Adding a variable for estimating minimal sign effect

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 1
Joined: Mon Apr 28, 2014 8:00 am

Adding a variable for estimating minimal sign effect

Postby EvHa » Mon Apr 28, 2014 8:39 am


I have a poweranalysis releated question. I have a baseline measure and I want to know what my minimal significant detectable effects are in four years. One of my outcomes is loneliness, its dichotomized in being lonely yes or no. With the baseline measure I have for example predicted the chance on being lonely now and being lonely in four years in my intervention and control group. Now I want to know what difference within difference effect I can detect significantly in four year (to see if i have to increase my N in the intervention group).
To clarify, if the chance on being lonely in the controle group increases from 0.32 to 0.42 the effect is +0.10, if the change of being lonely in the invervention group increases from 0.29 to 0.36 the effect is +0.07, the difference within difference effect is 0.10-0.07 is 0.03. I want to know what effect I can detect with my data, is a difference within difference of for example 0.045 significant.
How do I calculate the minimal difference within difference significant effect with spss? My plan was to use linear regression (the difference within difference as outcome) and adding a variable that gives the coefficient for alpha 0.05 and 0.01. But how do I do that?

Any suggestions, thanks

Who is online

Users browsing this forum: No registered users and 2 guests