## Applying T-test on fraction of respondents

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

Bartyboy
Posts: 13
Joined: Mon Apr 07, 2014 10:59 am

### Applying T-test on fraction of respondents

Hey there everyone.

I'm in the following situation:

100 respondents filled in my questionnaire:
--> The first 50 persons got Code 1 (They are assigned to Condition 1)
--> The last 50 persons got Code 2 (They are assigned to Condition 2)

I now want to execute an independent T-test within condition 1 ( 50 people) and within condition 2 ( also 50 people) SEPERATELY.
I have been searching for solutions but found nothing so far...

Could someone help me out ?
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

### Re: Applying T-test on fraction of respondents

use select cases.

You sure you want to do this though?
Bartyboy
Posts: 13
Joined: Mon Apr 07, 2014 10:59 am

### Re: Applying T-test on fraction of respondents

@ Gutnr

Hey,
I have to compare 2 means within each condition.

Within condition 1 ( first 50 people) there is a before(no manipulation of dependent variable) and after(manipulation of dependent variable) measurement which I have to compare for significant difference or not...

This also applies to condition 2. ( the other 50 people)

Is my reasoning correct in this case ?
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

### Re: Applying T-test on fraction of respondents

no, use repeated measures with pre/post as within subject and treatment condition as between subject factor.
Bartyboy
Posts: 13
Joined: Mon Apr 07, 2014 10:59 am

### Re: Applying T-test on fraction of respondents

Hey,

I'm using some scales (General sens of power of Anderson & Galinsky ) that have been tested in literature before.
But... here's my confusion:

Altough the scale items have been tested before (by other authors) on internal consistency via cronbach's alfa,
My SPSS output tells me a few scale items should better be removed.

Should I remove those scale items ?

Greetz
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

### Re: Applying T-test on fraction of respondents

that's not something you can answer easily, there is discussion about this.
Some people will indeed remove those items, others may just use the original scale.

just to be sure: did you recode the appropriate items?
and are you sure you should remove them (i.e., is it a strong effect?)
Bartyboy
Posts: 13
Joined: Mon Apr 07, 2014 10:59 am

### Re: Applying T-test on fraction of respondents

I did recode the appropriate items

Effects are in the range of 0,7 . Some are even less
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

### Re: Applying T-test on fraction of respondents

did you look at scale if item deleted?
Bartyboy
Posts: 13
Joined: Mon Apr 07, 2014 10:59 am

### Re: Applying T-test on fraction of respondents

Yes I did.

Gutnr I applied a paired samples t-test...
with 4 pairs:
1) Initial powerfeeling downstairs - powerfeeling upstairs (Condition 1)
2) initial powerfeeling upstairs - powerfeeling downstairs ( Condition 2)
3) Need for power downstairs - need for power upstairs (Condition 1)
4) Need for power upstairs - need for power downstairs (Condition 2)

I tried repeated measures but find it rather complex to use...

Does this gives me the same results as an repeated measures ?

Greetz
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

### Re: Applying T-test on fraction of respondents

you don't take into account the interdependency between some variables like this.
try to look at some case studies for a 2x2 design

### Who is online

Users browsing this forum: No registered users and 1 guest