Need urgent help with choosing the right statistical measure for my study

#1
Hello,

like mentioned in my title i need some help for choosing the right statistic procedure for my online study.

The topic of my study is gender-specific differences in aggressive behavior within a partnership, my hypotheses are:

1) Within a partnership there is a direct connection between gender and physical aggression: men are more likely to show physically aggressive behavior than women
2) There is a significant difference in indirect aggression between the sexes in the partnership: women prefer indirect aggression more than men
3) There is no significant difference in verbal aggression between men and women in partnerships

I was thinking of using independent sample t-tests but i was wondering if there are other possibilities to choose from.
Thank you in advance
Daniel
 

Karabiner

TS Contributor
#2
2) There is a significant difference in indirect aggression between the sexes in the partnership: women prefer indirect aggression more than men
BTW, the term "signficance" ( I guess you mean statistical significance) should not be part of a substantial hypothesis.
Statistical tests of significance are tools for data analysis, not part of theoretical considerations.
I was thinking of using independent sample t-tests but i was wondering if there are other possibilities to choose from.
It depends on your research design (who was observed/assessed/surveyed when and how), on the type of data collected,
and on sample size.

With kind regards

Karabiner
 
Last edited:
#3
BTW, the term "signficance" ( I guess you mean statistical significance) should not be part of a substantial hypothesis.
Statistical tests of significance are tools for data analysis, not part of theoretical considerations.
Thank you your answer.
I am using an validated self-report measure to assess aggression. Respondents simply rate each item on a 5-point scale ranging from “Not at all like me” to “Completely like me.”. The Sample Size is 153 and is almost equally split between men and women in a relationship.
 

Karabiner

TS Contributor
#4
If the observations are independent (i.e. this is not a study on couples, but on individuals unrelated to each other),
you can indeed consider the Welch t-test (one should use the Welch correction by default).

Keep in mind that a no-difference hypothesis (hypothesis 3) can not easily be investigated using significance
tests. A non-statistical result does not mean that there is "no difference". It does only indicate that it was not
possible to reject the no-difference hypothesis. You would need an "equivalence test" approach here.

With kind regards

Karabiner