Help with what test to run

#1
Hi all
I am currently planning a research project for my university dissertation but I am not clear on potential tests to run let me give you a brief overview.
First I will take 4 scenarios with 10 items measured on a Likert scale for each item (question) so from my understanding they would be the IV
the goal is to measure the each scenario against 1 scale as the DV.
From my understanding (with no data at present) a potential test to run would be in-between subjects ANOVA.
Am I correct or would this design be measured an alternative way?
Thanks
P.S sorry if this the incorrect place to ask this question:)
 

Karabiner

TS Contributor
#2
I am currently planning a research project for my university dissertation but I am not clear on potential tests to run let me give you a brief overview.
First I will take 4 scenarios with 10 items measured on a Likert scale for each item (question)
Well, that's not called Likert scale, but it is just the response format of a single item. A Likert scale is an instrument
composed of several Likert-type items, which jointly measure the same thing. It is now not clear whether you want
to analyse each of the 10 items separately, or as a sum scale.

Moreover, what do you actually mean by "I will take 4 scenarios with 10 items"? This is quite incomprehensible
(at least for me).

Could you describe the topic of your research, your research questions, the study design, the sample size,
the dependent variable, and whether these "4 scenarios" are either for 4 distinct groups, or for just one
group of people who will encounter all of them?

With kind regards

Karabiner
 
#3
Ok let me explain in more detail of my research.
My project is on "how do myths influence the identification of intimate partner violence (IPV) victims and support mechanisms available"
My plan is to take 4 victim types for example, female on female IPV, male on female IPV, female on male and male on male IPV.

Then I will use vignettes 5 for for each victim group so a total of 20 vignettes each response to the Vignette will be responded on a 7 point Likert scale.
The vignettes will be measure on 1 DV, which would be Scale items from the domestic violence myth association scale (DVMAS)

The sample size is unknown but data will be collected through online sampling.
Each participant will answer all the same questions 20 (vignettes and items from the DVMAS scale yet to be determined which items).

Therefore, I am taking 4 victim groups (male on male, female on male, male on female, female on female) and measuring against 1 DV which will be items from the DVMAS scale.
The research aims to understand how individuals perceive each victim group.

My understanding is there will be 4 IV one for each victim type.

I would like to use an ANOVA design but I'm not sure on which one to use.

Hope that makes more sense Thanks :)
 
Last edited:

Karabiner

TS Contributor
#4
The research aims to understand how individuals perceive each victim group.
As far as I can seel, you have one factor "vignette type" with 4 levels.

You could average the 5 measurements for each vignette type, and
perform a repeated-measures analysis of variance with "vignette type"
as within-subjects factor with 4 levels.

With kind regards

Karabiner
 
#5
Thanks for the help much appreciated.
I'm still a little confused though factorial ANOVA is not my favourite.
I have 1 factor which is (vignettes)
I now only have 4 subject factors (independent variables) one for each gender relationship scenario Within 4 levels (male-male, female-female, male-female-, female-male). These levels will constitute the independent variables.
Measured on 1 dependent variable.
Also this will not be a reapeted measures design.
I think that is correct
So is this a 4×4x4x4 factorial design? Seems kind of large to me
Sorry for being a pain.
Kind regards.
 
Last edited: