A researcher wants to determine whether a new tutoring program causes higher test scores. She recruits 60 students and lets them choose whether to join the tutoring group or the control group. What is the primary problem with this design?
AThe sample size is too small to detect an effect
BSelf-selection means the groups may differ in motivation before the program starts, confounding the results
CThe researcher should have used a within-subjects design instead
DField experiments cannot establish causation
Allowing participants to self-select into conditions destroys the logic of a true experiment. Students who volunteer for tutoring may already be more motivated or academically engaged than those who opt out. Any score difference at the end cannot be attributed to the tutoring program alone — it may reflect pre-existing differences. Random assignment prevents this by ensuring that motivation, ability, and all other individual differences are distributed evenly across groups by chance.
Question 2 True / False
Random assignment and random sampling are two names for the same procedure in experimental research.
TTrue
FFalse
Answer: False
Random sampling refers to how you select participants from the population to include in your study — it determines external validity (generalizability). Random assignment refers to how you allocate participants who are already in your study to experimental conditions — it determines internal validity (ability to infer causation). A study can have one without the other: a convenience sample (no random sampling) that is randomly assigned to conditions can still establish causation within that sample, just with limited generalizability.
Question 3 Short Answer
What is the difference between a between-subjects and a within-subjects experimental design, and give one advantage of each?
Think about your answer, then reveal below.
Model answer: Between-subjects: different participants are in each condition. Advantage: no carryover effects — participants cannot be influenced by exposure to other conditions. Within-subjects: the same participants experience all conditions. Advantage: each participant serves as their own control, eliminating individual-difference variability and requiring fewer total participants.
The choice between designs involves a trade-off. Between-subjects designs require more participants (separate groups for each condition) but avoid order effects, practice effects, and demand characteristics that arise when the same person completes multiple conditions. Within-subjects designs are statistically more powerful for detecting effects because individual differences are removed from the error term, but they require counterbalancing to control for carryover effects.