team. The secondary data of the managers truly reflects the raw data. It is complete and comprehensive, thus the researchers opted to use it. The complete data was filtered down to two (2) quarter, the second and third quarters of 2012, as data is complete for these. Due to the busy work schedule of the US counterparts of the teams during the first quarter, they were not able to audit cases. Meanwhile, the fourth quarter is still on-going. The secondary data were provided by the CasesGeneralist workflow team managers who sent the data through electronic mail. The Cases Generalist Team of Thomson Reuters was purposively chosen since one of the members of our group (Group 2 of Managerial Statistics class) belongs to it. In addition to the statistical exercise which results can be utilized by Thomson Reuters, we also want to have an

appreciation of what it is like to work in a multinational company, which handles foreign accounts, specifically federal and state cases. The Cases-Generalist Team is composed of 35 members. It is divided into two subteams: Team H and Team M. From each team, we purposely selected 14 members who have complete records comprising of audit and performance information. The following shows the list of variables, the way these will be measured, and the analysis that will be performed. Each heading (i.e., A, B, C) corresponds to an Excel file that was provided by Thomson Reuters CasesGeneralist Team Manager.

A. Individual Production Summaries Variable 1. Team 2. Name 3. Wave 4. Function 5. PS Level 6. Total Cases 7. Total Number of Pages 8. Production Hours 9. Average Pages per Hour 10. Transition Phase Benchmark Class Qualitative Qualitative Qualitative Qualitative Qualitative Quantitative (discrete) Quantitative (discrete) Quantitative (continuous) Quantitative (continuous) Quantitative (discrete) Level of Measurement Nominal Nominal Ordinal Nominal Ordinal Ratio Ratio Ratio Ratio Ratio

Definition 1. Team- Composed of Teams H and M 2. Name- Name of the Team Member 3. Wave- Indicates when the member was hired (wave 1 being the earliest, wave 7 being the last. The company hires until wave 9. 4. Function- Each team has three functions: a. SV - statute verification - process of verifying and styling statutes/laws cited in a case b. OV - opinion verification - process of verifying case laws cited in a case c. CP - copy preparing - process of copy editing and preparing the case for publication

5. PS Level- Publishing Specialist - position label or hierarchy. 6. Total Cases- Total number of cases processed 7. Total Number of Pages- Number of pages per case 8. Production Hours- Number of hours spent doing the function 9. Average Pages per Hour - computed as total number of pages divided by production hours 10. Transition Phase Benchmark daily performance metrics; the higher the position, the higher the benchmark is set. Measurement 1. Mean of total cases, total number of pages, production hours, per team 2. Mean of total cases, total number of pages, production hours, per wave 3. Mean of total cases, total number of pages, production hours, per function 4. Mean of total cases , total number of pages, production hours, per PS level

B. Audits/Analysis of Performance Errors Variable 1. Team 2. Name 3. Wave 4. Function 5. PS Level 6. Copy Preparing Errors 7. Opinion Verification Errors 8. Stylistic Errors 9. Content Errors 10. Major Content Errors 11. Overall Quality Rating Class Qualitative Qualitative Qualitative Qualitative Qualitative Quantitative (discrete) Quantitative (discrete) Quantitative (discrete) Quantitative (discrete) Quantitative (discrete) Quantitative (continuous) Level of Measurement Nominal Nominal Ordinal Nominal Ordinal Ratio Ratio Ratio Ratio Ratio Interval

Measurement 1. Means and standard deviation of all types of errors 2. Means and standard deviation of each type of error (Copy Preparing, Statute Verification, Opinion Verification)

3. Means and standard deviation of each subtypes of error (Stylistics, Content, Major Content) 4. Means of each type and subtype of error per team, per gender, per PS level, per age range, per tenure duration

C.

Member Information

Variable 1. ID 2. Name 3. Role 4. Wave 5. Team 6. Age 7. Tenure 8. Income Measurement 1. Count of members (by name)

Class Qualitative Qualitative Qualitative Qualitative Qualitative Qualitative Quantitative Quantitative

Level of Measurement Nominal Nominal Nominal Nominal Nominal Ordinal Interval Interval

2. Count of roles (copy preparing, statute verification, opinion verification) 3. Count of wave 4. Count of Team membership 5. Mean, median, frequency distribution, etc. of age 6. Mean, median, frequency distribution, etc. of tenure 7. Mean, median, frequency distribution, etc. of income

Analytical Procedure Descriptive statistics was used in analyzing the performance basis of each team. Frequency distributions and proportions based on the frequency tables were computed. The average output of each team (based on total cases, number of pages, production hours, average pages per hour and number of errors) were also calculated. To present a comparison on the performance of the two teams based on output and errors, two-sample hypothesis testing will be used. Two-sample hypothesis testing is statistical 8

analysis designed to test if there is a difference between two means from two different populations. In this case, a two-tailed test regarding the differences will be applied. Since the samples of size is less than 30 and are taken from normally-distributed populations, a t-test was used to test the difference between the population means 1 and 2: 1. Comparison of the performance of two teams by: a) Total Cases Submitted b) Total Number of Pages per case c) Production Hours d) Average Pages/Hour 2. Comparison of the performance of members Using Total Cases Submitted Average Pages per Hour and Total Number of Errors per Type of Error by:

a) Gender b) Age (24 Years Old and Below, Above 24 Years Old) c) Tenure (below 2.5 years, 2.5 years and more) d) PS Level A two-sample z-test was used to test the difference between two sample means and

when a large sample (at least 30) is randomly selected from each population and the samples are independent. When the samples are less than 30, t-test was used to test the difference between the sample means. 3. Total Number of Errors Committed by:

a) Team b) Gender c) PS Level d) Age (24 Years Old and Below, Above 24 Years Old) e) Tenure (below 2.5 years, 2.5 years and more)

To assess the performance (using the average pages per hour) of each team against the Transition Phase Benchmark (target), one-sample hypothesis testing was used. A one sample test is a hypothesis test for answering questions about the mean where the data are a random sample of independent observations from an underlying normal distribution N. In this case, an upper-tailed test regarding the differences was applied. The t-test for the mean is a statistical test for a population mean. The t-test can be used when the population is normal or nearly normal, is unknown, and n < 30. 1. Members of Statute Verification vs. Transition Phase Benchmark 2. Members of Copy Preparing vs. Transition Phase Benchmark 3. Members of Opinion Verification vs. Transition Phase Benchmark To check if there is a difference in total cases processed and average pages per hour of waves 3,5,6,7 and 8.5, analysis of variance (ANOVA) was used. ANOVA is used to compare means between three or more groups, in this case, the period when the employee was hired (wave). The ANOVA F-Test is a comparison of the average variability between groups to the average variability within groups. The variability within each group is a measure of the spread of the data within each of the groups. The variability between groups is a measure of the spread of the group means around the overall mean for all groups. In this case, a single-factor ANOVA analysis was used.

The Chi-Square statistic compares the counts of categorical responses between independent groups. In this study, The Chi-square distribution was used to test the independence between types as well as subtypes of errors committed vis--vis team membership, gender, PS level, age, and tenure. It should be noted that in comparing the counts of categorical responses between PS levels, PS level 3 (the highest level) was not included since there was only one person in this category who performed just one function. Further, the test was not used to compare categorical responses according to Wave since not all Waves perform all functions (i.e. copy preparing, statute verification, opinion verification).

10

Regression analysis was done to estimate or predict the value of the dependent variable (average pages per hour and total number of errors committed) on the basis of known independent variables (age, tenure, salary, gender). The equation that was used is given below: Y= f(X1, X2. Xn ), Where: Y=Average Pages per Hour Number of Errors Committed X1=Gender X2=Age X3=Income X4=Tenure All the data processing and analysis were done with the help of MS Excel and its add-in PH-Stat. In all the statistical tests performed in this study, a 95% CI is used, which indicates that we are 95% certain where the true unknown parameter lies.

11

DATA PRESENTATION
Descriptive Statistics of Key Variables

This part of the study is a graphical presentation of the different frequency distributions of the factors examined in the case. Mainly, the researchers used the Excel Chart Wizard to produce the figures below. In Figure 1, the number of people aged 24 years and below is faintly higher in Team H than Team M. However, those aged above 24 years are higher in Team M than Team H. Also, the bar graph depicts that there are more 24 years old and below employees than above 24 years old employees. Figure 1. Frequency Distribution by Age

Age
10 9 8 7 6 5 4 3 2 1 0 Team H Team M

24 years old and below Above 24 years old

Figure 2 portrays that there are more females than males in the Cases-Generalist workflow. Females are slightly higher in number in Team H than Team M, while there are slightly more males in Team M than Team H.

12

Figure 2. Frequency Distribution by Gender

Gender
12 10 8 6 4 2 0 Team H Team M Female Male

Figure 3 shows that income ranging from 19K-25K and above 25K are equally distributed in both Team H and Team M. There is no difference in the frequency distribution of the given range of incomes. Figure 3. Frequency Distribution by Income

Income
8 7 6 5 4 3 2 1 0 Team H Team M 19K-25K Above 25K

13

In Figure 4, Team H employees with tenure 2.5 years and above outweighs those with tenure below 2.5 years; however, distributions of below 2.5 years and 2.5 years and above are equivalent in Team M. Team H has lower number of employees who have worked below 2.5 years in the company than Team M, and it has higher number of employees who have worked 2.5 years and above in the company than Team M. Figure 4. Frequency Distribution by Tenure

Tenure
12 10 8 Below 2.5 years 6 4 2 0 Team H Team M 2.5 years and above

Figure 5 demonstrates a pie chart of the distribution of the three functions in the CasesGeneralist workflow. From the graph, copy preparers are significantly higher in count than the other two functions. Between the other two, opinion verifiers are greater in number than statute verifiers.

