What SPSS test do i need to use?

#1
Hi everyone

I am currently in my final year at uni and writing a paper on the effects of BMI on peak Vo2 verification using a supramaximal test in adolescent boys and girls. I have to use SPSS and I am specifically trying to analyse:
- peak Vo2 and supramaximal Vo2 in obese and non obese girls
-peak Vo2 and supramaximal vo2 in obese and non obese boys
-If there was any significance between the obese/non obese girls group and the obese/non obese boys group

If anyone knows what statistical analysis i need to run to analyse these, it would be much appreciated xx
 

Karabiner

TS Contributor
#2
Seemingly, you want to compare an interval scaled dependent variable between 2 groups.
If n > 30, you could use the Welch-corrected t-test (the Welch correction is automatically provided
in the t-test for independent groups procedure in SPSS). If n is smaller, you can instead use the
"nonparametric" Mann-Whitney U-test.

Dividing all human children into just 2 groups "obese/non-obese" sounds a bit crazy, though,
or at least crude. A boy with just 1 point away from the arbitrary BMI obesity cut-off point
is lumped together with a skinny underweight boy, but is separated from his direct
BMI "neighbour". Couldn't you use at least 3 categories (you could perform a oneway
analysis of variance with Welch correction in that case, or the non-parametric Kruskal-Wallis
H-test). Or, you could do an X-Y scatterplot of Vo2 versus BMI, and additionally calculate
the correlation between them (Pearson correlation and/or Spearman rank correlation).

Also, is it necessary to perform separate analyses for boys and girls? You could
perform a 2-factorial analysis of variance (in SPSS "general linear model" ->
"univariate analysis of variance", IIRC), with factors BMI group and gender, so
you'd have it within one analysis.


With kind regards

Karabiner