Full Answer Section

Article 2: Incorrectly Implying Causation (Example)

• Source: A cross-sectional study examining the relationship between social media use and self-esteem (replace with a real article if you find one within the last 10 years). For this example, let’s imagine the title is: “Self-Esteem in the Age of Instagram: A Correlational Study.”

• Summary: This hypothetical study surveys a group of young adults, measuring both their time spent on social media platforms like Instagram and their self-esteem levels (using a validated self-esteem scale). The researchers find a negative correlation: individuals who spend more time on social media tend to report lower self-esteem. They might be tempted to conclude that social media use causes low self-esteem. However, this conclusion is premature. The cross-sectional design only captures a snapshot in time. It’s equally plausible that individuals with pre-existing low self-esteem are more drawn to social media, perhaps seeking validation or comparison. Or, a third factor, like social anxiety, could be influencing both social media use and self-esteem. The study only establishes a correlation, not a causal link. The researchers might acknowledge this limitation in their discussion, but if they overstate the strength of the relationship or imply causation without sufficient evidence, they are misusing correlation analysis.

Why Causation Was Appropriate in Article 1 and Not in Article 2:

The key difference lies in the study design and the researchers’ approach.

• Article 1 (Meta-Analysis of Longitudinal Studies): The longitudinal nature of the included studies establishes temporal precedence. Childhood adversity precedes adult mental health problems. This is a critical piece of evidence for inferring causality. Furthermore, the meta-analysis would include studies that attempted to control for confounding variables, strengthening the argument for a direct effect of ACEs. While absolute proof is rarely possible in social science, the combined evidence from multiple longitudinal studies, coupled with control for confounders, provides strong support for a causal inference.

• Article 2 (Cross-Sectional Study): The cross-sectional design only shows a relationship at one point in time. It cannot tell us which came first: social media use or low self-esteem. The negative correlation could be interpreted in multiple ways. Because of this lack of temporal precedence and the inability to rule out other explanations, implying causation is inappropriate.

What Would Be Needed for Article 2 to Justify a Statement of Cause?

To establish a stronger case for causation, the researchers in Article 2 would need to conduct a longitudinal study. They would need to measure social media use and self-esteem at multiple points in time, tracking changes over time. If they found that increases in social media use predict subsequent decreases in self-esteem (and not the other way around), they would have stronger evidence for a causal link. They would also need to carefully consider and control for potential confounding variables, such as personality traits, social support, and life events. Even with a longitudinal design, it’s essential to be cautious about causal claims. Researchers often use language like “increases the risk” or “is associated with” rather than definitively stating “causes.” However, a longitudinal design and careful attention to confounders are essential steps towards justifying a causal inference.