14

Figure 5. Frequency Distribution by Function

Function

Statute Verifiers Opinion Verifiers Copy Preparers

Figure 6 presents the distribution by wave. The team does not have a member from wave 4 as being absent in the pie chart. The largest wave is wave 3 while the least member in a wave is wave 1 having only one sample. Waves 2 and 5 are gently comparable in proportion. There is a wave identified as 8.5 because members from this wave were hired in between waves 8 and 9. The name 8.5 is just an identifier.

Wave
8.5 7 1 2 wave 1 wave 2 wave 3 6 3 5 wave 4 wave 5 wave 6 wave 7

15

Figure 7 reveals that half of the Cases-Generalist team ranks PS 2. The number of PS 1 team members is not very far from that of PS 2. PS 3 level, which requires more tenure and more experience, has a minute number of members. Figure 7. Frequency Distribution by PS Level

PS Level

PS 1 PS 2 PS 3

Table 1 below shows the output computed by the mean number of cases, pages, production hours, and the average pages per hour as performed by Teams H and M. The researchers used excel function =AVERAGE(data_array) for computing the different means. Table 1. Team Performance by Output Team H 1,473.7857 13,316.5714 379.3571 38.0291 Team M 1,777.5714 12,842.6429 367.5536 35.3055

Mean number of cases Mean number of pages Mean production hours Average pages per hour

There were a total of 101 errors encountered in 168 instances whereby a team member can commit any number of errors or not commit any error at all. In this study therefore, zero (0) is absolute, indicating that a team member did not commit any error during performance on a case. The main types of errors are Copy Preparing errors which are errors committed in the process of copy editing and preparing the case for publication, Opinion Verification errors which 16

are errors committed in the process of verifying case laws cited in a case, and Statute Verification errors which are errors committed in the process of verifying and styling statutes and federal and state laws (as compared to case laws which result from promulgation of cases). The errors are further sub-categorized into stylistic, content, and major content errors. In tabular form, this is illustrated as follows:

Figure 8. Error Types and Sub-Types


Copy Preparing Stylistic Content Major Content Statute Verification Major Stylistic Content Content Opinion Verification Major Stylistic Content Content

Stylistic errors are errors with regards to writing style, for example in copy preparing, a stylistic error would be a wrong mnemonic or a heading not styled correctly. A content error is a minor error regarding content, for example paragraphs split or merged incorrectly. A major content error is a significant error with regards to content, for example missing text. Tables below show the numbers of errors and sub-type errors committed by team, gender, PS level, age and tenure. Table 2. Number of Errors Committed by Team
Team H M Total Copy Prep 20 44 64 Statute Ver 12 2 14 Opinion Ver 8 15 23 Total 40 61 101

Table 3. Number of Errors Committed by Gender


Gender Female Male Total Copy Prep 49 15 64 Statute Ver 3 11 14 Opinion Ver 13 10 23 Total 65 36 101

Table 4. Number of Errors Committed by PS Level


PS Level 1 2 Total Copy Prep 15 49 64 Statute Ver 3 11 14 Opinion Ver 19 4 23 Total 37 64 101

17

Table 5. Number of Errors Committed by Age


Error Copy Prep Statute Ver Opinion Ver Total 24 and Below 29 5 19 53 Above 24 35 9 4 48 Total 64 14 23 101

Table 6. Number of Errors Committed by Tenure


Error Copy Prep Statute Ver Opinion Ver Total Below 2.5 Yrs. 6 3 16 25 2.5 Yrs and Above 58 11 7 76 Total 64 14 23 101

Table 7. Number of Sub-Type Errors Committed by Team


Team H M Total Stylistic 26 40 66 Content 13 18 31 Major Content 1 3 4 Total 40 61 101

Table 8. Number of Sub-Type Errors Committed by Gender


Team Female Male Total Stylistic 46 20 66 Content 16 15 31 Major Content 3 1 4 Total 65 36 101

Table 9. Number of Sub-Type Errors Committed by PS Level


PS Level 1 2 Total Stylistic 19 47 66 Content 17 14 31 Major Content 1 3 4 Total 37 64 101

18

Table 10. Number of Sub-Type Errors Committed by Age


Error Stylistics Content Major Content Total 24 and Below 33 19 1 53 Above 24 33 12 3 48 Total 66 31 4 101

Table 11. Number of Sub-Type Errors Committed by Tenure


Error Stylistics Content Major Content Total Below 2.5 Yrs. 11 14 0 25 2.5 Yrs and Above 55 17 4 76 Total 66 31 4 101

STATISTICAL ANALYSIS
Analysis of Performance based on Output This analysis compared the number of errors between teams H and I, females and males, PS levels 1 and 2, members with ages 24 and >24 (median age is 23.5 years), and members with tenures of <2.5 and 2.5 (median tenure is 2.5 years). The comparison utilized the test for the differences between two population means. A two-sample t-test was used to test the difference between two sample means was less than 30. and in cases when at least one sample size

Two-tailed Hypothesis Testing: Performance Comparison by Team With regards to Team, results show that there is no significant difference between the total cases processed, total production hours and average pages per hour processed by Teams H and M as follows:

Total Cases Processed Ho : 1-2 = 0(There is no significant difference in the total cases processed between Team H and Team M)

19

Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.61038 with a p-value of 0.546907 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of Team H and Team M has no significant difference.

Total Production Hours Ho: 1-2 = 0 (There is no significant difference in the total production hours between Team H and Team M) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: 0.241579 with a p-value of 0.811001 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean production hours of Team H and Team M has no significant difference.

Average Pages per Hour Ho: 1-2 = 0 (There is no significant difference in the average pages per hour between Team H and Team M) Ha: 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: 0.233897 with a p-value of 0.816897 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The average pages per hour of Team H and Team M has no significant difference.

20

Two-tailed Hypothesis Testing: Performance Comparison by Gender With regards to gender, results show that there is no significant difference between the total cases processed, total production hours and average pages per hour processed by males and females as follows:

Total Cases Processed Ho: 1-2 = 0 (There is no significant difference in the total cases processed between males and females) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.90453 with a p-value of 0.374018 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of Male and Female has no significant difference.

Total Production Hours Ho: 1-2 = 0 (There is no significant difference in the total production hours between males and females) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: 0.276631 with a p-value of 0.784251 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean production hours of Male and Female has no significant difference.

Average Pages per Hour Ho: 1-2 = 0 (There is no significant difference in the average pages per hour between males and females) Ha: 1-2 0

21

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -1.50266 with a p-value of 0.144976 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The average pages per hour of Team H and Team M has no significant difference.

Two-tailed Hypothesis Testing: Performance Comparison by Age With regards to age, results show that there is no significant difference between the total cases processed, total production hours and average pages per hour processed by members with ages below 24 years old and those above 24 years old, as follows:

Total Cases Processed Ho: 1-2 = 0 (There is no significant difference in the total cases processed between Members 24 years old and below and Members above 24 years old) Ha: 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.19228 with a p-value of 0.849014 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of members 24 years old and below and members above 24 years old has no significant difference.

Total Production Hours Ho: 1-2 = 0 (There is no significant difference in the total production hours between members 24 years old and below and members above 24 years old) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.99193 with a p-value of 0.370185 Conclusion: We do not have sufficient evidence to Reject Ho 22

Interpretation: The mean production hours of members 24 years old and below and members above 24 years old has no significant difference.

Average Pages per Hour Ho : 1-2 = 0 (There is no significant difference in the average pages per hour between members 24 years old and below and members above 24 years old) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.10319 with a p-value of 0.918607 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The average pages per hour of members 24 years old and below and members above 24 years old has no significant difference.

Two-tailed Hypothesis Testing: Performance Comparison by Tenure With regards to tenure, results show that there is no significant difference between the total cases processed, total production hours and average pages per hour processed by members working for the company 2.5 years and below, and those working for more than 2.5 years, as follows:

Total Cases Processed Ho : 1-2 = 0 (There is no significant difference in the total cases processed between members working for 2.5 years and below and members working for more than 2.5 years) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: 0.49073 with a p-value of 0.627734 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of members working for 2.5 years and below and members working above 2.5 years has no significant difference. 23

Total Production Hours Ho : 1-2 = 0 (There is no significant difference in the total production hours between members working for 2.5 years and below and members working above 2.5 years) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: 0.901425 with a p-value of 0.375637 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean production hours of members working for 2.5 years and below and members working above 2.5 years has no significant difference.

Average Pages per Hour Ho : 1-2 = 0 (There is no significant difference in the average pages per hour between members working for 2.5 years and below and members working above 2.5 years) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.05 or t <-2.05 t Test Statistic: -0.25543 with a p-value of 0.800403 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The average pages per hour of members working for 2.5 years and below and members working above 2.5 years has no significant difference.

