RESEARCH METHODOLOGY

Secondary Data

Secondary (or historical) data is data which has been acquired through a research or any other effort in the past and for another purpose Depending on the secondary data needed, there may be numerous sources for acquiring this secondary

data

Advantages of Secondary Data

Pool of already existing data Cost-effective vis--vis primary data Quick availability No need for designing and undertaking a research study Collecting primary data may not be feasible (privacy issues) Many different sources are available

Disadvantages of Secondary Data

Relevance for researcher Does data apply to target population? Time-frame consistent with researchers requirement? (outdatedness) Variation in terms of definition of terms Different units of measurement (data conversion) Verification of datas accuracy (inherent biases due to vested interests) Reputation of the data-collecting organization and research design used may not be good

Evaluating Secondary Data

Do the data help answer questions set out in the problem definition? Yes Applicability to Current Project Do the data apply to the time period of interest? Yes No! Can data be reworked? No!

Do the data apply to the population of interest?

Yes Do other terms and variable classifications presented apply to the project? Yes Are the units of measurement compatible?

No! If yes, continue No!

No!

(Next Slide!)

Evaluating Secondary Data

Are the units of measurement compatible? Yes If possible, go to the original data source Yes Is the cost of data acquisition worth it? Yes Is there a possibility of bias?
Is using data worth the risk? No

No!

Yes! No! Inaccurate or unsure)

No

Can accuracy of data collection be verified?

Yes (accurate) Use Data

Yes No

Fact Finding Fact finding is about collecting secondary (or descriptive) data with a view to obtaining insights which can assist in decision-making Some examples: (1) Macroeconomic Data (2) Global Oil price trends (3) FDI in China over the period 2000-2005 (4) Indo-Pakistan trade statistics

(5) Salary structure across ten selected industries

Model Building

Model building is about specifying relationships between two or more variables Models can range from very simple ones, involving only two interrelated variables to complex ones involving several interrelated variables (multivariate statistical forecasting models)

Classification of Secondary Data: Internal Sources

Internal Sources (data already existing within an organization, for example, data collected and stored by various departments such as accounting and finance, marketing) Sometimes one unit in an organization is unaware that other units may be maintaining data which may be relevant for it and undertakes unnecessary and costly data collection

Classification of Secondary Data: External Sources

External Sources (data generated or provided by individuals and entities other than the researcher or his/her organization) There are many types of external sources of secondary data:
Books and Professional Journals Federal Government sources Regional / Provincial Publications Electronic and Print Media Sources Commercial Sources Internet, CDs

Secondary Data in the International Context

As economies and business becomes increasingly global in outlook, so does the need for secondary data from other regions and countries Potential Problems:
Unavailability of data in some countries Accuracy of data may be questionable Definitions and measurements of data may vary Comparability of data may be difficult

Surveys
A Survey is a research technique in which data is collected from a sample of people using an interview or questionnaire Surveys are a crucial tool of business research methods Surveys are undertaken using verbal or written means to obtain primary data for the research project

Surveys target individuals and/or organizations (respondents)

Surveys are often quantitative, occasionally also qualitative in their orientation Surveys are usually done for descriptive purposes and for ascertaining the characteristics of a group, to measure attitudes and determine behavioural patterns, and sometimes to explore ideas or provide causal explanations

Advantages of Undertaking Surveys

Surveys have a number of advantages in terms of collecting, analyzing and assessing information from the sampled population: Quickness Inexpensiveness Flexibility Efficiency Accuracy Helpful in the decision-making process

The advantages are only evident when surveys are properly conducted!

Potential Errors in Undertaking Surveys

Total Error
Acquiescence Bias Extremity Bias Interviewer Bias Auspices Bias Social Desirability bias

Random Sampling Error

Systematic Error (Bias)

Deliberate Falsification Unconscious Misrepresentation

Non-Response Error Respondent Error Response Bias

Random Sampling Error and Systematic Error

Random sampling errors arise when the group selected out of a population for the purpose of undertaking a survey is not totally representative of the population, i.e. a variation exists. Technically, a random sampling error is the difference between the results of a sample and the result of a census conducted using identical procedures. As the sample size is increased, the variation will decrease Systematic error results from some imperfect aspect of the research design which causes response error, or from a mistake in the execution of the research

Respondent Error
Surveys depend on individuals responding to the questions asked of them in written or verbal form. Respondents must thereby fulfill two preconditions:
Be cooperative Be truthful

If these two preconditions are not fulfilled, the survey is unlikely to achieve its goal. Two major problems resulting from the non-fulfillment of these two preconditions by respondents are:
Non-response Error Response Bias

Non-Response Error and Causes of Non-Responding

Non-Response error is defined as the statistical difference between the results of a survey that includes those individuals who responded, and a perfect survey in which all individuals would have responded
A consequent problem stemming from non-response error is that the survey-based research results may be unutilizable for decision-makers if those individuals who did respond to the survey are not representative of those who did not respond

There are many reasons for not responding to surveys, for example, lack of time and preoccupation with routine work, absence from home, lack of interest in the survey, cultural factors (e.g. Middle East)

