Why stratification does not prove causation
Comparing people within the same measured group can address one composition difference. Other important differences can still remain.
Readiness may influence both support participation and final assessment. Grouping by baseline readiness checks one plausible source of confounding. Motivation, time, unmeasured preparation and selection may still differ within groups. Higher conditional rates remain observational evidence unless a suitable design and assumptions support a causal interpretation.
Timing matters for adjustment too. Baseline readiness precedes support; tutoring attendance measured after support might itself be changed by the intervention. Automatically grouping by every available variable can distort the causal question. More detailed subdivision is not automatically more valid, and tiny cells can become unstable.
The descriptive pooled result can be relevant to the actual observed mix, while a conditional result answers a different comparison. Choose the view based on the question and data-generating process, rather than treating either the pooled or stratified result as universally correct.
Source: W5 PDF pp. 3-51, 53. The lecture uses both a constructed 200-student support example and a synthetic 4,665-registration course dataset. They are different tables with different units.
Changing groupby changes the comparison, not its causal strength. Course mix and preparation can still differ between programmes.
The lecture's claim ladder becomes a written answer: report the comparison, explain its limits and identify a useful next check.
The lecture's fences and whiskers become record-level checks. Programme summaries and two equal-mean/SD classes show why the whole shape matters.
Practice problems
Supplementary study questions on the lecture concepts — not verbatim exam questions. Each answer counts toward spaced review.
1. Conceptual understanding
Definitions and distinctions.
Does stratifying by readiness prove support causes success?
Which candidate adjustment requires particular caution?
2. Application
Apply the concept to a case or calculation.
Which variable is clearly measured before a support intervention?
After grouping by readiness, which factor might still differ between support and no-support students?
3. Interpretation and limitations
Evaluate assumptions and justify conclusions.
Why not automatically group by every available variable?
Which result describes the actual observed composition?
Explain it in writing
Self-checked, not marked. Write a few sentences, then compare with the example answer. Your writing is saved as you type.
Show an example answer
Baseline readiness precedes support and may help describe selection. Later attendance may be part of how support works, so adjusting for it can change or distort the causal question.
Check: did you answer the question, use evidence or working, and avoid claiming more than the evidence supports?
Show an example answer
Yes. The pooled result describes the actual observed mix; within-group results compare more specific groups. Choose the result that matches the question and explain its limits.
Check: did you answer the question, use evidence or working, and avoid claiming more than the evidence supports?