Two-tailed Hypothesis Testing: Performance Comparison by PS Level With regards to PS Level, results show that there is no significant difference between the total cases processed, and average pages per hour processed by members with age below 24 years old and those above 24 years old. However, these 2 PS Levels have significant difference in total production hours, as follows:

24

Total Cases Processed Ho : 1-2 = 0 (There is no significant difference in the total cases processed between PS Level 1 and PS Level 2) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.06 or t <-2.06 t Test Statistic: 0.6850 with a p-value of 0.499893 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of PS Level 1 and PS Level 2 has no significant difference.

Total Production Hours Ho : 1-2 = 0 (There is no significant difference in the total production hours

between PS Level 1 and PS Level 2) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.06 or t <-2.06 t Test Statistic: 2.6397 with a p-value of 0.014351 Conclusion: We have sufficient evidence to Reject Ho Interpretation: The mean production hours of PS Level 1 and PS Level 2 has significant difference.

Average Pages per Hour Ho : 1-2 = 0 (There is no significant difference in the average pages per hour between PS Level 1 and PS Level 2) Ha : 1-2 0

Decision Rule: Reject Ho if t>2.06 or t <-2.06 t Test Statistic: -0.5136 with a p-value of 0.612232 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The average pages per hour of PS Level 1 and PS Level 2 has no

25

significant difference.

Analysis of Performance based on Errors

This analysis compared the number of errors between teams H and M, females and males, PS levels 1 and 2, members with ages 24 and >24 (median age is 23.5 years), and members with tenures of <2.5 and 2.5 (median tenure is 2.5 years). The comparison utilized the test for the differences between two population means. A two-sample z-test was used to test the difference between two sample means and in cases when both sample sizes were

at least 30. When at least one of the sample sizes was less than 30, the t-test was used to test the difference between the sample means.

Two-tailed Hypothesis Testing: Performance Comparison based on Error by Team With regards to Team, results show that there is a significant difference between the means of the total number of errors by Team H and M as follows:

Ho: 1-2 = 0 (There is no difference between the means of errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -3.056953112 with a p-value of 0.002235992 Conclusion: We have sufficient evidence to Reject Ho Interpretation: There is a significant difference between the means of errors of Team H and M. Team M committed more errors on average compared to Team H.

Two-tailed Hypothesis Testing: Performance Comparison based on Sub-Type Error by Team However, we found no difference in the means of errors between Team H and M with regards to copy preparing, statute verification, opinion verification, as well as in the error sub- categories of stylistic, content and major content, as follows:

26

Copy Preparing Errors Ho: 1-2 = 0 (There is no difference between the means of Copy Preparing errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -1.056571785 with a p-value of 0.290707063 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Copy Preparing errors of Team H and I.

Statute Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Statute Verification errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.05 or t<-2.05 t-Test Statistic: 1.600846701 with a p-value of 0.120634727 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Statute Verification errors of Team H and M.

Opinion Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Opinion Verification errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.01 or t<-2.01 t-Test Statistic: -1.070744185 with a p-value of 0.289870562 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Opinion Verification errors of Team H and M. 27

Stylistic Errors Ho: 1-2 = 0 (There is no difference between the means of Stylistic errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -1.383697 with a p-value of 0.172144547 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Stylistic errors of Team H and M.

Content Errors Ho: 1-2 = 0 (There is no difference between the means of Content errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -0.671939362 with a p-value of 0.504487101 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Content errors of Team H and M. However, since the p-value is .5045 there is an even chance that there may be a difference between the means of content errors of Teams H and M.

Major Content Errors Ho: 1-2 = 0 (There is no difference between the means of Major Content errors of Team H and M) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -1.028998973 with a p-value of 0.308067294

28

Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Major Content errors of Team H and M.

Two-tailed Hypothesis Testing: Gender Performance Comparison based on Error and Sub-Type Errors by Gender With regards to gender, we found no differences in the means of all errors by females and males, and no differences in the means of copy preparing, statute verification, and opinion verification errors between females and males. We also did not find any differences in the means of stylistic, content, and major content errors between them.

All Errors Ho: 1-2 = 0 (There is no difference between the means of errors of Females and Males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -0.526909716 with a p-value of 0.598256286 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There no difference between the means of errors of females and males. However, the p-value of .60 indicates a more than an even chance that the means of errors of males and females are different.

Copy Preparing Errors Ho: 1-2 = 0 (There is no difference between the means of Copy Preparing errors of females and males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>1.98 or t<-1.98 t-Test Statistic: -0.480414919 with a p-value of 0.632124523 Conclusion: We do not have sufficient evidence to Reject Ho

29

Interpretation: There is no difference between the means of Copy Preparing errors of females and males. However, the p-value of .63 signifies a high probability that there is a difference between the means of copy preparing errors of females and males.

Statute Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Statute Verification errors of females and males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.05 or t<-2.05 t-Test Statistic: -1.131657389 with a p-value of 0.2673778 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Statute Verification errors of females and males.

Opinion Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Opinion Verification errors of females and males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.013 or t<-2.013 t-Test Statistic: -0.429993966 with a p-value of 0.669207779 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Opinion Verification errors of females and males. However, with a p-value of almost .70, there is a high probability that there is a difference in the opinion verification errors of females and males.

Stylistic Errors Ho: 1-2 = 0 (There is no difference between the means of Stylistic errors of females and males) 30

Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: 0.252687417 with a p-value of 0.801469191 with a p-value of 0.669207779 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Stylistic errors of females and males. As above, a p-value of .70 indicates that there may be a difference between the stylistic errors of males and females.

Content Errors Ho: 1-2 = 0 (There is no difference between the means of Content errors of females and males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -1.471678375 with a p-value of 0.146911968 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Content errors of females and males.

Major Content Errors Ho: 1-2 = 0 (There is no difference between the means of Major Content errors of females and males) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: 0.311984936 with a p-value of 0.75625345 Conclusion: We do not have sufficient evidence to Reject Ho

31

Interpretation: There is no difference between the means of Major Content errors of females and males. However, the p-value of .75 points to a high probability that the means of major content errors of females and males may be different.

Two-tailed Hypothesis Testing: Performance Comparison based on Error and Error Sub-Type by PS Level With regards to PS level, we found no differences in the means of all errors by PS 1 and 2, and no differences in the means of copy preparing, statute verification, and opinion verification errors between PS 1 and 2. We also did not find any differences in the means of content and major content errors. However, we found a significant

difference in the means of stylistic errors between the two.

All Errors Ho: 1-2 = 0 (There is no difference between the means of errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -1.699715694 with a p-value of 0.089184416 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There no difference between the means of errors of PS 1 and 2.

Copy Preparing Errors Ho: 1-2 = 0 (There is no difference between the means of Copy Preparing errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -1.297438559 with a p-value of 0.194480333 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Copy Preparing errors of PS 1 and 2.

32

Statute Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Statute Verification errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.07 or t<-2.07 t-Test Statistic: -1.848642734 with a p-value of 0.077995146 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Statute Verification errors of PS 1 and 2.

Opinion Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Opinion Verification errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.01 or t<-2.01 t-Test Statistic: 0.613122416 with a p-value of 0.542815173 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Opinion Verification errors of PS 1 and 2. A p-value of .54, however, may mean a difference between the means of opinion verification errors of PS 1 and 2.

Stylistic Errors Ho: 1-2 = 0 (There is no difference between the means of Stylistic errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.006 or t<-2.006 t-Test Statistic: -5.260204902 with a p-value of 2.75379E-06 Conclusion: We have sufficient evidence to Reject Ho

33

Interpretation: There is a significant difference between the means of Stylistic errors of PS 1 and 2. PS 2 commits more error of average than PS 1.

Content Errors Ho: 1-2 = 0 (There is no difference between the means of Content errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.006 or t<-2.006 t-Test Statistic: 0.560134181 with a p-value of 0.577793178 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Content errors of PS 1 and 2. However, a p-value of .58 may mean that there is a difference in the content errors of PS 1 and 2.

Major Content Errors Ho: 1-2 = 0 (There is no difference between the means of Major Content errors of PS 1 and 2) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.006 or t<-2.006 t-Test Statistic: -0.953126014 with a p-value of 0.344935928 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Major Content errors of PS 1 and 2.

Two-tailed Hypothesis Testing: Performance Comparison based on Error and SubType Errors by Gender With regards to age, we found no differences in the means of all errors by members with ages 24 and >24, and no differences in the means of copy preparing, statute

34

verification, and opinion verification errors. We also did not find any differences in the means of stylistic, content, and major content errors between them.

All Errors Ho: 1-2 = 0 (There is no difference between the means of errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -1.204555793 with a p-value of 0.228374831 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There no difference between the means of errors of members with ages 24 and >24.

Copy Preparing Errors Ho: 1-2 = 0 (There is no difference between the means of Copy Preparing errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: 1.208035 with a p-value of 0.367459911 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Copy Preparing errors of members with ages 24 and >24.

Statute Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Statute Verification errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.05 or t<-2.05 t-Test Statistic: -1.50449125 with a p-value of 0.143652565 35

Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Statute Verification errors of members with ages 24 and >24.

Opinion Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Opinion Verification errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.01 or t<-2.01 t-Test Statistic: 0.613122863 with a p-value of 0.542814879 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Opinion Verification errors of members with ages 24 and >24. A p-value of .54 may mean that there is a difference in the opinion verification errors of members with ages 24 and >24.

Stylistic Errors Ho: 1-2 = 0 (There is no difference between the means of Stylistic errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -1.432828166 with a p-value of 0.157671142 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Stylistic errors of members with ages 24 and >24.