Response Bias
A response bias occurs when survey respondents tend to answer the questions posed to them in a certain direction, thereby consciously or unconsciously, or intentionally and inadvertently, misrepresenting the truth Response bias has been found to depend on factors such as the income or social class of respondents and their ethnic background. Example: Mayoral and gubernatorial elections in the USA (white respondents and their supposed choice of candidates)

Deliberate Falsification by Survey Respondents

Deliberate falsification occurs when survey respondents deliberately give false answers. There are many reasons for this, for example, when respondents want to appear more intelligent, to avoid embarrassment, and to conceal personal information

Examples: Survey about shopping habits (respondent has forgotten expenses paid and does not want to admit this; employees are asked to give their opinion about their employers and give false answers because they fear adverse consequences; respondents wants to please the interviewer and give answers they think will bring this about; survey respondents want to appear average so that they dont stand out too much)

Unconscious misrepresentation by Survey Respondents

Unconscious misrepresentation occurs because of the specific situation or stimulus a survey respondent finds himself or herself in, or because of the nature of the questions asked, even though the respondent is trying to be truthful and cooperative Examples: Respondent has forgotten dates and details and gives a best guess estimate, which may not be accurate; respondent doesnt have enough time to think about the answer to a question; feelings often cannot be expressed accurately in words)

International surveys are particularly susceptible to unconscious misrepresentation by survey respondents due to cultural and communicational differences

Types of Response Bias (1)

Acquiescence Bias Tendency of survey respondents to agree with the questions asked of them, especially in regard to surveys relating to new or unfamiliar programs, products or ideas. Acquiescence bias can, on the other extreme, result in a respondent disagreeing with all questions asked of them

Extremity Bias Tendency of survey respondents to use extremes when responding to questions, unlike other respondents who may adopt a neutral stance Interviewer Bias Tendency of survey respondents to give untrue answers because of the presence or influence of interviewers on them and their interest in appearing intelligent, more affluent or just willing to please

Types of Response Bias (2)

Auspices Bias Tendency of survey respondents to give answers to questions based on their perception of the organization which is undertaking the survey. For example, employees of CIIT may give quite different responses to an in-house CIIT survey than they would, were they asked by a neutral or unrelated organization Social Desirability Bias Tendency of survey respondents to give answers that put them in a favourable light with the interviewer. For example, people may claim to be more socially active than they really are because being socially active is considered a positive activity, or inflate their education and income levels to save face and gain prestige

Potential Errors in Undertaking Surveys

Total Error

Random Sampling Errror

Systematic Error (Bias)

Data Processing Error Sample Selection Error Interviewer Error Interviewer Cheating

Administrative Error

Administrative Errors (1)

Administrative Errors are the consequence of the improper administration or execution of the research task Administrative errors can reduce the value of the research and hence, its usefulness as a tool for decision makers There are many causes of administrative errors, including, for example, carelessness, confusion, neglect or omission

Four major types of administrative errors are data-processing errors, sample selection errors, interviewer errors and interviewer cheating

Administrative Errors (2)

Data-processing Errors occurs when data are inaccurately entered by people into the computer. Such errors can be reduced or minimized by creating and applying careful processes and procedures for verifying each stage of data computerization Sample Selection Errors occur due to an improper or non-representative - sampling of individuals for the survey Interviewer Errors occurs when interviewers misrecord or fail to record responses due to inability, lack of experience, personal biases and preferences or some other reason

Interviewer Cheating occurs when an interviewer falsifies questionnaires or fills in the answers himself or herself to selected questions or skips questions to avoid asking sensitive questions

Techniques for Estimating Systematic Error

Estimating systematic error can be quite a difficult undertaking. In practice, many business researchers use the rules-of-thumb method, meaning that they use past experience to try to quantify how much survey results differ from actual results Example: Approximately 40% of survey respondents who claim they will definitely purchase Products X,Y and Z within the next 7 days will actually go out to the market and carry out this purchase, while only 10% who stated that they may purchase Products X, Y and Z will actually do as they have said To reduce the likelihood and extent of systematic errors, much care has to be given to designing a proper questionnaire, adequate training of interviewers and selecting the appropriate samples

Types of Survey Research Methods (1)

Surveys can be classified according to the mode of communication with the respondent (personal interviews, telephone interviews, mail surveys, internet surveys etc.) Surveys can be classified according to the type of questions asked of respondents (structured questions, disguised questions). Structured questions impose a limit on the number of permissible responses, while disguised questions try to hide the purpose of the study from the respondents and get information that respondents may otherwise be reluctant to give. Surveys often incorporate both structured and disguised questions

Types of Survey Research Methods (2)

Surveys can be classified according to their time frame (cross-sectional studies, i.e., where data is collected from respondents at a single point in time, and longitudinal studies, i.e., where data is collected from a group of respondents over a time interval, with a view to examining the level of continuity or change over time If data is collected from the same sample of individuals, the longitudinal study is called a panel study. A method for documenting data in panel studies is to use diaries for tracking

