Abstract
A sequential design utilizing identical samples was used to classify mixed methods studies via a two-dimensional model, wherein sampling designs were grouped according to the time orientation of each study's components and the relationship of the qualitative and quantitative samples. A quantitative analysis of 121 studies representing nine fields in the social or health sciences revealed that more studies utilized a sampling design that was concurrent (66.1%) than sequential (33.9%). Also, identical sampling designs were the most prevalent, followed by nested sampling, multilevel sampling, and parallel sampling, respectively. Qualitative analysis suggested that across a number of studies the researchers made statistical generalizations that were not sufficiently warranted—culminating in interpretive inconsistency and contributing to crises of representation, legitimation, integration, and politics.
| Original language | English |
|---|---|
| Pages (from-to) | 267-294 |
| Number of pages | 28 |
| Journal | Journal of Mixed Methods Research |
| Volume | 1 |
| Issue number | 3 |
| DOIs | |
| Publication status | Published - Jul 2007 |
| Externally published | Yes |
Keywords
- generalization
- interpretive consistency
- representation
- sample size
- sampling
ASJC Scopus subject areas
- Education
- Social Sciences (miscellaneous)
- Statistics, Probability and Uncertainty