Content Errors Ho: 1-2 = 0 (There is no difference between the means of Content errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 36

Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: 0.048934198 with a p-value of 0.96115228 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Content errors of members with ages 24 and >24. However, a p-value of .96 points out that it is very likely that there is a difference between the content errors of members with ages 24 and >24.

Major Content Errors Ho: 1-2 = 0 (There is no difference between the means of Major Content errors of members with ages 24 and >24) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -1.522039965 with a p-value of 0.1338339 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Major Content errors of members with ages 24 and >24.

Two-tailed Hypothesis Testing: Performance Comparison based on Error and SubType Errors by Gender

With regards to tenure, we found no differences in the means of all errors by members with tenure of <2.5 years and 2.5 years, and no differences in the means of copy preparing, statute verification, and opinion verification errors. We also did not find any differences in the means of content and major content errors. However, we found a significant difference in the means of stylistic errors between the two.

All Errors Ho: 1-2 = 0 (There is no difference between the means of errors of members with tenure of <2.5 years and 2.5 years)

37

Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if z>1.96 or z<-1.96 Z-Test Statistic: -1.778162781 with a p-value of 0.075377124 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There no difference between the means of errors of members with tenure of <2.5 years and 2.5 years.

Copy Preparing Errors Ho: 1-2 = 0 (There is no difference between the means of Copy Preparing errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.5 or t<-2.5 t-Test Statistic: -1.50187087 with a p-value of 0.13671039 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Copy Preparing errors of members with tenure of <2.5 years and 2.5 years.

Statute Verification Errors Ho: 1-2 = 0 (There is no difference between the means of Statute Verification errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.05 or t<-2.05 t-Test Statistic: -1.131657389 with a p-value of 0.2673778 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Statute Verification errors of members with tenure of <2.5 years and 2.5 years.

Opinion Verification Errors

38

Ho: 1-2 = 0 (There is no difference between the means of Opinion Verification errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.05 or t<-2.05 t-Test Statistic: 0.508575431 with a p-value of 0.613480475 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Opinion Verification errors of members with tenure of <2.5 years and 2.5 years. A p-value of .61, points out the more than even probability that there is a difference between the opinion verification errors of members with tenure of <2.5 years and 2.5 years.

Stylistic Errors Ho: 1-2 = 0 (There is no difference between the means of Stylistic errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -2.71691399 with a p-value of 0.008835663 Conclusion: We have sufficient evidence to Reject Ho Interpretation: There is a significant difference between the means of Stylistic errors of members with tenure of <2.5 years and 2.5 years. Members with tenure of 2.5 years commit more error on average.

Content Errors Ho: 1-2 = 0 (There is no difference between the means of Content errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: 0.823073277 with a p-value of 0.414086436

39

Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Content errors of members with tenure of <2.5 years and 2.5 years.

Major Content Errors Ho: 1-2 = 0 (There is no difference between the means of Major Content errors of members with tenure of <2.5 years and 2.5 years) Ha: 1-2 0 = 0.05 Decision Rule: Reject Ho if t>2.005 or t<-2.005 t-Test Statistic: -0.616452087 with a p-value of 0.540186401 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: There is no difference between the means of Major Content errors of members with tenure of <2.5 years and 2.5 years. A p-value of .54 indicates at least an even chance that there is a difference between the means of major content errors of members with tenure of <2.5 years and 2.5 years.

Analysis of Performance based on Output vs. Benchmark

This analysis assessed the performance (using the average pages per hour) of each team against the Transition Phase Benchmark (target) using One-Sample Hypothesis Testing. In this case, an upper-tailed test regarding the differences was applied. The t-test was used since, is unknown, and n < 30. Upper-tailed Hypothesis Testing: Performance of members of each function vs. Benchmark All Members of Statute Verification Process were able to meet the required number of pages per hour.

40

Statute Verification Ho : 1 50 pages per hour (Members of this function does not meet the transition

phase benchmark) Ha : 1 50 pages per hour

Decision Rule: Reject Ho if t > 6.3137 t Test Statistic: 6.8421 with a p-value of 0.046195261 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Members of Statute Verification are able to comply with the benchmark of 50 pages per hour.

Copy Preparing All Members of Copy Preparing Process were able to meet the required number of pages per hour.

Ho : 1

17 pages per hour (Members of this function does not meet the transition

phase benchmark) Ha : 1 17 pages per hour

Decision Rule: Reject Ho if t > 1.833 t Test Statistic: 7.608 with a p-value of 1.64992E-05 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Members of Copy Preparing are able to comply with the benchmark of 17 pages per hour.

Ho : 1

14 pages per hour (Members of this function does not meet the transition

phase benchmark) Ha : 1 14 pages per hour

Decision Rule: Reject Ho if t > 2.1318 t Test Statistic: 4.3442 with a p-value of 0.006154527

41

Conclusion: We have sufficient evidence to Reject Ho Interpretation: Members of Copy Preparing are able to comply with the benchmark of 14 pages per hour.

Opinion Verification Not all members of Opinion Verification Process were able to meet the required number of pages per hour. New hires with a benchmark of 9 pages per hour, were not able to comply.

Ho : 1

23 pages per hour (Members of this function does not meet the transition

phase benchmark) Ha : 1 23 pages per hour

Decision Rule: Reject Ho if t > 6.3138 t Test Statistic: 10.2018 with a p-value of 0.031101994 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Members of Opinion Verification are able to comply with the benchmark of 23 pages per hour.

Ho : 1

18 pages per hour (Members of this function does not meet the transition

phase benchmark) Ha : 1 18 pages per hour

Decision Rule: Reject Ho if t > 2.1318 t Test Statistic: 6.5727 with a p-value of 0.001386506 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Members of Opinion Verification are able to comply with the benchmark of 18 pages per hour.

Ho : 1

9 pages per hour (Members of this function does not meet the transition

phase benchmark)

42

Ha : 1 9 pages per hour

Decision Rule: Reject Ho if t > 6.3138 t Test Statistic: 2.4599 with a p-value of 0.123534919 Conclusion: We dont have sufficient evidence to Reject Ho Interpretation: Members of Opinion Verification are not able to comply with the benchmark of 9 pages per hour. Analysis of Performance based on Independence between Error Types

The Chi-square distribution test was utilized to determine the independence between types as well as subtypes of errors and team membership, gender, PS level, age, and tenure. We found out that type of error is related to team membership as follows:

Ho: Type of error and team are independent. Ha: Type of error and team are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 14.53533309 with a p-value of 0.000697738 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Type of error and team membership are related.

Comparing the observed and expected frequencies, it seems that Team H is prone to committing statute verification errors while Team M is prone to committing copy preparing and opinion verification errors. On the other hand, it was found out that there is no relation between error subtypes (stylistics, etc.) and team membership. However, the p-value (.80) is big enough to put the result into question, as follows.

Ho: Error subtype and team are independent. Ha: Error subtype and team are dependent or related. = 0.05 Decision Rule: Reject Ho if 43

Chi-Square Test Statistic: 0.428329067 with a p-value of 0.80721556 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: Error subtype and team membership are not related.

It was also found out that type of error is related to gender. By comparing observed frequencies to expected frequencies, females tend to commit copy preparing errors while males tend to commit status verification and opinion verification errors.

Ho: Type of error and gender are independent. Ha: Type of error and gender are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 16.01916677 with a p-value of 0.000332263 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Type of error and gender are related.

Again, it was found out that error subtypes and gender are not related. Ho: Error subtypes and gender are independent. Ha: Error subtypes and gender are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 3.212824168 with a p-value of 0.200606082 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: Error subtypes and gender are related.

With regards to PS level, we found out that type of error is related to a members PS level as follows: Ho: Type of error and PS level are independent. Ha: Type of error and PS level are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 27.13810142 with a p-value of 1.27949E-06 44

Conclusion: We have sufficient evidence to Reject Ho Interpretation: Type of error and PS level are related.

Also, error subtypes and PS level were found to be dependent or related. Ho: Error subtypes and PS level are independent. Ha: Error subtypes and PS level are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 6.409321769 with a p-value of 0.040572658 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Error subtypes and PS level are related.

Comparing observed and expected frequencies, PS level 2 members tend to commit copy preparing errors while PS level 1 members tend to commit opinion verification errors. Further, PS level 2 members tend to commit stylistic errors while PS level 1 members tend to commit content errors.

Type of error and age was also found to be related or dependent. Members who are 24 years old and below are prone to committing opinion verification errors while members who are more than 24 years old are prone to committing copy preparation and statute verification errors.

Ho: Type of error and age are independent. Ha: Type of error and age are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 11.26805616 with a p-value of 0.003574149 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Type of error and age are related.

On the other hand, error subtypes and age were found to be independent. Ho: Error subtypes and age are independent. Ha: Error subtypes and age are dependent or related. 45

= 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 2.338852328 with a p-value of 0.310545092 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: Error subtypes and age are not related.

Type of error and tenure was also found to be related or dependent. Comparing observed and expected frequencies, members with tenure of below 2.5 years tend to commit opinion verification errors while members who are with the company for 2.5 years or more tend to commit copy preparing errors. The Chi-Square test result for error subtypes and tenure is not presented because expected frequency assumptions were not met. Ho: Type of error and tenure are independent. Ha: Type of error and tenure are dependent or related. = 0.05 Decision Rule: Reject Ho if Chi-Square Test Statistic: 33.00653339 with a p-value of 6.80334E-08 Conclusion: We have sufficient evidence to Reject Ho Interpretation: Type of error and tenure are related.

