Multivariate Analysis with Linear Regression

When you have found correlated relationships among the variables, that's a great news. You can use the information to explain what is happening and predict what is going to happen. But, you don't stop there. More often than not, that correlation you found could involve something called 'confounding', which is a fancy word for a 'hidden cause' - another variable that creates such relationship.

And this is where you want to perform Multivariate Analysis. It can help you investigate further to see if there are such 'confounding' elements among the variables and find out how much of the effect the prediction variables have on the target variable independently.

In this seminar, Kan introduces Multivariate Analysis by using Linear Regression models, discusses why and when you need, and walks you through how you can perform it in Exploratory.

Agenda:

• Correlation and Confounding
• Difference between Univariate and Multivariate Analysis
• How to Interpret Prediction Models for Multivariable Analysis
• Limitation

Learning Statistics Series

• Part 1 - Introduction to Hypothesis Testing - Link
• Part 2 - t-Test: Comparing Two Means - Link
• Part 3 - Kruskal Wallis Test: Comparing Several Means with No Assumption - Link
• Part 4 - Chi-Square Test - Are These Two Categorical Variables Related? - Link

Past Seminars

You can check other past seminars on this page.

For the upcoming seminars, check out our Exploratory Online Seminar page!