Hello, I hope you are all well. I am finishing my MSc and am struggling quite a bit with organising/ analysing my data.
I have conducted a study looking at the effect of different model sizes on the recall of clothing advertisements. First, participants were exposed to 45 images of clothes on 3 different types of models (Thin, average, plus = 15 in each). They were then shown images of the clothes the thin/average/plus size models wore + some extras (15 in each) and required to indicate whether they have seen them before. Lastly, we asked them to input their BMI (split into thin, average, plus).
The aims are as follows: 1) to determine whether the diversity of the model impacted recall 2) whether a higher percentage of recall occurs when participants view models of the same BMI. 3) whether people recall more for thin vs other
Unfortunately, the participant BMI split is very uneven, with most being average - this makes it hard to test the second aim.
i hoe this make sense :) any help is very appreciated!
https://www.ibm.com/docs/en/spss-statistics/24.0.0?topic=coach-statistics
Click the link above and follow the steps. Good luck!
Do you have the concrete BMI numbers and you could use them as a numerical covariate?
If not, and you only have the three categories, where almost all are "average", then forget about using the BMI as a variable. If you build 3 groups, but if the "thin" and "plus" are extremely small, then there is no value and comparing these three groups is quite non-representative.
Natasha Fulton
First, identify your dependent (DV) and independent variables (IDV).
After that, check the type of your variables- categorical( nominal and ordinal) or Continuous ( ratio or interval).
Now, check whether DV and IDV are parametric (metric) or not. [Otherwise, they are non-metric (non-parametric)]
if variables are non-metric (both DV/IDV) then apply crosstabs (chi-square).
if DV= metric and IDV= non-metric, use hypothesis testing.
if DV= metric, IDV=metric; use Correlation and regression.
if DV= non-metric, IDV= metric, use discriminant analysis or binary regression.
N.B.: Above mentioned techniques are for Bi-variate analysis.
Apart from that, you can use mean, median, mode counts, and percentages. (For univariate analysis).
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers