Designing the Questionnaire or Observation Form
Chapter 14
Stages in the Research Process
Importance of Research Hypotheses in Questionnaire Design
Hypotheses guide questionnaire design by determining what information will be sought and from whom
Hypotheses establish the relationships between variables that the researcher wants to investigate
This affects the type of questions to be asked and the response options to be coupled with the questions themselves
Key Steps in Questionnaire Development
Step 1: Specify What Information Will Be Sought
The first step should be relatively easy, assuming that the researchers have done a good job at earlier stages in the research process
Hypotheses, dummy tables, etc., make it clear what information is needed
Hypotheses are the researcher’s expectations of what relationships they will find between variables.
Collect only the information that is vital to the decision problem and research problems. Researchers should not look to obtain that will simply be interesting or nice to know.
This lengthens questionnaire, lowers response rate, increases subject fatigue
Step 2: Determine Method of Administration
The degree of structure and disguise influence this decision, as does the specific research situation
Personal Interview
Telephone Interview
Mail Questionnaire
Internet-based Questionnaire
Consider the specific research project and the advantages and disadvantages of each data collection method to determine what is best
May consider length of questionnaire (too long for telephone?), structure (high or low structure?), resources (available $), complexity of questions (may need explanation?), disguise?
Step 3: Determine Content of Individual Questions
Driven mostly by decisions in Steps 1-2
What information is necessary to collect?
What degree of structure is necessary to collect
it?
Does the research need to be disguised?
Some key issues for question formation
Is the question necessary? Include as few questions as possible, but make sure you completely address all informational needs
Are several questions needed instead of one? If question may be interpreted differently by respondents, researchers may need to ask multiple, more specific questions that will address all potential responses
Do respondents have the necessary information? They will answer any question asked, even if they have no clue as to the correct answer. May ask filler/qualifying questions to weed out potential inaccurate responses.
Will respondents give the information? May vary due to how much effort is needed to answer, sensitivity of the issue
Step 3: Telescoping and Recall Error
Respondents’ ability to answer accurately may be a function of:
Telescoping Error
Type of error resulting from people remembering an event as having occurred more recently than it did
Recall Error
Type of error caused by a respondent’s forgetting that an event happened at all
The impact of these two types of error depends on the length of the time period in question.
Longer periods = Smaller telescoping error; larger recall error
Shorter periods = Larger telescoping error; smaller recall error
Step 3: Handling Sensitive Questions
Don’t ask sensitive questions unless absolutely necessary! If you have to:
Guarantee anonymity
Place sensitive questions near end of questionnaire – develop rapport first, and if they decide to stop answering they already answered most questions.
Include a counter-biasing statement – something that will normalize a potentially embarrassing response