Sunteți pe pagina 1din 9

6/28/2017

Planning for Analysis


Analysis, Interpretation &
Presentation of Data Type of Type of
Data Formatting

Type of
Analysis

Planning for Analysis Type of Data & Formatting Technique


 A sound research plan successfully matches  Quantitative Data
these elements with the proper techniques  Must “quantify” the data
 Convert (“data reduce”) from collection format into
 Collect the type of data that is most appropriate numeric database
to answering your question and fits the other  Qualitative Data
parameters of your project (budget, personnel,  Must process the data (type/enter/describe)
etc.)  Convert from audio/video to text
 Combination
 Process each element as appropriate

Type of Data & Analysis Quantifying Data


 Coding
 Quantitative Data
 Counts, frequencies, tallies  Processing
 Statistical analyses (as appropriate)
 Qualitative Data
 Coding
 Patterns, themes, theory building
 Combination
 Process each element as appropriate

1
6/28/2017

Quantifying Data Quantifying Data


 Before we can do any kind of analysis, we need Some transformations are simple:
to quantify our data  Assign numeric representations to nominal or
 “Quantification” is the process of converting ordinal variables:
data to a numeric format  Turning male into “1” and female into “2”
 Convert social science data into a “machine-  Assigning “3” to Very Interested, “2” to Somewhat
readable” form, a form that can be read & Interested, “1” to Not Interested
manipulated by computer programs
 Assign numeric values to continuous variables:
 Turning born in 1973 to “35”
 Number of children = “02”

Developing Code Categories Coding Quantitative Data


Some data are more challenging. Open-ended  Goal – reduce a wide variety of information to a
responses must be coded. more limited set of variable attributes:
 “What is your occupation?”
 Use pre-established scheme: Professional, Managerial,
 Two basic approaches: Clerical, Semi-skilled, etc.
 Create a scheme after reviewing the data
 Begin with a coding scheme derived from the
 Assign value to each category in the scheme: Professional
research purpose. = 1, Managerial = 2, etc.
 Generate codes from the data.  Classify the response: “Secretary” is “clerical” and is
coded as “3”

Coding Quantitative Data Codebook Construction


 Points to remember: Purposes:
 If the data are coded to maintain a good amount  Primary guide used in the coding process.
of detail, they can always be combined (reduced)  Should note the value assigned to each variable
later attribute (response)
 However, if you start off with too little detail, you
 Guide for locating variables and interpreting
can‟t get it back
codes in the data file during analysis.
 If you‟re using a survey / questionnaire, it‟s a good
idea to do your coding on the form so that it can  If you‟re doing your own input, this will also
be entered properly (i.e. create a “codebook”) guide data set construction

2
6/28/2017

Entering Data Quantitative Analysis


 Data entry specialists enter the data into an SPSS  You should choose a level of analysis that is
data matrix, Excel spreadsheet, or ASCII file. appropriate for your research question
 Typically, work off a coded questionnaire
 You should choose the type of statistical analysis
appropriate for the variables you have
 Nominal/Categorical, Ordinal, or Continuous

Quantitative Levels of Analysis Univariate Analysis


 Univariate – simplest form, describe a case in  Describing a case in terms of the distribution of
terms of a single variable attributes that comprise it
 Bivariate – subgroup comparisons, describe a
case in terms of two variables simultaneously. Example:
 Multivariate – analysis of two or more variables  Gender - number of women, number of men
simultaneously
 You should ALWAYS begin your analysis by running
the basic univariate frequencies and checking to
be sure data were entered properly

Univariate Analysis Presenting Univariate Data


 Frequency distributions Goals:
 Provide reader with the fullest degree of detail
 Measures of central tendency regarding the data
 Mean, Median, Mode  Present data in a manageable from
 Simple and straightforward

3
6/28/2017

Subgroup Comparisons Bivariate Analysis


 Describe subsets of cases, subjects or  Describe a case in terms of two variables
respondents simultaneously
 Example:
 Gender
Examples
 Attitudes toward equality for men and women
 “Collapsing” response categories:  How does a respondent‟s gender affect his or her attitude
 Age categories, Open responses, etc toward equality for men and women?
 Crosstabulations/Correlations
 Handling “don't knows”
 Code separately, make missing if appropriate

Constructing Bivariate Tables Bivariate Analysis


 Divide cases into groups according to the  Bivariate Tables / Crosstabs are appropriate for
attributes of the independent variable. all types of variables, but the proper inferential
 Describe each subgroup in terms of attributes of statistic will vary by variable type
the dependent variable.
 Read the table by comparing the independent  Continuous variables are typically made into
variable subgroups in terms of a given attribute of categorical variables for this type of analysis
the dependent variable.  Recode variables
 DV goes in the rows, IV goes in the columns  Example: Create “Age” (18-34, 35-50, 51-65, 66+)

Appropriate Types of Analysis Bivariate Analysis: Correlations


 Bivariate correlation analysis is appropriate for
continuous variables (interval, ratio)
 Other types of variables are often recoded into
„Dummy‟ variables (value 0 or 1) for these
purposes
 Example: Gender becomes two variables „Male‟
(1=yes) & „Female‟ (1=yes)
 Present in Correlation Matrix

4
6/28/2017

Multivariate Analysis Multivariate Analysis


 Analysis of more than two variables  Ordinal (technically inappropriate but it
simultaneously happens), continuous, dummy variables
 Can be used to understand the relationship  Type of regression analysis will depend on the
between multiple variables more fully type of variables
 Most typical: Regression analysis  OLS (continuous)
 Logistic (other types)