Analysis of Performance based on Output of Multiple Sample Means

A single-factor ANOVA Analysis was used to check if there is a significant difference in total cases processed and average pages per hour of waves 3, 5, 6, 7 and 8.5. Analysis of Variance: Performance by Wave There is no significant difference in the performance among the 6 waves classified.

Total Cases Processed Ho: 1= 2= 3= 4= 5= 6 (Mean number of cases processed by all waves are the same) Ha : At least one 1 differs from the others

46

Decision Rule: Reject Ho if F 2.68 F Test Statistic: 0.904793 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean total cases processed of Waves 2, 3, 5, 6, 7 and 8.5 are the same.

Average Pages per Hour Ho: 1= 2= 3= 4= 5= 6 (Mean pages per hour processed by all waves are the same) Ha : At least one 1 differs from the others

Decision Rule: Reject Ho if F 2.68 F Test Statistic: 0.734984 Conclusion: We do not have sufficient evidence to Reject Ho Interpretation: The mean pages per hour processed of Waves 2, 3, 5, 6, 7 and 8.5 are the same.

Analysis of Factors Affecting the Performance based on Output and Error

In analyzing how the performance of the two teams are influenced by different factors. Factors identified as Age, Gender, Income and Tenure were used.

Regression Analysis: Average Pages per Hour Regression Statistics R Square 0.168122876 P-value 0.677799 0.486851 0.31748 0.138449 0.185843

Intercept Gender: (0-F), (1-M) Age Income Tenure (Years)

47

Interpretation: 16.82% of the variation in the Average Pages per Hour processed is explained by the independent variables: Gender, Age, Income and Tenure. This indicates that these predictor variables explain very little variation in the response variable.

Gender: p-value=0.486851 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Average pages per hour processed and the members Gender. Age: p-value=0.31748 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Average pages per hour processed and the members Age. Income: p-value=0.138449 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Average pages per hour processed and the members Income. Tenure: p-value=0.185843 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Average pages per hour processed and the members Tenure.

Regression Analysis: Total Number of Errors Committed Regression Statistics R Square 0.114443 P-value 0.235803 0.29054 0.231104 0.14784 0.705145

Intercept Gender: (0-F), (1-M) Age Income Tenure (Years) Interpretation:

11.44% of the variation in the Total errors committed is explained by the independent variables: Gender, Age, Income and Tenure. This indicates that these predictor variables explain very little variation in the response variable.

48

Gender: p-value=0.29054 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Total Errors Committed and the members Gender. Age: p-value=0.231104 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Total Errors Committed and the members Age. Income: p-value=0.14784 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Total Errors Committed and the members Income. Tenure: p-value=0.705145 is greater than =0.05, we fail to find significant statistical evidence for a linear relationship between the Total Errors Committed and the members Tenure.

Conclusion

The researchers found out that the performance of the two teams by Age, Gender, Tenure, PS Level, based on their outputs has no significant difference. Being mostly fresh graduates upon entry, team members are afraid to commit mistakes and aim to do good work to fit in the corporate world. Also, the company values highlight quality of work among its employees, thus affect the performance to be at the same caliber. Further, with the performance metrics comparison among teams and with the fresh working attitude of these members who are mostly young professionals, members tend to be in a competitive mode, thus they try to keep up with the performance of one another. There is no high variation among the age, income and tenure of the members. Thus, performance of the members based on output and errors are not affected much by these factors. The researchers merit these results to the delicateness of the product. Information is the product of the company. Legal departments products are cases used by professionals in courts, corporate counsels, governments, and schools. Upon arriving in the company, employees undergo one-week on-boarding that discusses the companys history, products, and the 49

American legal system. After which, employees are distributed to their teams and go through a thorough training in their specific functions for two straight months. For the Cases-Generalist workflow, there are several reviews held per jurisdiction per employee until he or she passes the review period. If he or she fails to pass the quality control, he or she remains on-review for the specific jurisdiction he or she is assigned to. Ultimately, from the first week up to the walls of the company, the tag line We help the legal system perform better. Every day. Worldwide. and its values are continually instilled to the minds of the employees. Thus, there is no significant difference on the outputs and with their performance vis--vis their level of hierarchy in the company, income, tenure, or age. The team knows and understands the product of the company and what the company wants from him or her.

50

RECOMMENDATIONS
It is recommended that review should be conducted for Team H on statute verification and for Team M on opinion verification and copy preparing. It is also suggested that review on stylistics should be conducted for PS level 2 and for members with tenure of more than 2.5 years. Though stylistic errors may be characterized as human-natured errors because people tend to see past the details when working quickly, the team must devise a strategy or train the members to have better eye for details. Better yet, in the future, members may suggest to construct a button in their work tool that would help minimize stylistic errors. Moreover, more intensive training should be conducted for new hires on opinion verification as this seems to be a weakness that should be addressed. A higher sample size for the analysis could provide a variation among the age and income and could give a sound analysis of the working group. A higher sample for the audited cases may also be recommended to increase the true mean of errors.

51

APPENDICES
Audit Criteria

Stylistic Missing Punctuation Judges names not capped Wrong mnemonic used Titles not marked as italic N.J.-book titles arent italic CT.- forgot to add at in partial cites DEL.- forgot to add both AP1 lines Spacing errors Cite markup in wrong place Forgot to add Full names to Judges/Justices that are on our list Statutes not in correct order in headnotes Incorrect tagging Headings not styled correctly (Bold, Italic, etc.) Forgot to add Assigned, etc. to synopsis

Content Cite has wrong serial no. Missing case itself references Missing lower court info. in syn. Missing prior reports Missing attorneys Forgot to markup cites in syn. & headnotes Wrong title used for Judge/Justice Forgot to add cities to attorneys Paragraphs split or merged incorrectly Missing fileline information

Major Content Wrong judge/justice Appendix missed Dissent or Concur. Opinions missing Wrong statute used in headnotes Wrong court in synopsis Missing Text Released a case when it should have been held Missing a title Wrong volume or page information in citations Wrong Courtline

Statute should have had a year in parenthesis Partial cite, missed markup Popular name incorrect or incomplete Statute style incorrect Statute Template style invalid Spell check errors/not corrected Incorrect parallel added or not deleted (ex. Ala & WestlawOnly cites)

Missing Star-paging

Ratings:
A rating of 1 to 5 is given with 5 being the highest. 0 Errors = 5 1-2 Stylistic Errors = 4 3 Stylistic and/or 1 Content Error = 3 >3 Stylistic and/or 2 Content Errors = 2 >2 Content and/or 1 Major Content Errors = 1 >1 Major Content Error = 0 Difficulty of Case = High, Medium or Low

52

PH-Stat Results

Two-tailed Hypothesis Testing Results


Performance Comparison by Team (Total Cases) (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 14 1473.786 1006.569 14 1777.571 1566.729

13 13 26 1733910 -303.785 -0.61038

-2.05553 2.055529 0.546907

Performance Comparison by Team (Total Number of Pages) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 14 Sample Mean 13316.57 Sample Standard Deviation 9570.975 Population 2 Sample Sample Size 14 Sample Mean 12842.64 Sample Standard Deviation 10351.36 Intermediate Calculations Population 1 Sample Degrees of Freedom

13

53

Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

13 26 99377108 473.928 0.125782

-2.05553 2.055529 0.900872

Performance Comparison by Team(Total Production Hours) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 14 Sample Mean 379.357 Sample Standard Deviation 149.122 Population 2 Sample Sample Size 14 Sample Mean 367.553 Sample Standard Deviation 105.77 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

13 13 26 16712.33 11.804 0.241579

-2.05553 2.055529 0.811001

Performance Comparison by Team(Average Pages/Hour) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 14 Sample Mean 38.029 Sample Standard Deviation 30.682

54

Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

14 35.305 30.943

13 13 26 949.4272 2.724 0.233897

-2.05553 2.055529 0.816897

Performance Comparison by Gender(Total Cases) (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 19 1472.052 1145.244 9 1950 1609.508

18 8 26 1705101 -477.948 -0.90453

-2.05553 2.055529 0.374018

55

Performance Comparison by Gender(Total Number of Pages) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 19 Sample Mean 11093.95 Sample Standard Deviation 6759.974 Population 2 Sample Sample Size 9 Sample Mean 17271.56 Sample Standard Deviation 13828.24 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

18 8 26 90473505 -6177.61 -1.60501

-2.05553 2.055529 0.12057

Performance Comparison by Gender(Total Production Hours) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 19 Sample Mean 378.105 Sample Standard Deviation 140.303 Population 2 Sample Sample Size 9 Sample Mean 363.639 Sample Standard Deviation 99.93 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means

18 8 26 16700.65 14.466

56

t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0.276631

-2.05553 2.055529 0.784251

Performance Comparison by Gender(Average Pages/Hour) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 19 Sample Mean 29.341 Sample Standard Deviation 24.714 Population 2 Sample Sample Size 9 Sample Mean 47.496 Sample Standard Deviation 39.026 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

18 8 26 891.4731 -18.155 -1.50266

-2.05553 2.055529 0.144976

Performance Comparison by Age (Total Cases) (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean

