What is the need to look beyond p-value? How do we check the practical importance of the statistically significant result? Identify the effect needed for practical importance Use Confidence Intervals Conclusion Sources What is the need to look beyond p-value? In my last post, I mentioned how we can use p-values to select important variables in a linear model. In general, this sounds like an easy basis of testing a hypothesis, but that is not where the story ends.
Introduction Is there a relationship between the medical charges and the predictors? Which variables have a strong relation to medical charges? Are there any multicollinear features? Is the relationship linear? Non-constant variance of error terms Correlation of error terms Interpretations Conclusion Sources Introduction Regression analysis is a powerful statistical process to find the relations within a dataset, with the key focus being on relationships between the independent variables (predictors) and a dependent variable (outcome).