Randomization gives the strongest evidence of cause
The hospital wants to know whether the program itself causes a change in exercise habits. To show cause and effect, the evaluation must rule out other explanations for any change. Random assignment to the program or to usual care makes the two groups alike in both known and unknown factors, so a difference afterward can be attributed to the program. This is a randomized controlled design, the strongest design for testing whether an intervention works.
Why other designs are weaker
Every other design leaves room for the groups to differ in ways besides the program. A matched comparison group from another clinic is balanced on the matched factors, such as age, but may still differ in motivation, access to exercise space, or clinic culture. Repeated measures before and after the program without a control group show change over time but cannot separate the program's effect from other influences such as season, news reports, or natural improvement. Self-selected comparison groups, women who chose not to join, are the weakest: women who choose to attend are likely already more motivated, so they would exercise more even without the program.
| Design | Control group | Random assignment | Strength of causal evidence |
|---|
| Randomized to program or usual care | Yes | Yes | Strongest |
| Matched group from another clinic | Yes | No | Moderate |
| Repeated measures, before and after | No | No | Weaker |
| Participants versus women who declined | Yes | No (self-selected) | Weakest |
Key terms behind the answer
Confounding occurs when a third factor, such as motivation, affects both who receives the program and the outcome. Selection bias occurs when the way people enter groups makes the groups different from the start. Randomization is the main tool to prevent both. In a hospital program, usual care for the control group can include standard clinic advice, and the program can be offered to them later, which addresses ethical concerns about withholding a benefit.
Application
For the next cycle, the hospital could randomly assign eligible postmenopausal women with low bone mass either to the eight-session program or to usual care, measure exercise habits in both groups at the same time points, and compare the results. A difference between the groups can then be credited to the program with confidence.
Exam takeaway
Key point Does the program cause the change? Choose random assignment with a control group. Matching, before-and-after measures, and self-selected groups give weaker evidence.