Discriminant Analysis in SPSS (DV with Three Levels) with Assumption Testing

This video demonstrates how to conduct and interpret a Discriminant Analysis (Discriminant Function Analysis) in SPSS using a dependent variable with three levels. The assumptions for Discriminant Analysis are reviewed. Discriminant Analysis is used to determine the probability of categorical group membership using interval or ratio predictor variables.

[@danial_amini]

Thanks Dr. Grande, that was very informative! I'm ashamed to say that as a civil engineer I always carry out these calculations without really caring about the underlying assumptions and the normality requirements. Generally, I would carry out the analysis and if the results are good, then I'll see if the requirements are met.

[@shwanahassan3860]

Thank you for sharing your knowledge.

[@rijamanandhar9281]

Thank you so much Dr. Grande. Really helpful.

[@DrGrande]

You're welcome - thank you for watching.

[@fathimasahira1896]

@@DrGrande can you provide this data set

[@aprilbrooks8672]

Looking at the classification results will show us the best results and if the group membership was collected and analyzed correctly. I assume this is why we look for the discriminant analysis to calculate the probability of membership.

[@drrachitkashyap7231]

Dr Grande... thank u for making this video... its very helpful ... but could you tell me that what is function 1 and function 2.

[@fatimagillani7410]

Hi Dr Grande. What does it mean if I have one of my discriminant functions non significant? What does it imply for the groups?

[@shrutiraut7862]

This was a great video! Todd, if we have NPS as dependent variable and other drivers which respondents rated on a point scale (IVs). Which technique of regression is more suitable?

[@shaliquems3968]

It would be really helpful if you can add another video on test data based on the discriminant function arrived at.

[@smiley1136]

Thank you :)

[@tetlleyplus]

Great video! thanks Dr. Grande!, I just want to know if the predictors are not independent, can I use PCA as predictors (components are orthogonal, i.e. independent)?

[@hoshaminhaji3693]

Hi dear Dr. Todd Grande
, i have some questions from the assumption of DA. 1. can i compare the sig. value of Box's M with 0.01 or not? 2. If the independent variables didn't distribute normal. what should i do? . 3. Is it possible if the independent variable is categorical variable?

[@sarahburrous7374]

This is a lot of data and analysis to look at. I found it interesting that the scatter plots weren't great as well as the Structure Matrix results. Is this okay to still go with the data if things don't look very good?

[@perrycatlin7353]

What does it mean if your Shapiro-Wilks is

[@sedatbatmaz9226]

How do you write up the results of a discriminant analysis? Which parts of the SPSS output need to be presented in the results section of our manuscript?

[@cortexcortexx1340]

+Sedat Batmaz I too would like to know this

+Sedat Batmaz i also want to know that, if u learn from anywhere please inform me too

[@zohredehghannezhad7384]

Yes

[@joydeepmondal3491]

you didnot tell what variables are best predictor among all the variables taken. Can you tell that please?

[@anjanbasumatary1510]

can we have multiple dependant variable?

[@amandachaffin6636]

Once the data had been adjusted on the SPSS main screen it became overwhelming! It was hard to discriminate between the probabilities!

[@cassieperoulis4881]

+Amanda Sutton I agree that it became a little over-whelming at one point. This is a video that I have to review for sure.

[@cardinalgrad11]

I found this concept very confusing and overwhelming. The assumptions as well as the extent of the results of the data was difficult to differentiate between all of the probabilities as well as the various analysis tools.

[@youngsuksuh9511]

It could be better...