Trying to understand hypothesis testing in psychology, but I can't work out what tests to use for my hypotheses.
One example hypothesis is comparing three Independent variables with one dependent variable.
> low social support (satisfaction and number), low adaptive coping (seeking social support, stoicism, and self-care), and high perceived stress would be associated with high depressive symptomatology.
I think you would separate each IV and compare them individually with the DV.
>IV1xDV
>IV2xDV
>IV3xDV
So I was thinking a t test or maybe an ANOVA.
The second hypothesis is comparing a nominal variable with two dependent variables (i think?).
> rural participants will have lower perceived stress and depressive symptomatology levels than their urban counterparts.
I was thinking chi squares and an ANOVA?
I've looked through my books but it doesn't tell me what tests I need to perform, just that it requires pearson's r and stuff like that.
General SPSS thread.
>>7993516
anova is for testing groupd differences, you want linear regression
>>7993536
oh okay thanks
>>7993516
What hypothesis test are you using? There's plenty, but I know you psychologists only use the most basic babby statistics
>Nominal variable
This should be banned.
>>7993545
Trying to work out what tests to use. Multiple regression looks like it fits for Hypothesis 1 though.
No idea what to do for Hyp 2
>>7993551
Categorical*?
Don't even bother. You will never understand it and it will only lead to flawed research
http://www.nature.com/news/first-results-from-psychology-s-largest-reproducibility-test-1.17433
>>7993608
It's an assignment so I kinda have to do it. My fault for bumming out in stat class last year.
>>7993608
To be fair, without looking at the article I can tell it's gonna be something to do with the significance level. The 0.05 significance level should be considered more of a guideline than an absolute threshold. I'm sure with a lot of studies in psychology you can get decent p-values when trying to reproduce them, but sure they may not reach the significance level.
>>7993516
One-way ANOVA should be good for what you want to do.
T-test is decent if you're comparing one group against another with a small sample size, otherwise if you have a sample size of n > 30, you would use a z-test. You can form a hypothesis about the mean of the population in this case: "is x less than y?" or something. Keep it simple, and really try to understand why you're doing what you're doing.
>>7993608
>Don't even bother. You will never understand it and it will only lead to flawed research
i'd give you the same advice on getting a girlfriend u memer
