5.2 Research Designs and Causal Logic

Choosing a Defensible Design

Evaluation design is the plan for comparing what happened with what would reasonably have happened without the program. In practice, health education specialists often work in schools, clinics, worksites, and community agencies where perfect control is not possible. The CHES exam expects you to select a design that is ethical, feasible, and strong enough for the decision.

An experimental design uses random assignment to place eligible participants into intervention and control conditions. Random assignment helps reduce selection bias because known and unknown differences should be distributed across groups. This design can support strong causal claims, but it may be impractical or unethical when services cannot be withheld or when community partners require universal access.

A quasi-experimental design includes comparison but not random assignment. A school may compare one campus receiving a peer education program with a similar campus that starts later. A county may compare pre-policy and post-policy clinic visits. Quasi-experimental designs are common in public health because they fit real programs, but they require attention to baseline differences, timing, contamination, and outside events.

A one-group pretest-posttest design measures the same participants before and after an intervention. It is simple and useful for learning whether participants changed, but it cannot prove the program caused the change. A local news campaign, seasonal trend, or repeated exposure to the test could explain some improvement. In an exam item, this design is often acceptable for a small quality improvement decision but weak for a major causal claim.

Cross-sectional designs measure variables at one point in time. They are useful for describing needs, behaviors, attitudes, or associations. They do not establish temporal order. A cross-sectional survey can show that adults with higher perceived risk report more screening, but it cannot prove perceived risk caused screening.

Repeated measures, interrupted time series, matched comparison groups, and wait-list designs can improve field evaluations. For example, monthly clinic referral data for a year before and after a new navigation protocol can show whether a shift occurred beyond normal variation. Matching sites by size, demographics, and baseline rates can reduce obvious selection problems.

Study threats matter because they shape conclusions. History means another event occurred during the evaluation period. Maturation means participants changed naturally over time. Attrition means people dropped out, possibly in a patterned way. Testing means the pretest affected posttest responses. Selection means groups differed before the program began. The best CHES answer often names the threat and selects a practical way to reduce it.

For CHES preparation, do not memorize design names in isolation. Practice linking each design to the claim it can support. A randomized study can support stronger causal language when implemented well. A quasi-experiment can be persuasive when comparison and baseline information are strong. A descriptive design can still be exactly right when the purpose is assessment, monitoring, or program improvement rather than proof of causation.

Scenario Review Checklist

• Identify the relevant CHES Area of Responsibility.
• Locate the program stage in the scenario.
• Match the answer to evidence, stakeholders, and ethics.
• Reject choices that are premature, unsupported, or outside scope.