Data Collection & Sampling Flashcards

When designing a questionnaire, a question like 'Don't you agree that our product quality is excellent?' is an example of:

Answer: Leading question

A leading question steers respondents toward a particular answer through its phrasing, compromising the objectivity of the data.

Which sampling approach is most appropriate when the researcher needs to ensure representation of rare subgroups that would be underrepresented in a simple random sample?

Answer: Stratified oversampling

Stratified oversampling deliberately selects more respondents from rare subgroups than their proportion warrants, ensuring sufficient data for subgroup analysis.

A researcher collects transaction records from a retailer's database without any interaction with customers. This is an example of:

Answer: Secondary data

Secondary data is information that was originally collected for another purpose and is being reused for the current research.

What does 'coverage error' refer to in survey methodology?

Answer: The discrepancy between the target population and the sampling frame

Coverage error occurs when the sampling frame does not perfectly represent the target population, leaving some members with no chance of selection.

In focus group research, what is the primary role of a moderator guide?

Answer: To outline topics and probes while allowing flexible exploration of emerging themes

A moderator guide structures the discussion with key topics and probing questions while allowing the moderator flexibility to explore relevant themes as they arise.

Which of the following best describes 'random digit dialing' (RDD) as a sampling method?

Answer: Generating phone numbers randomly to reach households without listed numbers

RDD generates phone numbers randomly, allowing researchers to reach both listed and unlisted households, improving coverage of the general population.

A researcher asks respondents to rate their agreement with 10 statements, each on a 5-point scale. After analysis, the researcher finds responses are highly correlated and all measure a single underlying construct. This demonstrates:

Answer: Scale dimensionality suggesting the items form a single construct

When multiple items correlate strongly and appear to measure one underlying construct, factor analysis or similar techniques reveal the scale's dimensionality.