ISAT - International Student Admissions Test International Student Admissions Interpreting Scientific Data 1 — Questions and Answers
Question 1: A study comparing two pain medications reports: Drug A reduced pain scores by 18 points (95% CI: 14–22), Drug B reduced pain scores by 15 points (95% CI: 11–19). What can researchers most accurately conclude?
- Drug A is significantly better than Drug B because it has a higher mean reduction
- The two drugs cannot be compared because confidence intervals were used
- The results are inconclusive because the confidence intervals overlap (Correct answer)
- Drug B should be discarded since it has a lower mean reduction
Correct answer: The results are inconclusive because the confidence intervals overlap
Overlapping confidence intervals indicate that the difference between the two drugs may not be statistically significant. When CIs overlap, we cannot confidently conclude one treatment is superior to the other without further testing.
Question 2: Researchers test whether sleep duration affects exam scores. They find r = –0.61 between hours of sleep deprivation and exam performance. Which conclusion is best supported?
- Sleeping more causes students to perform better on exams
- There is a strong negative association between sleep deprivation and exam scores (Correct answer)
- Sleep deprivation explains 61% of the variation in exam scores
- Students who sleep less will always score lower on exams
Correct answer: There is a strong negative association between sleep deprivation and exam scores
r = –0.61 indicates a moderately strong negative correlation — as sleep deprivation increases, exam scores tend to decrease. Correlation does not imply causation, and r² (not r) gives the proportion of variance explained (~37%).
Question 3: A scatter plot shows the relationship between study hours and test scores for 40 students. The line of best fit has a positive slope, but several data points are widely scattered around the line. What does this pattern suggest?
- Study hours is a perfect predictor of test scores
- There is no relationship between study hours and test scores
- Study hours is a positive but imperfect predictor of test scores (Correct answer)
- The data contains an error and should be discarded
Correct answer: Study hours is a positive but imperfect predictor of test scores
A positive-sloped best-fit line indicates a positive trend, meaning more study hours is generally associated with higher scores. However, wide scatter around the line shows that study hours alone does not fully predict outcomes — other variables play a role.
Question 4: In a longitudinal study, Group A (n=100) receives a dietary intervention and Group B (n=100) receives no intervention. After 6 months, Group A shows a 5% reduction in cholesterol and Group B shows a 4.8% reduction. The p-value is 0.74. What should the researchers conclude?
- The dietary intervention is effective because Group A had a larger reduction
- There is insufficient evidence that the intervention caused a significant difference (Correct answer)
- The study is invalid because both groups showed cholesterol reduction
- A larger study would definitely prove the intervention is effective
Correct answer: There is insufficient evidence that the intervention caused a significant difference
A p-value of 0.74 is far above the conventional threshold of 0.05, meaning the difference between groups is not statistically significant. The minimal difference in outcomes (5% vs 4.8%) combined with the high p-value suggests the intervention may not have a meaningful effect.
Question 5: A table shows bacterial colony counts at three temperatures: 25°C = 340 colonies (SD=45), 35°C = 890 colonies (SD=52), 45°C = 210 colonies (SD=48), with n=20 samples each. Which conclusion is best supported by the data?
- Bacteria grow best at 35°C, and this difference is likely meaningful given the large gap and small SDs (Correct answer)
- Temperature has no effect on bacterial growth because all SDs overlap
- The experiment is flawed because three conditions were tested simultaneously
- Bacterial growth at 25°C and 45°C are statistically identical
Correct answer: Bacteria grow best at 35°C, and this difference is likely meaningful given the large gap and small SDs
The mean at 35°C (890) is dramatically higher than at 25°C (340) or 45°C (210), and the standard deviations (~45–52) are small relative to these differences. This large signal-to-noise ratio strongly suggests the 35°C advantage is a real and meaningful effect.
Question 6: A clinical study reports that patients taking Drug X have a 2% absolute risk reduction in heart attacks compared to a placebo. The relative risk reduction is reported as 40%. A journalist headlines: 'Drug X cuts heart attack risk by 40%.' What is the most accurate critique of this headline?
- The headline is accurate because 40% risk reduction is statistically significant
- The headline is misleading because it uses relative risk reduction without providing the absolute baseline risk (Correct answer)
- The headline should instead report the p-value to be meaningful
- Both the absolute and relative risk figures are equally informative for the general public
Correct answer: The headline is misleading because it uses relative risk reduction without providing the absolute baseline risk
Relative risk reduction (40%) sounds dramatic but can be misleading without the baseline. A 40% relative reduction from a 5% baseline to 3% is very different from a 40% reduction from a 50% baseline. The absolute risk reduction (2%) provides critical context for evaluating real-world impact.
ISAT - International Student Admissions Test International Student Admissions Interpreting Scientific Data 1
A study comparing two pain medications reports: Drug A reduced pain scores by 18 points (95% CI: 14–22), Drug B reduced pain scores by 15 points (95% CI: 11–19).
What can researchers most accurately conclude?