1

I ran a regression in which I have an interaction term of two factor variables.

The one of the variables has 2 levels: Condition_A & Condition_B (therefore the first one is the reference category) but the other one is called Group.Membership has 3 levels: Group_1, Group_2 & Group_3 (therefore the first one is the reference category).

What I am not sure about is how to interpret the output. What I get as part of the lm() ouput is : 

    ConditionB:Group.Membership.L         (Beta and p-value)
    ConditionB:Group.Membership.Q         (Beta and p-value)

Should I interpret the .L as the coefficient for Group_2 compared to the reference category (Group_1), and the same goes for .Q? Or there is another interpretation of this type of output?

Thank you for your help!

EDIT ----------------

I have another question. I have tested the interaction term of the Condition variable with many other factor variables. Why sometimes I get in the ouput the .L and .Q in the end, whereas in other occasions I get the levels of the second variable of the interaction term without the .L and Q.? Am I doing something wrong in my lm() ?

Pulse
  • 175
  • 2
  • 5
  • The question nominated as the duplicate was NOT about ordered factors which in R have a default of polynomial contrasts. – DWin Nov 15 '18 at 16:57

1 Answers1

4

You should provide str() from that data.frame. One of these variables is an ordered factor. Ordered factors will cause a polynomial contrast to be estimated. The L stands for linear and the Q stands for quadratic. Interpretation will require that you provide more details about the variables under study, and will probably need predictions in some sort of tabular format. Your really cannot interpret interactions just from the coefficients in a summary report.

DWin
  • 7,005
  • 17
  • 32
  • You were right, you should buy a boat :-) I will create a reproducible example tomorrow. Basically, when the variable is ordered it gives the .L and .Q whereas when it is a factor variable just the comparisons with the reference category. – Pulse Mar 26 '14 at 00:30