0 0.05 17 1586.941 1301.376 11 1685.545

57

Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

1362.553

16 10 26 1756261 -98.604 -0.19228

-2.05553 2.055529 0.849014

Performance Comparison by Age(Total Number of Pages) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 17 Sample Mean 12285.12 Sample Standard Deviation 8853.542 Population 2 Sample Sample Size 11 Sample Mean 14307.46 Sample Standard Deviation 11418.84 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

16 10 26 98387014 -2022.34 -0.5269

-2.05553 2.055529 0.602732

Performance Comparison by Age(Total Production Hours) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05

58

Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

17 355.794 121.029 11 400.75 136.978

16 10 26 16230.69 -44.956 -0.91193

-2.05553 2.055529 0.370185

Performance Comparison by Age(Average Pages/Hour) (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value

0 0.05 17 34.529 34.559 11 35.702 18.248

16 10 26 863.0418 -1.173 -0.10319

-2.05553 2.055529

59

p-Value Do not reject the null hypothesis

0.918607

Performance Comparison by Tenure(Total Cases) (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 9 1803.556 1078.581 19 1541.421 1414.241

8 18 26 1742619 262.135 0.49073

-2.05553 2.055529 0.627734

Performance Comparison by Tenure(Total Number of Pages) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 9 Sample Mean 13344.11 Sample Standard Deviation 8170.03 Population 2 Sample Sample Size 19 Sample Mean 12954.32 Sample Standard Deviation 10673.06 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom

8 18

60

Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

26 99401884 389.795 0.096618

-2.05553 2.055529 0.923771

Performance Comparison by Tenure(Total Production Hours) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 9 Sample Mean 405 Sample Standard Deviation 98.474 Population 2 Sample Sample Size 19 Sample Mean 358.513 Sample Standard Deviation 138.387 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

8 18 26 16242.09 46.487 0.901425

-2.05553 2.055529 0.375637

Performance Comparison by Tenure(Average Pages/Hour) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 9 Sample Mean 32.948 Sample Standard Deviation 27.649 Population 2 Sample

61

Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

19 36.133 32.122

8 18 26 949.5596 -3.185 -0.25543

-2.05553 2.055529 0.800403

Performance Comparison by PS Level(Total Number of Cases) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 12 Sample Mean 1681.75 Sample Standard Deviation 973.582 Population 2 Sample Sample Size 14 Sample Mean 1358.143 Sample Standard Deviation 1363.856 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

11 13 24 1441993 323.607 0.685022

-2.0639 2.063899 0.499893

Performance Comparison by PS Level(Total Number of Pages) (assumes equal population variances)

62

Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis 0 0.05 12 13083.5 7309.261 14 11386.5 10229.96

11 13 24 81173105 1697 0.478789

-2.0639 2.063899 0.636422

Performance Comparison by Position(Total Production Hours) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 12 Sample Mean 427.645 Sample Standard Deviation 122.063 Population 2 Sample Sample Size 14 Sample Mean 309.286 Sample Standard Deviation 106.652 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic

11 13 24 12990.15 118.359 2.639746

63

Two-Tail Test Lower Critical Value Upper Critical Value p-Value Reject the null hypothesis

-2.0639 2.063899 0.014351

Performance Comparison by PS Level (Average Pages/Hour) (assumes equal population variances) Data Hypothesized Difference 0 Level of Significance 0.05 Population 1 Sample Sample Size 12 Sample Mean 30.594 Sample Standard Deviation 24.155 Population 2 Sample Sample Size 14 Sample Mean 36.815 Sample Standard Deviation 35.447 Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

11 13 24 948.0197 -6.221 -0.51359

-2.0639 2.063899 0.612232

Comparison of All Errors of Team H and M Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Population Standard Deviation Population 2 Sample Sample Size Sample Mean Population Standard Deviation

0 0.05 84 0.457143 0.943102 84 0.943102 1.110549

64

Intermediate Calculations Difference in Sample Means Standard Error of the Difference in Means Z-Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Reject the null hypothesis

-0.48596 0.158968 -3.05695

-1.95996 1.959964 0.002236

CP Errors by Team Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Population Standard Deviation Population 2 Sample Sample Size Sample Mean Population Standard Deviation Intermediate Calculations Difference in Sample Means Standard Error of the Difference in Means Z-Test Statistic

0 0.05 42 0.428571 1.069045 48 0.6875 1.255632

-0.25893 0.245065 -1.05657

Two-Tail Test Lower Critical Value -1.95996 Upper Critical Value 1.959964 p-Value 0.290707 Do not reject the null hypothesis

Opinion Ver Errors by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation

0 0.05 24 0.333333 0.701964

65

Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

24 0.625 1.13492

23 23 46 0.890398 -0.29167 -1.07074

-2.0129 2.012896 0.289871

Statute Ver Errors by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value

0 0.05 18 0.666667 1.028992 12 0.166667 0.389249

17 11 28 0.702381 0.5 1.600847

-2.04841 2.048407

66

p-Value Do not reject the null hypothesis

0.120635

Opinion Ver Errors by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 24 0.333333 0.701964 24 0.625 1.13492

23 23 46 0.890398 -0.29167 -1.07074

-2.0129 2.012896 0.289871

Stylistics Error by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation

0 0.05 28 0.928571 1.245096 28 1.428571 1.451144

67

Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

27 27 54 1.828041 -0.5 -1.3837

-2.00488 2.004879 0.172145

Content Errors by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 28 0.464286 0.838082 28 0.642857 1.129218

27 27 54 0.988757 -0.17857 -0.67194

-2.00488 2.004879 0.504487

68

Major Content Errors by Team (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 28 0.035714 0.188982 28 0.107143 0.31497

27 27 54 0.06746 -0.07143 -1.029

-2.00488 2.004879 0.308067

CP Errors by Gender (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of

0 0.05 72 0.680556 1.196549 18 0.833333 1.248529

71

69

Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

17 88 1.456282 -0.15278 -0.48041

-1.98729 1.98729 0.632125

SV Errors by Gender (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 12 0.25 0.452267 18 0.611111 1.036901

11 17 28 0.733135 -0.36111 -1.13166

-2.04841 2.048407 0.267378

Opinion Ver Errors by Gender (assumes equal population variances) Data

70

Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 30 0.433333 0.8172 18 0.555556 1.149026

29 17 46 0.908937 -0.12222 -0.42999

-2.0129 2.012896 0.669208

Stylistic Error by Gender (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance

0 0.05 38 1.210526 1.417228 18 1.111111 1.278275

37 17 54 1.890622

71

Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0.099415 0.252687

-2.00488 2.004879 0.801469

Content Errors by Gender (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 38 0.421053 0.792927 18 0.833333 1.294786

37 17 54 0.958577 -0.41228 -1.47168

-2.00488 2.004879 0.146912

Major Content Error by Gender (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean

0 0.05 38 0.078947

72

Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0.273276 18 0.055556 0.235702

37 17 54 0.068659 0.023391 0.311985

-2.00488 2.004879 0.756253

Copy Prep Errors by PS Level Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Population Standard Deviation Population 2 Sample Sample Size Sample Mean Population Standard Deviation Intermediate Calculations Difference in Sample Means Standard Error of the Difference in Means Z-Test Statistic

0 0.05 30 0.5 0.973795 60 0.816667 1.295254

-0.31667 0.244071 -1.29744

Two-Tail Test Lower Critical Value -1.95996 Upper Critical Value 1.959964 p-Value 0.19448 Do not reject the null hypothesis

73

Statute Ver Errors by PS Level (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0.05 12 0.25 0.452267 12 0.916667 1.1645

11 11 22 0.780303 -0.66667 -1.84864

-2.07387 2.073873 0.077995

Opinion Ver Errors by PS Level (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of

0 0.05 36 0.527778 1.027789 12 0.333333 0.651339

35 11

74

Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

46 0.905194 0.194445 0.613122

-2.0129 2.012896 0.542815

Stylistic Errors by PS Level (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Reject the null hypothesis

0 0.05 26 0.730769 1.002305 28 2.586207 1.516662

25 27 52 1.677356 -1.85544 -5.2602

-2.00665 2.006647 2.75E-06

Content Error by PS Level (assumes equal population variances) Data Hypothesized Difference Level of Significance

0 0.05

75

Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

26 0.653846 1.16421 28 0.5 0.83887

25 27 52 1.017012 0.153846 0.560134

-2.00665 2.006647 0.577793

Major Content Error by PS Level (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic

0 0.05 26 0.038462 0.196116 28 0.107143 0.31497

25 27 52 0.070002 -0.06868 -0.95313

76

Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

-2.00665 2.006647 0.344936

All Errors by Age Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Population Standard Deviation Population 2 Sample Sample Size Sample Mean Population Standard Deviation Intermediate Calculations Difference in Sample Means Standard Error of the Difference in Means Z-Test Statistic

0 0.05 102 0.519608 1.05051 66 0.727273 1.116939

-0.20767 0.1724 -1.20456

Two-Tail Test Lower Critical Value -1.95996 Upper Critical Value 1.959964 p-Value 0.228375 Do not reject the null hypothesis

Copy Prep Errors by Age Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Population Standard Deviation Population 2 Sample Sample Size Sample Mean Population Standard Deviation

0 0.05 48 0.604167 1.198218 42 0.833333 1.208035

77

Intermediate Calculations Difference in Sample Means Standard Error of the Difference in Means Z-Test Statistic

