Some researchers/reviewers don't like primary data (questionnaire). What are the reasons and what can be the main issues with acquiring or analyzing the primary data. Your answers and guidance will be highly appreciated.
I myself don’t like many of the questionnaires I’ve been asked to fill. The people who design them often don’t seem to realize that their questions are ambiguous, vague, or make unwarranted presuppositions. Moreover, with multiple-choice questions they may force an incorrect answer because they don't have “none of these” as an option. Such shortcomings compromise the value or interpretation of the data. To use a familiar caricature, what can one conclude from the answer to "When did you stop beating your wife?" when the answerer either has no wife or else has a wife but has never beaten her, if the only available choices are between times?
Surveys or questionnaires are in different firms (e.g., paper, electronic, in person interviews). Each has potential issues. Karl Pfeier articulated some of these concerns.
Moreover, bias might be induced by the way a question is worded (or asked in an oral interview). Even subtle hints of an expectation of a preferred answer might change a respondent's answer.
You always have to be careful of sampling bias. Are the people taking the survey actually representative of the population? Are they self-selected (maybe the respondents have an interest)? Have you contacted people on a medium with an unrepresentative set of users? Are working people unavailable because you did the survey on a street corner in the middle of the day? Are they confident that their answers are anonymous (people often conceal unpopular opinions or preferences)?
Hi,
Pleasw refer tis excellent source on issues in collecting primary data. We can prevent several mistakes by designing the questionnaire carefully and pretesting it.
https://ori.hhs.gov/education/products/n_illinois_u/datamanagement/dctopic.html
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning