The study proposed a novel two-step approach for using multiple linear regression with binary variables: first, estimating probabilities using logistic regression or OLS, and then using the predicted probabilities as the dependent variable in multiple linear regression. This approach enables the modeling of risk on a linear scale, which is easier to interpret compared to the log-odds of logistic regression. However, this method relies on the initial estimation step and carries a risk of data looping; therefore, it is primarily suitable for exploratory analysis and does not replace logistic regression for probability prediction.