Personal Interviews
A personal interview is a form of direct communication in which an interviewer asks respondents in a face-to-face conversational situation Personal interviews can take place in various locations, often at the respondents place of residence or in their workplaces (door-to-door interviews), in shopping malls and in supermarkets (mall intercept interviews) and in other high-traffic areas (this has the advantage of lower cost but, on the downside, it has a higher refusal rate due to the respondents time limitation and there may be sampling issues to consider) Personal interviews have a number of advantages and disadvantages for business researchers

Advantages of Personal Interviews (1)

Opportunity for Feedback Interviewer can provide direct feedback to the respondent, give clarifications and help alleviate any misconceptions or apprehensions over confidentiality that the respondent may have in answering the interviewers questions Probing Complex Answers Interviewers can probe if the respondents answer is too brief or unclear. This gives interviewers some flexibility in dealing with unstructured questions and is especially suited for handling complex questions

Length of Interview If the questionnaire is very lengthy, the personal interview is the best technique for getting respondents to cooperate, without overtaxing their patience

Advantages of Personal Interviews (2)

Complete Questionnaires Personal ensures ensure that the respondent will answer all questions asked, unlike in telephone interview where the respondent may hang up or in mail questionnaire where some questions may go unanswered Props & Visual Aids Interviewers have the opportunity of showing respondents items such as sample products, graphs ands sketches, which can aid in their answers High Participation Interviewing respondents personally can increase the likelihood of their participation, as many people prefer to communicate directly verbally and sharing information and insights with interviewers

Disadvantages of Personal Interviews (1)

Cost Personal interviews are usually more expensive than mail, telephone and internet surveys. Factors influencing the cost of the interview include the respondents geographic proximity, the length and complexity of the questionnaire, and the number of non-respondents Lack of Anonymity Respondents are not anonymous in a personal (faceto-face) interview and may be reluctant to disclose certain information to the interviewer. Hence, considerable must be expended by the interviewer when dealing with sensitive questions to avoid bias effects on the respondents part Necessity for Callbacks When a person selected for interview cannot be reached the first time, a callback has to be scheduled which result in extra cost and time spent

Disadvantages of Personal Interviews (2)

Variance Effects It has been shown that the demographic characteristics of the interviewer can influence the answers of the respondents. In one study, male interviewers had a much larger variance of answers than female interviewers in a sample of most female individuals Dishonesty Interviewers cheat to make their life easier and save time and effort Personal Style The interviewers individual questioning style, techniques, approach and demeanor may influence the respondents answers Global Considerations Cultural aspects may influence peoples willingness to participate in an interview (e.g. repressive Middle Eastern cultures discourage females from being questioned by male interviewers)

Telephone Interviews
In telephone interviews, respondents are contacted by telephone in order to collect data for surveys Telephone interviewing has been used for decades and, in some ways, has advantages over other methods of undertaking surveys With improvements in the IT-field, computers can be used to assist in telephone interviewing, and answers given by respondents can be entered by interviewers directly into the computer, saving effort, time and cost

Advantages of Telephone Interviews (1)

Speed Speed is a major advantage of telephone interviewing, enabling data to be collected on very short notice (example: a union decides whether to organize a strike by telephone-interviewing members over a one-day period) Cost Telephone interviews are comparatively cheaper to conduct than personal interviews. No travel time and travel cost is involved. No Face-to-Face Contact Because telephone interviews lack the direct element of interaction, respondents may be more willing to provide certain information that they would be reluctant to disclose in a personal (face-to-face) interview

Advantages of Telephone Interviews (2)

Cooperation People may be reluctant to allow interviewers into their homes, but they may be willing to cooperate by letting themselves be interviewed over the telephone Callbacks Telephone callbacks are easier to perform than personal interview callbacks

Disadvantages of Telephone Interviews (1)

No Face-to-Face Contact Interviewer may not be able to record the respondents data fast enough and the respondent, who cannot see this, may continue to add data. Also, due to the visual communication gap, there is a greater tendency for interviewers to record no-answers and incomplete answers than in a personal interview Cooperation Research shows that response rates in telephone interviews are declining with the passage of time and the availability of respondents has also declined for various reasons. Also, reaching executives in workplaces can be very difficult due to tight schedules and the work load

Disadvantages of Telephone Interviews (2)

Lack of Visual Mediums Visual aids cannot be used by interviewers in telephone interviews, hence, surveys which need visual aids to help respondents cannot be undertaken with this survey method Limited Duration Length of interview time in a telephone interview is limited. Too long interview times may result in exasperated respondents hanging up the telephone or refusing to answer questions

Representative Samples Using the telephone directory as the basis for sampling can be problematic in the sense that many persons are unlisted or do not have telephones, but whose opinions are nevertheless important
Global Considerations In many countries, people are reluctant to divulge information over the telephone

Self-Administered Questionnaires
Self-Administered Questionnaires

Printed Questionnaires

Electronic Questionnaires

Mail

E-Mail

In-Person Drop-Off
Inserts Fax

Internet Website
Interactive Kiosk