Planning Your Analysis Planning your analysis


 Leave enough time for data entry and data  If your research question requires a level of
formatting analysis your variables won‟t allow, you‟ll need
 Can take much longer than you expect to transform them
 In your codebook – note the TYPE of variable for  Create „dummy‟ variables
each measurement/question  Collapse categories

 This will allow you to plan the proper levels and  Determine the level of significance acceptable &
types of analysis apply proper tests

Planning your analysis


 Proper planning will make things easier later  Data Presentation
 Take good notes on any transformations, etc.
that you do
 Save all the elements of your analysis programs

5
6/28/2017

Summarizing Data Basic guidelines


 Tables  Ensure graphic has a title
 Simplest way to summarize data  Label the components of your graphic
 Data are presented as absolute numbers or  Indicate source of data with date
percentages
 Provide number of observations (n=xx) as a
 Charts and graphs
reference point
 Visual representation of data
 Add footnote if more information is needed
 Data are presented as absolute numbers or
percentages

31 32

Tables Tables
 Frequency distribution: set of categories with  Relative Frequencies
numerical counts
number of values within an interval x 100
total number of values in the table
Year Number of births
1900 61
Year # births (n) Relative frequency
1901 58
(%)
1902 75
1900–1909 35 27
1910–1919 46 34
1920–1929 51 39
Total 132 100.0
33 34

Tables Charts and Graphs


 Percentage of births by decade between 1900
 Charts and graphs are used to portray:
and 1929
 trends,
 relationships, and
Year Number of births Relative frequency  comparisons
(n) (%)
 The most informative are simple and self-
1900–1909 35 27 explanatory
1910–1919 46 34
1920–1929 51 39
Total 132 100.0

35 36

6
6/28/2017

Use right type of graphics Bar chart


 Comparing categories
 Charts and graphs
 Bar chart: comparisons, categories of data 6

 Line graph: display trends over time 5


 Pie chart: show percentages or proportional 4
share
Site 1
3
Site 2
2 Site 3

0
Quarter 1 Quarter 2 Quarter 3 Quarter 4
37 38

Has the program met its goal?


Percentage of new enrollees tested for HIV at each
site, by quarter Percentage of new enrollees tested for HIV at each
site, by quarter
6
% o f new enrollees tested for

60%
% of new enrollees tested

5
50%
4
40%
for HIV

3
HIV

30% Site 1
2 20% Site 2
Site 1
1 Site 3
Site 2 10%
0 Site 3 0%
Q1 Quarter
Jan–March 1 Q2Quarter
Apr–June 2 Q3 July–Sept3
Quarter Q4 Oct–Dec4
Quarter Quarter 1 Quarter 2 Quarter 3 Quarter 4
Months

39 40

Stacked bar chart


 Represent components of whole & compare wholes  Displays trends over time
 Number of months patients have been enrolled in
HIV care Number of Clinicians Working in Each Clinic During
Years 1–4* (* includes doctors & nurses)
6
Females 4 10
5
Number of clinicians

0-14 years 4
15+ years Clinic 1
3
Males 3 6 Clinic 2
2 Clinic 3
1
0 5 10 15
0
41 Number of months patients have been enrolled in HIV care 42
Year 1 Year 2 Year 3 Year 4

7
6/28/2017

Number of Clinicians Working in Each Clinic During


Years 1–4* Pie chart
(* includes doctors & nurses)
Contribution to the total = 100%
6

5 Percentage of All Patients Enrolled by Quarter


Number of clinicians

8%
4
Clinic 1 10%
3
Clinic 2
1st Qtr
2 Clinic 3
2nd Qtr

1 3rd Qtr
23% 59% 4th Qtr
0
Y1 1995
Year 1 Y2Year
19962 Y3Year
19973 Y4 1998
Year 4

43 44
(Source: NHRC, 2015)

Interpretation Interpretation: Relevance of finding


 Adding meaning to information by making  Adding meaning to information by making
connections and comparisons and exploring connections and comparisons and exploring
causes and consequences causes and consequences

Conduct
Relevance of Reasons for Consider
further
finding finding other data Conduct
research Relevance of Reasons for Consider
further
finding finding other data
research

45 46

Interpretation: Relevance of finding Interpretation: Possible causes?


 Does the indicator meet the target? • Supplement with expert opinion
 How far from the target is it? • Others with knowledge of the program or target
 How does it compare (to other time periods, population
other facilities)?
 Are there any extreme highs and lows in the
data?

Conduct
Relevance of Reasons for Consider
further
finding finding other data
research

47 48

8
6/28/2017

Interpretation: conduct further


Interpretation: consider other data research
 Use routine data to clarify questions  Data gap conduct further research
 Methodology depends on questions being asked
 Use other data sources
and resources available
 Situation analyses
 Performance improvement data

Conduct
Relevance of Reasons for Consider
further
finding finding other data
Conduct research
Relevance of Reasons for Consider
further
finding finding other data
research

49 50

Key Messages
 Use the right graph for the right data
 Tables – can display a large amount of data
 Graphs/charts – visual, easier to detect patterns
 Label the components of your graphic
 Interpreting data adds meaning by making
connections and comparisons to program
 Service data are good at tracking progress &
identifying concerns – do not show causality

51

S-ar putea să vă placă și