0

I would like to run a regression to test a hypothesis that alpha = 0 and Beta = 1. I understand minitab or SPSS regression will test whether alpha or beta is 0.

How do I test whether beta is equal to 1 using either one of the softwares?

Without needing to do extra calculation or changing the equation, is there a way to do regression(to test beta=1) using either one of these softwares ? I understand it can be done with eviews. Any “tricks” available with SPSS or Minitab?

Ferdi
  • 4,882
  • 7
  • 42
  • 62
Sheila
  • 11
  • 2
  • Look and see if the software lets you specify an offset. If it does there is a solution depending on exactly how it implements it but detailed coding advice is off-topic here. – mdewey Mar 07 '19 at 13:52
  • Thanks for your reply. Any idea where I might find a solution? Thanks. – Sheila Mar 07 '19 at 15:07
  • Would it be possible to run a normal regression (test beta is zero), then based the results manually recalculate t test to see if beta is one ? Slope minus 1 divided by standard error. – Sheila Mar 07 '19 at 15:08

1 Answers1

0

You can do this in SPSS in the GLM or UNIANOVA procedures pretty easily. In the menus, click Analyze>General Linear Model>Univariate. Specify the dependent variable and specify the independent variable as a covariate. Click the Paste button. If y is the dependent and x the predictor, you'll see something like:

UNIANOVA y WITH x /METHOD=SSTYPE(3) /INTERCEPT=INCLUDE /CRITERIA=ALPHA(0.05) /DESIGN=x.

Edit it to read:

UNIANOVA y WITH x /METHOD=SSTYPE(3) /INTERCEPT=INCLUDE /CRITERIA=ALPHA(0.05) /DESIGN=x /LMATRIX x 1 /KMATRIX 1.

With the cursor anywhere in that command, click the Run button.

The LMATRIX subcommand specifies a custom hypothesis test, with 1 times the beta coefficient for the x predictor as the estimand to be tested. The KMATRIX says to compare that to 1 rather than the default value of 0.

Look in the Custom Hypothesis Tests #1 section of the output for your results.

David Nichols
  • 840
  • 2
  • 9