Week 4: Questionnaire Design, Slides of Design

Download Week 4: Questionnaire Design and more Slides Design in PDF only on Docsity!

Week 4: Questionnaire Design

Andrew Stokes

February 12, 2018

Review: Week 1

I (^) Research best practices I (^) Value of mixed methods approaches I (^) Importance of replication and reproducibility I (^) Practical implication of these principles for implementing your study

Overview of Today

I (^) Mode of data collection I (^) Types of questions I (^) Asking effective questions I (^) Survey Design I (^) KoBoToolbox Demo

Mode of data collection

Survey data are collected in a variety of ways

I (^) Face-to-face interviews (e.g. NHANES) I (^) Telephone (e.g. BRFSS) I (^) Mail I (^) Internet (e.g. GBD Disability Weights Survey)

Strengths of interview administered questionnaires

I (^) Allow participation of illiterate people I (^) Allow clarification of ambiguity I (^) May yield higher quality data (e.g. research has shown that weight data collected in the NHANES has less bias than that in the BRFSS)

Mode of data collection continued

The approach used in a given context depends on,

I (^) Type of population being sampled I (^) Resources that are available I (^) Nature of the data being collected.

The basic principles of questionnaire design we will discuss below are similar across different modes of collecting data.

Types of questions

I (^) Open vs. closed questions I (^) Recall questions I (^) Likert scales

Open vs. closed questions

The two basic types of questions are open-ended and closed questions

Closed questions

I (^) Respondents are given a list of possible answers from which they must choose I (^) Appropriate for well-defined studies in which the range of responses is known.

Closed questions: advantages

I (^) Easy and quick to fill in I (^) Reduce challenges for those who are less literate I (^) Easy to code, record and analyze results quantitatively I (^) Easy to report results

Recall questions

I (^) Survey questions can ask about the past as well as the present I (^) Recall questions are an inexpensive way to collect historical data on respondents, but potential for recall error should be carefully evaluated I (^) Recall bias tends to increase as the time horizon grows and depends on the salience of the event to the individual.

Recall questions: Examples

I (^) DHS birth histories. In birth history modules, questions relating to children are asked for each child, including time of birth and survival status. I (^) NHANES weight histories. These ask respondents to report their weight and height at various past ages, such as at age 25. The survey also asks respondents about weight ten years ago and at maximum. I (^) Smoking histories are used in many surveys to determine when an individual began smoking and when they quit (if they quit). Information is also frequently gathered on intensity of smoking in the past. I (^) Others you can think of?

Likert scales: Example

Self-reported health is a very common variable collected in health surveys. This information is usually collected using a Likert Scale. For example, see this question from the NHANES.

Likert scales: Example

“Statistics is an interesting topic” (strongly disagree/disagree/cannot decide/agree/strongly agree)