Ex-post-facto designs ("after the fact")

Much less able to determine causality than true experiments but these are necessary and important research designs.

Necessity for:1. ethical reasons or

2. an interest in organismic variables

2 main types:

Prospective and Retrospective designs

• find naturally occurring groups (thus, "after the fact") and follow them forward (prospective) or trace their histories (retrospective)

Problems:

1. subjects are not randomly assigned to treatments, as a result there will be inherent confounds in the populations studied (this is the most serious problem)
2. sampling problems (often a convenient sample):
3. dropouts in prospective studies
4. detection bias (equally likely to detect in both groups?)

Matching:

1) subject for subject (preferable but more difficult) or

2) distribution by distribution

• in both cases can selectively drop individuals and bias the sample further

Measuring: so will:

1) know if potential confounds (uncontrolled or extraneous variables) are confounded, and

2) to statistically control for these variables

Retrospective studies have additional problems in that they rely on memory so the partial solutions are more difficult to employ successfully

Have the advantage (over prosepective designs) in that they are more efficient (cheaper and faster) May be necessary with very rare grouping variables of interest (e.g., rare diseases)

(The additional problems of retrospective designs were well illustrated by McFarland's (1988) study of cyclical variability in moods)

DVs used in Ex-post-facto studies

• relative risk ratio (prospective studies) - illustrated by breast cancer data
• relative odds ratio (approximates the relative risk) -- retrospective studies

Problem with both in that absolute risks are hidden, both (absolute and relative risks) should be reported

Causality and ex-post-facto designs. Although no one (or few) quasi-analytic experiment will unambiguously show a causal relationship, with converging evidence from many such studies (5, 10 or 100?) can make causal statements (like "smoking causes cancer").