Correlation and Linear Regression Analysis for Continuous Outcome

This chapter discusses the principles and methods of correlation and regression analysis, specifically focusing on linear regression for continuous outcomes. It explains the correlation tests—Pearson’s, Kendall’s, and Spearman’s—and the conditions under which each should be used. Emphasis is placed on assumptions such as normality and linearity, which are necessary for accurate statistical interpretation. The linear regression analysis section describes how to model relationships between dependent and independent variables. It explains simple and multiple linear regression and how to test assumptions like normality of residuals, homoscedasticity, and the absence of collinearity. Visualizations such as scatter plots are recommended to check these assumptions. Statistical measures like beta coefficients, R-squared values, and F-statistics are calculated to evaluate model fit. The chapter also touches on how to perform stepwise regression to optimize models. It also highlights the importance of reporting assumptions and ensuring they are satisfied for regression analysis to be valid.