The study relies on self-reported survey measures of fear, preventative behaviors, and risk perceptions. Such self-reported data may be less objective or precise than physiological measures (e.g., for fear) or behavioral data (e.g., for risk-taking). Furthermore, survey responses are susceptible to biases like social desirability bias, where participants may respond in ways they perceive as more acceptable or favorable, rather than their true beliefs or behaviors.

The survey data is cross-sectional, meaning the data is collected at a single point in time from three waves. This design makes determining causality problematic because changes observed over time might correlate with but not necessarily cause each other (correlation does not equal causation). In this case, the researchers observe that gender differences in preventative health behaviors are reduced or disappear when controlling for self-reported fear of the COVID-19 pandemic. However, because the data is cross-sectional, it remains unclear whether women engaged in more preventative behaviors because of higher levels of fear or if something else explains both the fear and preventative actions.

The study investigates gender differences in behavior and risk perception only during the initial stages of the COVID-19 pandemic. This time constraint is a limitation because fear of the pandemic, preventative behaviors, and risk perceptions may have evolved as the crisis unfolded. Therefore, the conclusions of the study might not generalize well over time or to later stages of the pandemic.