This tutorial shows you how to test for multicollinearity in a linear regression using SPSS.
For context, for models with more than one predictor, multicollinearity describes a strong correlation between two or more predictors.
Needless to say, you don't want that because your coefficients will be untrustworthy, and the model cannot attribute an explanation of variance to the predictors.
You have three possibilities to check for multicollinearity. I) A scatter plot matrix is an easy way to start, especially when you have only variables on an interval or ratio scale. II) Additionally, a bivariate correlation matrix can be requested. III) Finally, the third method allows for a very precise identification of potential sources for multicollinearity using VIF values / tolerance.
Usually only VIF is interpreted, the so called variance inflation factors. Higher values indicate that this predictor could be a problematic source in regard to multicollinearity. A very conservative approach states that the values should not exceed 2. However, in reality, higher values can be tolerated. If you have values above 10 this should concern you.
🎥 Recommended Videos:
======================
[ Ссылка ]
Literature:
========
📚 Field, Andy (2018), Discovering Statistics using SPSS, p. 402
⏰ Timestamps:
==============
0:00 Introduction
0:24 Scatterplot matrix
0:55 Correlation matrix
2:18 VIF-values
If you have any questions or suggestions regarding multicollinearity in linear regression in SPSS, please use the comment function. Thumbs up or down to decide if you found the video helpful.
#statorials
Support channel? 🙌🏼
===================
Paypal donation: [ Ссылка ]
Amazon affiliate link: [ Ссылка ]
Ещё видео!