# Which statistical test should I be using?

#### RichTS7

##### New Member
Hi, im currently doing a dissertation on how people body condition score underweight dogs differently to overweight dogs so I have conducted a survey showing 5 pictures of dogs (1 severely underweight, 1 underweight, 1 severely overweight, 1 overweight, 1 average) They are then asked to grade this dogs on a scale from 1-5. 1 being extremely poor condition, 5 being optimal condition.
My results so far show that picture 1 (underweight dog) should be graded at a number 2 (as by a professional) and I have gathered results showing that 65% scored it as 1, 25% scored it as 2, 1% scored it as 3, 4% scored it as 4 and 5% scored it as 5.
For the next picture (overweight dog should be scored at 2 also)
47% scored it as a 1, 42% as a 2, 9% as a 3, 1% as a 4 and 1% as a 5

Is there a test that I can use to link these results or even the individual sets of results to see if there is a significant difference between whether people score overweight dogs worse or underweight dogs worse?

Thanks for your time - Rich

#### Karabiner

##### TS Contributor
Each participant has 2 scores, one for each dog.
You could use a statistical test which is feasible
for a) dependent measures and b) an ordinal score.
This could be the sign test. It will show you whether
participants tend to score dog A higher (or lower)
than dog B.

With kind regards

K.