Survey Construct Validity and Question Design
Construct Validity of Surveys
- Construct validity ensures surveys measure intended concepts and respondents answer truthfully.
- Question type (open-ended vs. forced choice) doesn't determine construct validity.
- Critical factor: Question wording.
Framing
- Framing: How question wording subtly influences interpretation, similar to a picture frame.
- Framing effects: Wording affects answers due to human psychology.
Leading Questions
- Leading questions imply the desired answer, potentially skewing responses.
- Example: A question that prefaces with negative statements about a topic before asking for approval.
- Another Example: Citing experts to sway opinion.
Framing Effects
- Framing: How choices are presented affects decisions, even with equal outcomes.
- Example: "80% fat-free" vs. "20% fat."
- Example: "Avoid a penalty" vs. "get a discount" (avoiding penalty more effective).
- Superficial aspects of questions can sway people.
Complicated Questions
- Avoid questions that complicate the survey.
- Complicated questions make answering harder.
Double-Barreled Questions
- Double-barreled questions combine two questions into one, limiting response options.
- Example: A question asking about agreement with two distinct aspects of the second amendment.
Negatively Worded Questions
- Negatively worded questions are harder to process than positive ones.
- Example: Confusing wording regarding belief about the Holocaust.
- Revised question: Only 1% indicated the Holocaust didn't occur.
- Example: "Abortion should never be restricted" involves a double negative.
- Alternative: "I favor strict restrictions on abortion."
- If unsure, ask both versions and compare responses.
- Multiple negative questions can exhaust survey takers.
Question Order
- Question order affects responses; earlier questions influence later ones.
- Example: Jealousy survey with scenarios in ascending or descending order of intensity.
- Same question can elicit different responses based on its position in the survey.
- Test different question orders to assess impact.
Shortcuts
- Respondents may take shortcuts due to survey fatigue.
Response Sets
- Response set: Answering multiple questions with the same response.
- Acquiescence: Repeatedly agreeing due to it being the nicer answer.
Fence Sitting
- Fence sitting: Choosing the neutral option to avoid taking a stand.
- Example: A scale about a higher power with a neutral midpoint.
- Solution: Remove the neutral option to force a choice.
- Disadvantage: Eliminates the ability to express a truly neutral stance.