-0.22917 0.254278 -0.90124

Two-Tail Test Lower Critical Value -1.95996 Upper Critical Value 1.959964 p-Value 0.36746 Do not reject the null hypothesis

Statute Ver Errors by Age (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 18 0.277778 0.460889 12 0.75 1.215431

17 11 28 0.709326 -0.47222 -1.50449

-2.04841 2.048407 0.143653

Opinion Ver Errors by Age (assumes equal population variances) Data Hypothesized Difference Level of Significance

0 0.05

78

Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

36 0.527778 1.055221 12 0.333333 0.492366

35 11 46 0.905193 0.194445 0.613123

-2.0129 2.012896 0.542815

Stylistic Error by Age (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic

0 0.05 34 0.970588 1.336783 22 1.5 1.371478

33 21 54 1.82353 -0.52941 -1.43283

79

Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

-2.00488 2.004879 0.157671

Content Error by Age (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 34 0.558824 1.049998 22 0.545455 0.911685

33 21 54 0.99698 0.013369 0.048934

-2.00488 2.004879 0.961152

Major Content Error by Age (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean

0 0.05 34 0.029412

80

Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0.171499 22 0.136364 0.35125

33 21 54 0.065954 -0.10695 -1.52204

-2.00488 2.004879 0.133834

Copy Prep Errors by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value

0 0.05 18 0.333333 0.766965 72 0.805556 1.274218

17 71 88 1.423612 -0.47222 -1.50187

-1.98729

81

Upper Critical Value p-Value Do not reject the null hypothesis

1.98729 0.13671

Statute Ver Errors by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 12 0.25 0.452267 18 0.611111 1.036901

11 17 28 0.733135 -0.36111 -1.13166

-2.04841 2.048407 0.267378

Opinion Ver Errors by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation

0 0.05 30 0.533333 1.105888 18 0.388889 0.607685

82

Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

29 17 46 0.907488 0.144444 0.508575

-2.0129 2.012896 0.61348

Stylistic Errors by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Reject the null hypothesis

0 0.05 20 0.55 0.825578 36 1.527778 1.482972

19 35 54 1.665226 -0.97778 -2.71691

-2.00488 2.004879 0.008836

83

Content Error by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

0 0.05 20 0.7 1.260743 36 0.472222 0.810154

19 35 54 0.984671 0.227778 0.823073

-2.00488 2.004879 0.414086

Major Content Error by Tenure (assumes equal population variances) Data Hypothesized Difference Level of Significance Population 1 Sample Sample Size Sample Mean Sample Standard Deviation Population 2 Sample Sample Size Sample Mean Sample Standard Deviation Intermediate Calculations Population 1 Sample Degrees of

0 0.05 20 0 1 36 0.111111 0.318728

19

84

Freedom Population 2 Sample Degrees of Freedom Total Degrees of Freedom Pooled Variance Difference in Sample Means t Test Statistic Two-Tail Test Lower Critical Value Upper Critical Value p-Value Do not reject the null hypothesis

35 54 0.417696 -0.11111 -0.61645

-2.00488 2.004879 0.540186

Upper-tailed Hypothesis Testing Results


Statute Verification(Benchmark:50 pages/hour) Data =

Null Hypothesis Level of Significance Sample Size Sample Mean Sample Standard Deviation

50 0.05 2 118.875 14.236

Intermediate Calculations Standard Error of the Mean 10.06637214 Degrees of Freedom 1 t Test Statistic 6.842087602 Upper-Tail Test Upper Critical Value p-Value Reject the null hypothesis Calculations Area For one-tailed tests: TDIST value 0.046195 1-TDIST value 0.953805

6.313751515 0.046195261

Copy Preparing(Benchmark: 17 pages/hour) Data Null Hypothesis = Level of Significance Sample Size Sample Mean Sample Standard Deviation

17 0.05 10 21.648 1.932

Intermediate Calculations Standard Error of the Mean 0.610952044 Degrees of Freedom 9

85

t Test Statistic Upper-Tail Test Upper Critical Value p-Value Reject the null hypothesis

7.607798429 Calculations Area For one-tailed tests: TDIST value 1.65E-05 1-TDIST value 0.999984

1.833112933 1.64992E-05

Copy Preparing(Benchmark:14 pages/hour) Data Null Hypothesis = Level of Significance Sample Size Sample Mean Sample Standard Deviation

14 0.05 5 24.337 5.333

Intermediate Calculations Standard Error of the Mean 2.384990105 Degrees of Freedom 4 t Test Statistic 4.33418989 Upper-Tail Test Upper Critical Value p-Value Reject the null hypothesis Calculations Area For one-tailed tests: TDIST value 0.006155 1-TDIST value 0.993845

2.131846786 0.006154527

Opinion Verification(Benchmark: 23 pages/hour) Data Null Hypothesis = Level of Significance Sample Size Sample Mean Sample Standard Deviation

23 0.05 2 39.671 2.311

Intermediate Calculations Standard Error of the Mean 1.634123771 Degrees of Freedom 1 t Test Statistic 10.20179762 Upper-Tail Test Upper Critical Value p-Value Reject the null hypothesis Calculations Area For one-tailed tests: TDIST value 0.031102 1-TDIST value 0.968898

6.313751515 0.031101994

86

Opinion Verification(Benchmark: 18 pages/hour) Data Null Hypothesis = Level of Significance Sample Size Sample Mean Sample Standard Deviation

18 0.05 5 29.402 3.879

Intermediate Calculations Standard Error of the Mean 1.734741537 Degrees of Freedom 4 t Test Statistic 6.572737066 Upper-Tail Test Upper Critical Value p-Value Reject the null hypothesis Calculations Area For one-tailed tests: TDIST value 0.001387 1-TDIST value 0.998613

2.131846786 0.001386506

Opinion Verification(Benchmark: 9 pages/hour) Data =

Null Hypothesis Level of Significance Sample Size Sample Mean Sample Standard Deviation

9 0.05 2 20.583 6.697

Intermediate Calculations Standard Error of the Mean 4.735494114 Degrees of Freedom 1 t Test Statistic 2.445996072 Upper-Tail Test Upper Critical Value p-Value Do not reject the null hypothesis Calculations Area For one-tailed tests: TDIST value 0.123535 1-TDIST value 0.876465

6.313751515 0.123534919

87

Chi-square Test of Independence Results


Ho: Type of error and team are independent. Ha: Type of error and team are dependent or related. Observed Frequencies Type of Error CP SV H 20 12 I 44 2 Total 64 14

Calculations OP 8 15 23 Total 40 61 101 fo-fe -5.34653 6.455446 -1.10891 5.346535 -6.45545 1.108911

Team

Team

Expected Frequencies Type of Error CP SV OP H 25.34653 5.544554 9.108911 I 38.65347 8.455446 13.89109 Total 64 14 23

Total 40 61 101

(fo-fe)^2/fe 1.127785 7.515983 0.134998 0.739531 4.928513 0.088523

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 14.53533 p -Value 0.000698 Reject the null hypothesis Expected frequency assumption is met.
Ho: Error subtypes and team are independent. Ha: Error subtypes and team are dependent or related. Observed Frequencies Error Subtypes Stylistics Content Major Content H 26 13 1 I 40 18 3 Total 66 31 4 Expected Frequencies Error Subtypes Stylistics Content Major Content H 26.13861 12.27723 1.584158416 I 39.86139 18.72277 2.415841584 Total 66 31 4

Calculations Total 40 61 101 fo-fe -0.13861 0.722772 -0.58416 0.138614 -0.72277 0.584158

Team

Team

Total 40 61 101

(fo-fe)^2/fe 0.000735 0.04255 0.215408 0.000482 0.027902 0.141251

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 0.428329 p -Value 0.807216 Do not reject the null hypothesis Expected frequency assumption is met.

88

Ho: Type of error and gender are independent. Ha: Type of error and gender are dependent or related. Observed Frequencies Type of Error Gender CP SV Female 49 3 Male 15 11 Total 64 14

Calculations OV 13 10 23 Total 65 36 101 fo-fe 7.811881 -6.0099 -7.81188 6.009901 -1.80198 1.80198

Gender

Expected Frequencies Type of Error CP SV OV Female 41.18812 9.009901 14.80198 Male 22.81188 4.990099 8.19802 Total 64 14 23

Total 65 36 101

(fo-fe)^2/fe 1.481628 4.008802 0.219372 2.675162 7.238115 0.396087

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 16.01917 p -Value 0.000332 Reject the null hypothesis Expected frequency assumption is met.

Ho: Error subtypes and gender are independent. Ha: Error subtypes and gender are dependent or related. Observed Frequencies Error Subtypes Gender Stylistics Content Major Content Female 46 16 3 Male 20 15 1 Total 66 31 4 Expected Frequencies Error Subtypes Stylistics Content Major Content Female 42.47525 19.9505 2.574257426 Male 23.52475 11.0495 1.425742574 Total 66 31 4

Calculations Total 65 36 101 fo-fe 3.524752 -3.9505 0.425743 -3.52475 3.950495 -0.42574

Gender

Total 65 36 101

(fo-fe)^2/fe 0.292497 0.782257 0.070411 0.528119 1.412408 0.127131

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 3.212824 p -Value 0.200606 Do not reject the null hypothesis Expected frequency assumption is met.

