What Is Criterion Sampling?
Criterion sampling is a purposive strategy where the researcher selects all cases that meet a predetermined standard or criterion. Unlike other purposive methods that select particular types of cases (typical, extreme, intense), criterion sampling uses a clear threshold: everyone who qualifies gets included or is eligible for inclusion. The criterion can be a behavior (all customers who returned a product), an outcome (all patients who were readmitted within 30 days), a status (all employees who completed the certification), or any other measurable standard that defines a group of interest. It's the most straightforward purposive approach, define the bar, then study everyone who clears it. The method is heavily used in quality assurance, program evaluation, and clinical research where specific outcomes or conditions trigger deeper investigation.
Why Criterion Sampling Matters
When you need to study everyone who meets a specific condition, not a sample of them, but the full set, criterion sampling is the method. It ensures complete coverage of the target group, which matters in quality assurance (you want to review every adverse event, not a sample of them), regulatory contexts (every noncompliant case requires documentation), and situations where the qualifying population is small enough to study exhaustively.
How Criterion Sampling Works
The method is conceptually simple, but the quality of the criterion and the identification process determine whether it works in practice.
Defining the Criterion
The criterion must be specific, measurable, and defensible. "Customers who had a bad experience" is too vague, different people would apply it differently. "Customers who gave a CSAT score of 1 or 2 out of 5" is a criterion. "Patients with elevated readmission risk" is vague. "Patients readmitted within 30 days of discharge for the same diagnosis" is a criterion.
Good criteria share these properties: they can be applied consistently across all potential cases, they're based on documented data rather than subjective judgment, and they define a group that's analytically meaningful for your research question.
Identifying Qualifying Cases
Once the criterion is defined, systematically identify every case that meets it. This requires access to the data system that contains the qualifying variable, your CRM, medical records, program database, or survey responses. The identification process needs to be exhaustive: missing qualifying cases introduces bias because the missed cases may differ systematically from the ones you found.
Automated identification (database queries, flag-based triggers) is more reliable than manual review for large populations. For smaller populations or complex criteria that require judgment (e.g., qualitative review of incident reports to determine severity), use multiple reviewers and check inter-rater agreement.
Studying the Qualifying Cases
Depending on population size, you either study every qualifying case or draw a random sample from the qualifying group. If 15 patients were readmitted, study all 15. If 1,500 customers gave low CSAT scores, you might randomly sample 150 from that group for in-depth follow-up.
When the qualifying population is small and the stakes are high (patient safety, regulatory compliance, major account churn), exhaustive coverage is standard. When the population is large and the research is exploratory, random sampling within the criterion group is more practical.
Data Collection Approach
Criterion sampling often triggers intensive data collection because the qualifying cases represent important events that warrant deep understanding. Chart reviews, root cause analyses, extended interviews, and multi-source data triangulation are common. The question isn't just "who qualifies?" but "why did they qualify, and what can we learn from their experience?"
Iterative Criterion Refinement
Sometimes the initial criterion is too broad (too many cases to study meaningfully) or too narrow (too few cases for useful analysis). Refine the criterion after an initial scan. If 3,000 customers meet your CSAT criterion, add a second filter (e.g., high-value accounts only). If only 3 cases meet your criterion, consider broadening it or extending the time window.
When to Use Criterion Sampling
- Quality assurance and improvement where every case meeting a failure or adverse-event criterion needs review, product defects, service failures, safety incidents
- Program evaluation studying specific outcome groups, all dropouts, all high-achievers, all participants who didn't meet benchmarks
- Clinical research investigating specific patient populations defined by diagnoses, outcomes, or treatment responses
- Customer experience research targeting specific satisfaction thresholds, all detractors (NPS 0-6), all repeat returners, all churned accounts
- Compliance and audit research where regulatory requirements define which cases must be examined
Common Mistakes to Avoid
- Using vague or subjective criteria that different researchers would apply inconsistently. If two analysts reviewing the same cases would disagree on who qualifies, the criterion needs tightening.
- Missing qualifying cases due to incomplete data systems. If your criterion depends on a database that has missing data, unreported events, or inconsistent coding, your "all cases" sample is actually a convenience sample of documented cases.
- Confusing criterion sampling with screening. Screening filters a sample before data collection based on eligibility criteria. Criterion sampling defines the study population as all cases meeting a specific standard, the criterion is the research focus, not just an entry requirement.
How Quali-Fi Supports Criterion Sampling
Quali-Fi's response filtering and conditional logic tools let you define precise criteria within your survey data and automatically flag qualifying respondents for follow-up research, ensuring no qualifying case slips through the cracks. The platform's automated triggers can route criterion-qualifying respondents into separate analysis groups or follow-up survey flows in real time.
Frequently Asked Questions
How is criterion sampling different from quota sampling?
Quota sampling sets target numbers for predefined demographic groups and recruits until those targets are met. Criterion sampling selects all cases (or a random sample of cases) that meet a specific standard. Quota sampling is about filling cells; criterion sampling is about studying a defined population.
Can I use multiple criteria simultaneously?
Yes. Multi-criteria sampling (e.g., patients who were readmitted within 30 days AND had a complication AND received a specific treatment) narrows the qualifying population to a more specific group. Each additional criterion should be justified by the research question.
What if very few cases meet my criterion?
Small criterion groups are common in quality assurance and safety research. If the qualifying population is tiny (under 10), study all cases intensively. Consider expanding the time window, broadening the criterion slightly, or supplementing with near-miss cases that almost met the threshold.
Related Topics
Study the cases that matter most. Start a free trial with Quali-Fi and use automated response filtering, conditional triggers, and real-time flagging to identify and follow up with every qualifying case.