Society of Actuaries (SOA) PA Practice Exam

When conducting univariate analysis on a Continuous predictor variable, which aspect should be examined?

1. The correlation with the binary target

2. The histogram of the distribution

3. The count of observations for each factor

4. The box plot of the target variable

The correct answer is: The histogram of the distribution

When conducting univariate analysis on a continuous predictor variable, examining the histogram of the distribution is essential. A histogram provides a graphical representation of the data distribution, allowing one to visualize the shape, central tendency, and spread of the continuous variable. By analyzing the histogram, you can identify characteristics such as skewness, kurtosis, and the presence of outliers, which are crucial for understanding how the predictor might influence the target variable and for making decisions about further statistical analysis or modeling approaches. While other options may provide insights, they do not focus specifically on the distribution characteristics of the continuous variable itself. The correlation assessment with a binary target can indicate the strength and direction of the relationship but does not describe the predictor's distribution. The count of observations applies more to categorical variables, and while box plots can illustrate the relationship between a continuous predictor and the target, they do not depict the distribution of the predictor on its own. Therefore, the histogram is the most relevant aspect to examine in this context.