89

Ho: Type of error and PS level are independent. Ha: Type of error and PS level are dependent. Observed Frequencies Type of Error CP SV 1 15 3 2 49 11 Total 64 14

Calculations OV 19 4 23 Total 37 64 101 fo-fe -8.44554 -2.12871 10.57426 8.445545 2.128713 -10.5743

PS Level

PS Level

Expected Frequencies Type of Error CP SV OV 1 23.44554 5.128713 8.425743 2 40.55446 8.871287 14.57426 Total 64 14 23

Total 37 64 101

(fo-fe)^2/fe 3.042251 0.883539 13.27063 1.758801 0.510796 7.672084

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 27.1381 p -Value 1.28E-06 Reject the null hypothesis Expected frequency assumption is met.
Ho: Error subtype and PS level are independent. Ha: Error subtype and PS level are dependent. Observed Frequencies Error Subtype Stylistic Content Major Content 1 19 17 1 2 47 14 3 Total 66 31 4 Expected Frequencies Error Subtype Stylistic Content Major Content 1 24.17822 11.35644 1.465346535 2 41.82178 19.64356 2.534653465 Total 66 31 4

Calculations Total 37 64 101 fo-fe -5.17822 5.643564 -0.46535 5.178218 -5.64356 0.465347

PS Level

PS Level

Total 37 64 101

(fo-fe)^2/fe 1.109012 2.804561 0.147779 0.641148 1.621387 0.085435

Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

0.05 2 3 2

Results Critical Value 5.991465 Chi-Square Test Statistic 6.409322 p -Value 0.040573 Reject the null hypothesis Expected frequency assumption is met.

90

Ho: Age and type of errors are independent. Ha: Age and type of errors are dependent. Observed Frequencies Age Below Above Type of Error 24 24 Copy Prep 29 35 Statute Ver 5 9 Opinion Ver 19 4 Total 53 48 Expected Frequencies Age Type of Error Copy Prep Statute Ver Opinion Ver Total Data Level of Significance Number of Rows Number of Columns Degrees of Freedom Below 24 33.58416 7.346535 12.06931 53 Above 24 30.41584 6.653465 10.93069 48 Total 64 14 23 101 (fo-fe)^2/fe 0.625727 0.690907 0.7495 0.827572 3.979889 4.394461

Calculations Total 64 14 23 101 fo-fe -4.58416 4.584158 -2.34653 2.346535 6.930693 -6.93069

0.05 3 2 2

Results Critical Value 5.991465 Chi-Square Test Statistic 11.26806 p-Value 0.003574 Reject the null hypothesis Expected frequency assumption is met.

Ho: Age and subtype errors are independent. Ha: Age and subtype errors are dependent. Observed Frequencies Age Below Above Error Subtype 24 24 Stylistics 33 33 Content 19 12 Major Content 1 3 Total 53 48

Calculations Total 66 31 4 101 fo-fe -1.63366 1.633663 2.732673 -2.73267 -1.09901 1.09901

91

Expected Frequencies Age Error Subtype Stylistics Content Major Content Total Data Level of Significance Number of Rows Number of Columns Degrees of Freedom Below 24 34.63366 16.26733 2.09901 53 Above 24 31.36634 14.73267 1.90099 48 Total 66 31 4 101 (fo-fe)^2/fe 0.07706 0.085087 0.459049 0.506867 0.575425 0.635365

0.05 3 2 2

Results Critical Value 5.991465 Chi-Square Test Statistic 2.338852 p-Value 0.310545 Do not reject the null hypothesis Expected frequency assumption is met.

Ho: Tenure and type of error are independent. Ha: Tenure and type of error are related. Observed Frequencies Tenure Below 2.5 Years 2.5 Years and Above 6 58 3 11 16 7 25 76 Expected Frequencies Tenure Below 2.5 Years 2.5 Years and Above 15.84158416 48.15841584 3.465346535 10.53465347 5.693069307 17.30693069 25 76

Type of Error Copy Prep Statute Ver Opinion Ver Total

Total 64 14 23 101

Calculations fo-fe -9.84158 9.841584 -0.46535 0.465347 10.30693 -10.3069

Type of Error Copy Prep Statute Ver Opinion Ver Total Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

Total 64 14 23 101

(fo-fe)^2/fe 6.114084 2.011212 0.062489 0.020556 18.66003 6.138166

0.05 3 2 2

Results Critical Value 5.991464547 Chi-Square Test Statistic 33.00653339 p -Value 6.80334E-08 Reject the null hypothesis Expected frequency assumption is met.

92

Ho: Tenure and error subtype are independent. Ha: Tenure and error subtype are related. Observed Frequencies Tenure Error Subtype Below 2.5 Yrs 2.5 Yrs and Above Stylistic 11 55 Content 14 17 Major Content 0 4 Total 25 76 Expected Frequencies Tenure Error Subtype Below 2.5 Yrs 2.5 Yrs and Above Stylistic 16.33663366 49.66336634 Content 7.673267327 23.32673267 Major Content 0.99009901 3.00990099 Total 25 76 Data Level of Significance Number of Rows Number of Columns Degrees of Freedom

Total 66 31 4 101

Calculations fo-fe -5.33663 5.336634 6.326733 -6.32673 -0.9901 0.990099

Total 66 31 4 101

(fo-fe)^2/fe 1.7433 0.573454 5.216493 1.715952 0.990099 0.32569

0.05 3 2 2

Results Critical Value 5.991464547 Chi-Square Test Statistic 10.56498868 p -Value 0.005079744 Reject the null hypothesis Expected frequency assumption is violated.

Analysis of Variance Results Anova: Single Factor (TOTAL CASES) SUMMARY Groups WAVE 2 WAVE 3 WAVE 5 WAVE 6 WAVE 7 WAVE 8.5

Count 5 7 6 3 4 2

Sum Average Variance 2888 577.6 6656.3 12487 1783.857 3229110 10894 1815.667 810757.5 4856 1618.667 301240.3 7389 1847.25 1539018 2206 1103 356168

93

ANOVA Source of Variation Between Groups Within Groups Total

SS 6254007 29030775 35284781

df 5 21 26

MS F P-value F crit 1250801 0.904793 0.496467 2.684781 1382418

Anova: Single Factor (AVERAGE PAGES per HOUR) SUMMARY Groups WAVE 2 WAVE 3 WAVE 5 WAVE 6 WAVE 7 WAVE 8.5

Count 5 7 6 3 4 2

Sum 128.2404 346.1575 163.4089 84.91717 188.2334 39.46339

Average Variance 25.64808 88.32649 49.45107 2353.45 27.23482 52.2334 28.30572 52.4829 47.05836 1594.914 19.7317 44.84555

ANOVA Source of Variation Between Groups Within Groups Total

SS 3442.126 19669.73 23111.85

df

MS F P-value F crit 5 688.4252 0.734984 0.605473 2.684781 21 936.6536 26

Simple Linear Regression Results

SUMMARY OUTPUT: Average Pages per Hour Regression Statistics Multiple R 0.410027896 R Square 0.168122876 Adjusted R Square 0.023448593 Standard Error 29.91208168 Observations 28

94

ANOVA df Regression Residual Total 4 23 27 SS 4158.99827 5 20578.8505 24737.8487 7 Standard Error 53.6590278 4 12.4025924 8 2.46616818 5 0.00138605 3 13.3625853 MS 1039.75 894.7326 F 1.16207 9 Significance F 0.353226936

Coefficients Intercept Gender: (0-F), (1-M) Age Income* Tenure (Years) 65.32342749 13.41751539 -3.033567051 0.002075731 -5.119544828

t Stat 1.21738 1.081832 -1.23007 1.497585 -0.38313

P-value 0.23580 3 0.29054 0.23110 4 0.14784 0.70514 5

Lower 95% 45.67872882 12.23920195 8.135224635 0.000791537 32.76215861

Upper 95% 176.3255 838 39.07423 272 2.068090 533 0.004943 22.52306 895

SUMMARY OUTPUT: Total Errors Committed Regression Statistics Multiple R 0.338294 R Square 0.114443 Adjusted R Square -0.03957 Standard Error 3.252047 Observations 28 ANOVA df Regression Residual Total 4 23 27 SS 31.43495 243.2436 274.6786 Standard Error 5.833819 95 MS 7.858737 10.57581 F 0.74308 6 Significance F 0.572463

Coefficients Intercept -2.45493

t Stat -0.42081

P-value 0.67779 9

Lower 95% -14.5231

Upper 95% 9.61324

Gender: (0-F), (1-M) Age Income* Tenure (Years)

0.952925 0.273981 -0.00023 1.981245

1.348412 0.268122 0.000151 1.452783

0.706701 1.021851 -1.53491 1.363759

0.48685 1 0.31748 0.13844 9 0.18584 3

-1.83648 -0.28067 -0.00054 -1.02406

3.742328 0.828634 8.04E-05 4.986555

BIBLIOGRAPHY
Two-sample test: http://org.elon.edu/econ/sac/twosample.htm One-sample test: http://org.elon.edu/econ/sac/onesample.htm Anova: http://www.isixsigma.com/tools-templates/analysis-of-variance-anova/how-compare-data-setsanova/ Regression: http://dss.princeton.edu/online_help/analysis/interpreting_regression.htm http://blog.excelmasterseries.com/2010/03/how-to-quickly-read-output-of-excels.html http://www.obgyn.cam.ac.uk/cam-only/statsbook/esc.html

96

S-ar putea să vă placă și