Foundations of Experimental Research
A hospital tests whether adding music therapy during recovery improves patient mood compared to standard care. What is the independent variable in this study?
What is music therapy?
A rehabilitation clinic wants to track how patients recover following joint surgery. Researchers measure pain intensity at weeks 0, 3, and 6 to evaluate recovery. This setup would require this type of design.
What is a within-groups (repeated measures) design?
A study reports a p-value of 0.03. This means there is only a 3% probability the results are due to this.
What is chance?
A research study has participants complete Exercise A, then immediately Exercise B, then Exercise C in one session. By Exercise C, participants are exhausted and perform poorly. To control for this, researchers should use this technique.
What is counterbalancing?
A health blog reports that “vitamin D cures depression” based on a correlational finding. The claim misrepresents the evidence because the research did not demonstrate this.
What is causation (casual relationship)?
A nutrition researcher is studying the effect of a new diet plan on health outcomes. All participants receive the diet plan, but they choose their own exercise routines. Exercise represents this type of variable.
What is a confound?
This key feature of experimental research involves the researcher actively changing a variable to observe its effects.
What is manipulation (of the IV)?
A medication study shows statistically significant results with p = 0.02, but the mean difference between groups is only 2 points on a 100-point pain scale. The finding likely lacks this.
What is practical (clinical) significance?
Patients are tested on a cognitive task four times over two months. Their scores improve each time, but researchers suspect that the gains reflect test familiarity- a pattern representative of this threat.
What are practice effects?
Researchers originally planned a within-groups study with 20 participants trying three different medications in counterbalanced order. However, internal validity threats proved impossible to control, so they switched to a between-groups design. To maintain 20 participants per medication condition, they now need to recruit this many total participants.
A researcher finds that patients who attend more physical therapy sessions report less pain, but those same patients also tend to be more motivated to recover. Strengthening which element of experimental design would best address this problem?
What is control? (to address the 3rd variable problem)
A hospital wants to compare patients recovering from knee replacement with those recovering from hip replacement on a new rehab protocol. To make the groups as similar as possible on factors like age, fitness level, and comorbidities, researchers would need to use this kind of assignment.
What is matching?
A wound healing study reports medication effects were significant with p = 0.03, but researchers note that 40% of the medication group dropped out compared to 10% of the control group. These drop-out rates threaten the validity of the p-value because of this.
What is differential attrition (attrition bias)?
Hospital administrators feel bad that one unit isn't getting the new patient education program, so they give that unit extra nursing staff and resources. This represents this threat to internal validity.
What is compensatory equalization?
A study examines whether patients improved, stayed the same, or worsened across three different treatment protocols. To analyze their data, the researchers should use this type of statistical analysis.
What is a chi-square?
In an experiment testing whether a new medication lowers blood pressure, participants are randomly assigned to a treatment or placebo group, and only the medication group shows improvement. This combination of design features provides the evidence needed for this type of claim.
What is a cause-and-effect (causal) conclusion?
A rehabilitation researcher wants to test a new balance training program to see whether it outperforms the current standard therapy. Patients in one group complete the standard therapy, while others complete the new balance program. The standard-therapy group represents this type of group.
What is a no-treatment control group?
A cognitive training study shows significant improvement (p = 0.02) when comparing baseline to post-test scores. However, participants were selected for having unusually poor memory at baseline. Skeptics argue the improvement may simply reflect this.
What is regression to the mean?
A diabetes education study spans eight months. During this time, three participants' doctors change their medications, two participants start new jobs with different schedules, and one participant's spouse is hospitalized. Events like these represent this type of problem.
What are history effects (historical confounds)?
A study reports F(3, 116) = 4.82, p = 0.003, η² = 0.11. Although significant, critics argue we should still be worried about the fact that this much variance remains unexplained.
What is 89%?
A researcher measures anxiety before and after a stress-management workshop but forgets to record which sessions participants actually attended. Without knowing when or whether the treatment occurred, the study fails to establish this key feature of an experiment.
What is temporal precedence?
Because each participant provides data for all conditions, fewer total participants are required, and this quality is inherently higher.
What is (statistical) power?
Researchers test three different physical therapy techniques for shoulder mobility. Each participant completes only one technique. Results reveal a significant ANOVA. To determine which specific techniques differ from each other, follow-up analyses should involve multiple comparisons using this specific type of statistical test.
What is an independent t-test?
Researchers compare anxiety outcomes between two therapy types. Both groups average a 15-point reduction, but Group A's individual reductions range from 13-17 points while Group B's range from 2-28 points. Group A's results are more likely to reach statistical significance because it has this quality.
What is low (within group) variance?
A dataset is organized with one row per participant. The columns are labeled: Participant Number, Technique A, Technique B, and Technique C. Data organized in this manner represents this type of research design.
What is a within-groups (repeated measures) design?