A sociologist wants to understand why employees at one company leave their jobs at higher rates than at a comparable company. Which method is best suited to generating initial hypotheses about this?
AA large-scale national survey of worker satisfaction
BIn-depth interviews with current and former employees at both companies
CSecondary analysis of industry-wide turnover statistics
DA randomized controlled experiment assigning workers to different managers
When the goal is understanding meaning and process — why something is happening in a particular context — qualitative methods like in-depth interviews are best suited. They allow the researcher to discover unexpected factors and follow leads. A national survey would be too broad; secondary statistics wouldn't capture the specific companies; an RCT would be impractical and wouldn't generate explanatory hypotheses.
Question 2 True / False
A survey of 100,000 randomly selected people will generally produce more valid results than a survey of 1,000 people.
TTrue
FFalse
Answer: False
Sample size is only one factor in validity. The famous 1936 Literary Digest poll predicted a landslide loss for Roosevelt despite 2.4 million responses — because the sampling frame (car owners and phone subscribers) systematically excluded poorer voters. Question wording, response bias, and the fit between the sample and the target population matter at least as much as sheer size. A well-designed survey of 1,000 randomly selected people can be far more valid than a massive but poorly designed one.
Question 3 Short Answer
A researcher conducts an ethnographic study of a single high school for two years. A critic argues the findings cannot be generalized. How should the ethnographer respond?
Think about your answer, then reveal below.
Model answer: Ethnographic research aims for transferability, not statistical generalizability. The goal is thick description — detailed contextual findings that readers can assess for applicability to their own situations — not claims that the results hold across all high schools in a population.
Qualitative and quantitative research operate under different logics of rigor. Generalizability in the statistical sense requires representative sampling and is appropriate for quantitative research. Qualitative research instead seeks credibility (are the findings trustworthy?), transferability (can readers judge applicability to other contexts?), and reflexivity (did the researcher account for their own position?). Neither standard is superior — they answer different kinds of questions.