Crash Course in Causality Quiz
This article is a quiz for the previous articles related to Causality.
- Which of the following is an example of a mediator in a causal relationship?
- a) Smoking causing lung cancer
- b) Physical activity reducing the risk of heart disease
- c) Stress causing high blood pressure
- d) None of the above
2. Which of the following is a limitation of the counterfactual framework for causal inference?
- a) It assumes that all confounding variables have been measured and controlled for.
- b) It cannot account for the possibility of reverse causation.
- c) It cannot account for the possibility of spurious correlations.
- d) It assumes that the treatment effect is the same for all units in the population.
3. What is the difference between correlation and causation?
- a) Correlation involves a relationship between two variables, while causation involves one variable causing changes in another.
- b) Correlation involves one variable causing changes in another, while causation involves a relationship between two variables.
- c) Correlation and causation are interchangeable terms that mean the same thing.
- d) None of the above.
4. Which of the following is an example of a third variable that could affect the relationship between two variables?
- a) Age
- b) Gender
- c) Education level
- d) All of the above
5. Which of the following is an example of a confounding variable?
- a) The amount of sunlight a plant receives
- b) The type of fertilizer used on a plant
- c) The height of a plant
- d) None of the above
6. What is a potential outcome?
- a) The outcome of an experiment if a certain intervention is implemented
- b) The outcome of an experiment if no intervention is implemented
- c) The outcome of a study based on observations of a natural experiment
- d) None of the above
7. In causal inference, what is the purpose of propensity score matching?
- A) To balance covariates between treatment and control groups
- B) To estimate the average treatment effect
- C) To ensure random assignment of subjects to treatment and control groups
- D) To identify confounding variables
Explanation: Propensity score matching is a technique used in causal inference to balance covariates between treatment and control groups. It involves creating a score for each subject based on their probability of being assigned to the treatment group, given their covariate values. Subjects in the treatment and control groups with similar propensity scores are then matched, which helps to balance out any differences in covariate distributions between the groups. The purpose of this is to reduce the impact of confounding variables and improve the validity of the estimated treatment effect.
8. A pharmaceutical company is testing a new drug to treat a rare disease. In a clinical trial, patients with the disease are randomly assigned to receive either the new drug or a placebo, and their symptoms are monitored over the course of several weeks. Which of the following best describes the causal relationship being investigated in this scenario?
- A) The effect of the new drug on patients’ symptoms
- B) The effect of the disease on patients’ health
- C) The effect of the placebo on patients’ symptoms
- D) The effect of the clinical trial on patients’ outcomes
Explanation: In this scenario, the causal relationship being investigated is the effect of the new drug on patients’ symptoms. The drug is the treatment variable, and the symptoms are the outcome variable. By randomly assigning patients to receive either the drug or a placebo, the researchers are controlling for confounding variables and increasing the validity of the estimated treatment effect.
9. A company is testing the effectiveness of a new employee training program to increase productivity. The company measures productivity using the number of tasks completed per hour. They compare the productivity of employees who received the training to those who did not. However, they fail to account for the fact that employees who received the training were also given a new software tool to help them complete tasks more efficiently. This is an example of which common pitfall in causal inference?
- A) Failing to control for confounding variables
- B) Overfitting the model to the data
- C) Not collecting enough data to achieve statistical significance
- D) Ignoring outliers in the data
Explanation: This scenario is an example of failing to control for confounding variables, as the introduction of the new software tool could potentially influence productivity and bias the estimated treatment effect. To obtain a valid estimate of the causal effect of the training program, the company should control for the effect of the new software tool by adjusting for it in their analysis, either by using regression or another appropriate technique.
10. A study is conducted to investigate the effect of a new drug on blood pressure. The study involves 50 participants who were randomly assigned to either receive the drug or a placebo. After one week, their blood pressure was measured. The results are shown below:
- The mean blood pressure of participants who received the drug was 120 mmHg with a standard deviation of 5 mmHg.
- The mean blood pressure of participants who received the placebo was 130 mmHg with a standard deviation of 6 mmHg.
What is the estimated average treatment effect (ATE) of the new drug on blood pressure?
- A) 10 mmHg
- B) -10 mmHg
- C) 1 mmHg
- D) -1 mmHg
Explanation: The estimated ATE is calculated by taking the difference in mean blood pressure between the treatment and control groups. In this case, the estimated ATE is:
ATE = mean blood pressure (treatment) — mean blood pressure (placebo) ATE = 120 mmHg — 130 mmHg ATE = -10 mmHg
This means that, on average, participants who received the new drug had a blood pressure that was 10 mmHg lower than those who received the placebo.
