Source
doi:10.1177/1744987120927206 — original publication (opens in a new tab; the file is not redistributed)
Context: thesis · No NI feeds
Summary
Campbell et al. (2020) is a peer-reviewed methods paper from a University of Tasmania nursing-research group that sets out the nature and intent of purposive sampling and then illustrates it with three worked case studies drawn from the group’s own projects. Its central argument is that purposive sampling improves methodological rigour by better matching the sample to the study’s aims and objectives, and that this rigour should be judged against the four aspects of trustworthiness described by Lincoln and Guba: credibility, transferability, dependability and confirmability. The paper first distinguishes the main multi-case purposive strategies (stratified, cell, quota and theoretical sampling), then shows how each of three studies selected a strategy fitted to its methodology and question: a co-led stroke-service redesign using stakeholder-driven “Matrix” sampling; a child-and-family-health nursing study using inclusion/exclusion criteria under Interpretive Description; and a narrative-inquiry study of new mothers’ mental wellbeing that deliberately chose a simple criterion-based purposive design over a more complex one. For the thesis it is a methods-support citation for the sampling-strategy section: it supplies a compact typology of purposive strategies, worked “when to use which” examples, and the trustworthiness framework used to justify a purposive sample, complementing Nyimbili & Nyimbili (2024, LIT_105), which catalogues Patton’s finer-grained 16-type typology.
Key thesis insights
-
Purposive sampling exists to match the sample to the study’s aims, and its rigour is judged against four trustworthiness aspects. The paper defines purposive sampling as moving “away from any random form of sampling” to deliberately include specific kinds of cases that, given the study’s aims, hold different and important views on the questions at issue, selecting “respondents that are most likely to yield appropriate and useful information” and using limited research resources effectively. Making this selection logic explicit is what delivers methodological rigour, “as judged by the four aspects of trustworthiness” — credibility, transferability, dependability and confirmability (the Lincoln and Guba criteria). This is the justification frame the thesis methodology can adopt: state the sampling rationale from an ontological, epistemological and axiological perspective aligned to the study aims, then show it addresses each trustworthiness aspect. LIT_112
-
Four multi-case purposive strategies, each with a “when to use”. Stratified — select specific groups that must be in the sample and allocate a set number to each stratum (e.g. by age, family size), with a clear aims-linked reason each group differs. Cell — like stratified but the categories are discrete and may overlap “like a Venn diagram” (e.g. obese children, diabetic children, and the obese-and-diabetic overlap). Quota — more flexible: specify categories and a minimum number for each, monitoring fulfilment as the study proceeds; argued to be easier in recruitment terms than stratified or cell sampling. Theoretical — part of ongoing data collection and analysis (originating in Grounded Theory), where cases from new groups are added, or the sampling criteria are reshaped as the analysis proceeds, replacing the a-priori strategy. These labels give the thesis a vocabulary for naming and defending a multi-case selection design. LIT_112
-
Worked case examples show purposive sampling fitted to the methodology. (1) Stroke-service redesign used “Matrix sampling” (Reed et al. 1996; stakeholder sampling per Ovretveit 1998): stakeholders nominate participant categories they judge representative (e.g. 50% of stroke patients over 65, 50% aged 65 or under), with explicit exclusion criteria (transient ischaemic attacks, nursing-home residents, deceased patients) — a cost-effective, time-efficient method creating “maximum variability based on stakeholder knowledge.” (2) Child-and-family-health nursing study under Interpretive Description used inclusion/exclusion criteria (currently employed CFH nurses with ≥2 years recent experience) to recruit those “most likely to have in-depth knowledge and experience of the phenomenon,” explicitly mapping each design choice to credibility, transferability, dependability and confirmability. (3) New-mothers mental-wellbeing narrative-inquiry study chose specific inclusion/exclusion criteria (first-time mothers, singleton pregnancy, age over 18) for a longitudinal design of ~10 participants, judging a simple criterion-based purposive design more appropriate than a more complex one. LIT_112
-
All three “purposive” designs still carry a convenience element, and complexity is not automatically better. The authors note that because all consent is voluntary, “all of them, although purposive, have a convenience element” — the researcher is “at the mercy of the pool of potential participants.” Their conclusion is that the only firm requirement is that the sample be transparently justified so the reader can judge relevance, and that considering more complex purposive options (even when ultimately discarded for a simpler one) is itself part of building a high-quality design rather than “settling for standard approaches.” A methodology chapter can cite this both to acknowledge the convenience limitation candidly and to defend choosing a straightforward, well-reasoned purposive design over a needlessly elaborate one. LIT_112
-
Sample-size reasoning for small qualitative samples turns on data quality, not participant count. Drawing on Morse (2000), the paper argues that scope, topic, data quality, study design and “shadowed data” govern sample size, and that “there is an inverse relationship between the amount of useable data obtained from each participant and the number of participants” — so a longitudinal design with several interviews per person (e.g. four interviews with 10 participants) amasses ample data at a modest headcount. This supports a thesis using a small purposive expert-interview sample to justify its size by depth and repeated engagement rather than by number of participants. LIT_112
Notes
- Provenance. Authoritative first-party open-access PDF (Journal of Research in Nursing, SAGE, published under a Creative Commons reuse licence; available via Europe PMC OA).
- Identity confirmed against the raw. All eight authors (Steve Campbell, Melanie Greenwood, Sarah Prior, Toniele Shearer, Kerrie Walkem, Sarah Young, Danielle Bywaters, Kim Walker), all University of Tasmania, College of Health and Medicine; Journal of Research in Nursing 2020, 25(8):652-661; DOI 10.1177/1744987120927206; corresponding author steve.campbell@utas.edu.au — all match the citation recorded above. This is the Campbell et al. article, not the Denieffe commentary that appeared alongside it in the same issue.
- Complements Nyimbili (LIT_105). Nyimbili & Nyimbili (2024, LIT_105) catalogues Patton’s finer-grained 16-type purposive-sampling typology; Campbell et al. supplies the compact multi-case strategy set (stratified/cell/quota/theoretical), three worked case examples, and the four-aspect trustworthiness frame for justifying a purposive sample. The two are complementary and belong together.
- data_quality: high. Peer-reviewed methods paper in an established nursing-research journal, authored by an experienced multi-author group spanning PhD students to full professors; it restates and applies well-accepted methodological frameworks (Lincoln and Guba trustworthiness; Patton/Miles and Huberman typologies) with original worked examples rather than reporting new empirical measurements.
- Scope. A methods paper: no NZ-specific data and no NI calculation-cell exposure.
Connections
Links to
Sources (1): LIT_105
Referenced by
Concepts (1): Research Methodology