QUESTIONNAIRE DESIGN
Questionnaire is a tool to gather information about a respondent ( Self-administered versus interviewer-administered questionnaire Before writing questionnaire researcher needs to have clear understanding of the purpose of survey, the list of research variables involved, and a plan of analysis, all of which in turn are derived from the research problem.
Quantitative Survey Questions are worked out deductively from the conceptual framework and/or hypothesis THEORETICAL/ CONCEPTUAL FRAMEWORK RESEARCH QUESTION THEORY SEARCH HYPOTHESIS QUESTIONS INDICATORS VARIABLES
Qualitative Survey Questions are worked out based on a preliminary idea of the structure of the problem (what the problem involves, what relationships might occur etc). Questions provide stimuli for the respondent to express their thoughts, experiences, ideas that are relevant to the preliminary structure, quite possibly adding new insights into the problem Researcher continuously analyze the responses and display the findings, and improve on the questions. DATA COLLECTION DATA DISPLAY DATA REDUCTION CONCLUSIONS: DRAWING/VERIFYING
Clear variables should guide the researcher in writing questions posed to respondents to meet following standards: Write exactly what is to be said to or read by the respondent The question should mean the same thing to every respondent The question should be answerable The question should be one that all respondents would be willing to answer
Before designing questionnaire, researcher should conduct focused discussions with a small group of respondents to get an idea of how to pose questions on specific variables important to the researcher After designing questionnaire, researcher should test it by administering it to several respondents to see whether all questions can produce the correct information.
Questions should be able to extract correct information needed to assign a value to a specific variable: Example: income What is meant by income: salary? Take home pay? Money received each month from all sources? Income of head of household or income of all members? How do we classify income: how many classes? High-middle-low? Or numerical grouping? How important is class detail in analysis? How do we pose the question that people will anwer truthfully? Do we need to ask multiple questions or just one question on income? Do we need to mask the question?
Physical Design Questionnaire should be designed for simplicity and clarity: interviewers should have clear instructions on what to say to respondents; respondents should have clear instructions on how to answer questions Not too small type, not too small spacings, not too much clutter Use visual cues to direct respondents/ interviewers to the next questions Not too thin pages, easier to handle at least 80 gram paper, preferably heavier paper Not too many pages, no standard of appropriateness 30 minutes to 1 hour’s time needed for completion
Substantive Design Introduction: Must clearly state the purpose of the questionnaire, its importance to some theoretical or policy issues Very politely ask for the respondents time and effort to fill in the questionnaier or to be interviewed, stating that the respondents’ participation would greatly enhance the achievement of the said purpose Must give assurance of confidentiality of respondents’ identities and information given Give thorough instructions about how to answer questions, repeat instructions for every question if necessary
Substantive Design Part I: demographic information Name (optional) Address (if necessary, as locational variable) Age, sex, marital status Part II: social economic status (if research variable) Level of education Occupation Income Etc Part III: research variables (need to be grouped according to categories of questions)
Each question must be very clear, using precise words; eliminate chances of any questions about the question Bad question: a. Your age? b. What is your income? c. How do you spend your money each month? Good question a. How old would you be this year? b. How much do you earn each month from your primary and supplemental jobs? c. What is your estimated spending for food this month? What is your estimated spending for fuel this month? What is your estimated spending for clothing this month? What is your estimated spending for medicine and health this month?
What are good questions? Questions that are well articulated, clear, and have very specific single meaning; leave no chance of misinterpretation Questions that mean the same to everyone (every respondent should experience the same stimulus to make a response that are comparable among other respondents) If questions have multiple-choice answers, the choices must be clearly separated and structured, no double or ambiguous choices
What are good questions? Must be answerable by the respondents: Respondents must understand the questions Respondents must be willing to answer Respondents must know the answer Respondents must be willing to provide the answer
Structured Choice Answers Grouping: to give respondents easier way to estimate answers Nominal scale answers (sex, occupation etc) Interval scale answers must be categorized to facilitate analysis (income, age, etc) Ordinal scale answers: useful to get perceptual information about respondents choice, attitudes, or feelings (Agree-disagree items) From 2 categories to 5 categories Positive feelings --- negative feeling Good --- not good Good --- fair --- poor Very good --- good --- fair --- poor Excellent --- very good --- good --- fair --- poor
Steering your respondent to finish the questionnaire Use visual aids to lead the respondents to the next question, especially if they need to jump to a certain point after answering a question. Boxes and arrows are commonly used Instructions must be very clear and explicit