Question 26·Medium·Evaluate Statistical Claims: Observational Studies and Experiments
A nutrition company conducted a study to test the effectiveness of a new dietary supplement on reducing low-density lipoprotein (LDL) cholesterol levels. One hundred adult volunteers were randomly assigned to two groups: 50 took the supplement daily for 3 months, and 50 took a placebo. At the end of the study, the supplement group’s mean LDL cholesterol had decreased by 12 milligrams per deciliter (mg/dL), while the placebo group’s mean LDL cholesterol had decreased by 3 mg/dL. The difference between the two mean decreases was found to be statistically significant.
Which of the following conclusions is best supported by the results of the study?
For questions about conclusions from studies, first identify whether it’s an observational study or a randomized experiment: only randomized experiments justify cause-and-effect statements, and only random sampling (from a population) justifies generalizing to that whole population. Next, pay close attention to phrases like “statistically significant” (which indicates real evidence of an effect) and check whether each choice stays within what the data actually support—avoid answers that overreach, such as claiming effects for all people or specific percentages when the study only reports averages for a limited group.
Hints
Focus on the study design
Is this an observational study or an experiment? Pay attention to whether the researchers assigned treatments or just observed what people did.
Think about what random assignment allows
When participants are randomly assigned to treatment and control groups and a difference is statistically significant, what kind of conclusion (causation vs. just association) is justified for the people in the study?
Watch out for overgeneralizing
Look for answer choices that talk about all adults or specific percentages of all people. Ask yourself: Did the study actually collect enough information to make such broad claims, or are those conclusions too strong?
Use the phrase 'statistically significant'
If a difference is statistically significant, does that support 'no evidence of an effect,' or does it suggest there is evidence of a real effect in the study?
Step-by-step Explanation
Identify the type of study
Notice that 100 adult volunteers were randomly assigned to two groups: a supplement group and a placebo group. This is an experiment (not just an observational study) because the researchers actively assigned the treatment (the supplement). Random assignment allows us to talk about cause and effect for the people in the study.
Use the information about the results
The supplement group’s mean LDL decreased by 12 mg/dL and the placebo group’s mean LDL decreased by 3 mg/dL. The difference in mean decreases is mg/dL. The problem says this difference is statistically significant, which means it is very unlikely to be just due to random chance in how people were assigned to groups. So the supplement appears to have a real effect in this study.
Decide what we can and cannot conclude
Because of random assignment and the statistically significant result, we can say that, for the volunteers in this experiment, the supplement caused a larger average decrease in LDL than the placebo. However, we cannot:
- Claim the exact same 9 mg/dL effect for all adults everywhere (the volunteers were not a random sample of all adults).
- Guarantee that a certain percentage of all individuals who take the supplement will see a decrease.
- Say there is no evidence of an effect, because statistical significance is evidence that there is an effect in this study.
Match the valid conclusion to the answer choices
Now compare each answer choice with what is justified:
- Claims about all adults or a fixed percent of adults go beyond the study’s scope.
- Claiming no evidence contradicts statistical significance.
- The only statement that stays within the study’s volunteers and describes a causal, average effect is: For the volunteers in the study, taking the supplement caused a larger average decrease in LDL cholesterol than taking the placebo.