how to interpret results

[msherman]

 

 

Dear list: I just conducted a regression analysis where I introduced a moderator variable and the interaction terms with the moderator. In a number of places the interaction was significant which indicates to me that the relation between X1 and Y are statistically different.  In addition, I did find a number of places where the interaction was not significant but for one level of the moderator variable the relation between X2 and Y was significant and for the other level of the moderator variable the relation between X2 and Y was not significant.  My question is how to interpret/report these results.

For example.

Group 1  BETA =  .13  p < .01

Group 2 BETA =  .02   ns

 

Interaction indicates that the two BETAs are not statistically different.  So what do we really have?    Thanks,   martin Sherman

[ANDRES ALBERTO BURGA LEON]

Hello to everybody:

Im compariung column proportions using the Costum tables module.

The problem is that I need to report the Z and p-value, and the module anly shows if the diferenc is significant

Is there any way to get this values in SPSS?

I wonder why they are not reported

Andrés

Mg. Andrés Burga León
Coordinador de Análisis e Informática
Unidad de Medición de la Calidad Educativa
Ministerio de Educación del Perú
Calle El Comercio s/n (espalda del Museo de la Nación)
Lima 41
Perú
Teléfono 615-5840

[statisticsdoc]

Martin
The beta of .13 may be significantly different from 0 (the value proposed by the Null hypothesis) and at the same time the difference between a beta of .15 and .02 may not be significantly different from 0 (as proposed by a Null Hypothesis about group differences in b). Yoiu can conclude that there is a sighificant and positive relationship between X and Y in Gtroup 1 but you cannot make an assertion about a significant difference in BETA between Group 1 and Group 2 because the interaction term is not significant.
HTH
Steve Brand

www.StatisticsDoc.com

---

From: Martin Sherman <[hidden email]>

Sender: "SPSSX(r) Discussion" <[hidden email]>

Date: Fri, 25 Jun 2010 15:27:18 -0400

To: <[hidden email]>

ReplyTo: Martin Sherman <[hidden email]>

Subject: how to interpret results

 

 

Dear list: I just conducted a regression analysis where I introduced a moderator variable and the interaction terms with the moderator. In a number of places the interaction was significant which indicates to me that the relation between X1 and Y are statistically different.  In addition, I did find a number of places where the interaction was not significant but for one level of the moderator variable the relation between X2 and Y was significant and for the other level of the moderator variable the relation between X2 and Y was not significant.  My question is how to interpret/report these results.

For example.

Group 1  BETA =  .13  p < .01

Group 2 BETA =  .02   ns

 

Interaction indicates that the two BETAs are not statistically different.  So what do we really have?    Thanks,   martin Sherman