Hi. As with regression, tests of the intercept test the hypothesis that an obtained value (e.g., a single mean for univariate ANOVA, a combination of means in MANOVA) differs from 0. This test only makes sense to interpret if 0 has a meaning for your dependent measures; that is, if a subject can actually have the value 0.
------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------
Original Message:
Sent: Wed January 04, 2023 04:40 PM
From: Kevin Duff
Subject: intercept in MANOVA
I am running a MANOVA to compare two groups on 17 different dependent variables. In my output below, I see the variable that shows the group effect (for NeuroDX_record_change_1). But what does the intercept in this model mean? This is also statistically significant, but it is unclear to me what this is testing. Any thoughts? Thanks, Kevin
Multivariate Testsa |
Effect | Value | F | Hypothesis df | Error df | Sig. |
Intercept | Pillai's Trace | .745 | 101.294b | 6.000 | 208.000 | <.001 |
Wilks' Lambda | .255 | 101.294b | 6.000 | 208.000 | <.001 |
Hotelling's Trace | 2.922 | 101.294b | 6.000 | 208.000 | <.001 |
Roy's Largest Root | 2.922 | 101.294b | 6.000 | 208.000 | <.001 |
NeuroDX_record_change_1 | Pillai's Trace | .182 | 7.689b | 6.000 | 208.000 | <.001 |
Wilks' Lambda | .818 | 7.689b | 6.000 | 208.000 | <.001 |
Hotelling's Trace | .222 | 7.689b | 6.000 | 208.000 | <.001 |
Roy's Largest Root | .222 | 7.689b | 6.000 | 208.000 | <.001 |
a. Design: Intercept + NeuroDX_record_change_1 |
b. Exact statistic |
------------------------------
Kevin Duff
------------------------------
#SPSSStatistics