How to create a Coefficent table for regression analysis

I want to create a Coefficient table showing the Standard errors, P-value,t-statistics and the coefficient.

I have used the following algorithm:

D = x2fx(X, 'interaction'); b = regress(Y,D);

I tried using the following but I got an error message:

format short e t = stats.tstat; CoeffTable = dataset({t.beta,'coef'},{t.se,'StdErr'},{t.t,'tStat'},{t.pval,'pVal'})

Thank you

[Chimezie Umezie]

Here is the data: x1=[2198,6000,2198,6000, 2198,6000,2198,6000, 2198,6000,2198,6000, 2198,6000,2198,6000]'; x2=[2348,2348,4000,4000, 2348,2348,4000,4000, 2348,2348,4000,4000, 2348,2348,4000,4000]'; x3=[2348,2348,2348,2348,8000,8000,8000,8000, 2348,2348,2348,2348,8000,8000,8000,8000]'; x4=[2348,2348,2348,2348, 2348,2348,2348,2348,4078,4078,4078,4078, 4078,4078,4078,4078]'; X=[x1 x2 x3 x4 ]; Stress=[248,269,250,271,240,247,244,268,249,278,251,283,240,253,246,270]';

Below is the algorithm: D = x2fx(X, 'interaction'); b = regress(Stress,D); format short e t = stats.tstat; CoeffTable = dataset({t.beta,'coef'},{t.se,'StdErr'},{t.t,'tStat'},{t.pval,'pVal'})

The error message is: ??? Undefined variable "stats" or class "stats.tstat".

Thanks Tom

Answer 1

[Tom Lane]

What error message? You didn't define the variable stats in the code you provided, so I could imagine that might be a problem.

If you used regstats to calculate stats, then another potential problem is that x2fx inserts a column of ones, but regstats also inserts its own column of ones.