#1
Hello everyone,

I am studying the effects of framing on food neophobia and the purchase likelihood of novel foods.

Currently, I have made a pre-test where I have 4 conditions and 1 control condition where they are pictures with messages framed a certain way (emotion-promotion, emotion-prevention, information-promotion, information-prevention, and a neutral frame). I expose participants to only 2 of these conditions, followed by two 7-point semantic differential scales (likert), where I ask if they thought the message was either emotion or information, and if the message was promotion or prevention.

I need to find if there are significant differences between the scores of each condition. What statistical analysis should I run and how? I use SPSS 23. Personally, I think I just need an independent samples t-test but I do not know how to compare the scores if participants were shown different sets of conditions.

Thanks in advance,
 

Karabiner

TS Contributor
#2
I expose participants to only 2 of these conditions,
How large was your sample size? And how were participants allocated to conditions?

followed by two 7-point semantic differential scales (likert),
Either you use semantic difererential, or you have Likert scale. A measurement
cannot be both. But seemingly you have neither, but just 7-point rating scales,
which are ordinal scaled. You can treat them as interval scaled if it is sufficiently
plausible that steps between answering options are about even.

With kind regards

Karabiner
 
#3
Hi Karabiner, thanks for responding to my message!

My sample size is 50 and they were randomly allocated (evenly exposing at least 10 individuals to each condition) to two conditions.

Apologies on my confusion between the two because I thought semantic differential was under the umbrella term of a Likert scale. I used a semantic differential scale, where one scale ranged from emotion to information, and another scale that ranged from promotion to prevention. Attached is an image of what they look like:

Screen Shot 2021-01-13 at 6.10.31 PM.png

What are your recommendations for the statistical analyses I will need to run in order to show that there is a difference in ratings between these conditions? Please let me know.

Kind regards,

Consumerisms

How large was your sample size? And how were participants allocated to conditions?


Either you use semantic difererential, or you have Likert scale. A measurement
cannot be both. But seemingly you have neither, but just 7-point rating scales,
which are ordinal scaled. You can treat them as interval scaled if it is sufficiently
plausible that steps between answering options are about even.

With kind regards

Karabiner
 

Karabiner

TS Contributor
#4
Well, I don't know., ufortunately. I am not familiar with incomplete mixed
designs (i.e. a mix between repeated measures and independent samples).
I just wanted to clarify some things, in case someone else might want to
give it a try.

With kind regards

Karabiner