Hi all,

In prep for analyzing my MSc dataset (bird breeding and feeding ecology data from the Arctic), I'm trying to understand discriminant function analysis (among other things!). I've digested most of what my biostats text has to say, and have a basic understanding. I've spent today and yesterday going through examples and tutorials on Minitab, the software my university provides. I've poked around online reading what I can to enhance what I've learned.

Here's my current problem. In the Statshelp file for discriminant functions Minitab 16 provides, the problem is: assign students into one of 3 groups based on 2 variables: IQ and motivation. I'm having a hard time interpreting the output that Minitab gives (the description in the Statshelp file isn't helpful).

In the output, one of the lines is "Linear Discriminant Function for Groups". I was expecting to look for the discriminant functions so I could find the standardized coefficients that would indicate which variables were most important to each discriminant function. Alternatively, I hoped to see the loadings. Maybe they're both there, but I'm missing them. Here's the output:

The conclusion for from this forum is this:

Equation 1: Likelihood of a given student being put in group 1 = -237.85 +1.57*IQ + 5.15*Motivation

Equation 2: Likelihood of student x being put in group 2 = 170.58 + 1.19*IQ + 4.66*Mot.

etc.

So, does the first column, which I've turned into the group 1 equation, give us the 1st discriminant function, and the 2nd column the second, etc.?

If I'm right, then from there, can I conclude that motivation contributes most to each DF, and therefore to each classification of the student?

Thanks for any help you can provide! It's really tricky to apply these tools to my own data when I can't even interpret what the Minitab Statshelp file gives me!

In prep for analyzing my MSc dataset (bird breeding and feeding ecology data from the Arctic), I'm trying to understand discriminant function analysis (among other things!). I've digested most of what my biostats text has to say, and have a basic understanding. I've spent today and yesterday going through examples and tutorials on Minitab, the software my university provides. I've poked around online reading what I can to enhance what I've learned.

Here's my current problem. In the Statshelp file for discriminant functions Minitab 16 provides, the problem is: assign students into one of 3 groups based on 2 variables: IQ and motivation. I'm having a hard time interpreting the output that Minitab gives (the description in the Statshelp file isn't helpful).

In the output, one of the lines is "Linear Discriminant Function for Groups". I was expecting to look for the discriminant functions so I could find the standardized coefficients that would indicate which variables were most important to each discriminant function. Alternatively, I hoped to see the loadings. Maybe they're both there, but I'm missing them. Here's the output:

The conclusion for from this forum is this:

Equation 1: Likelihood of a given student being put in group 1 = -237.85 +1.57*IQ + 5.15*Motivation

Equation 2: Likelihood of student x being put in group 2 = 170.58 + 1.19*IQ + 4.66*Mot.

etc.

So, does the first column, which I've turned into the group 1 equation, give us the 1st discriminant function, and the 2nd column the second, etc.?

If I'm right, then from there, can I conclude that motivation contributes most to each DF, and therefore to each classification of the student?

Thanks for any help you can provide! It's really tricky to apply these tools to my own data when I can't even interpret what the Minitab Statshelp file gives me!

Last edited: