What are some statistical tests for the relationship between two variables?

#1
Hi, I posted previously, but I haven't got any response, I assume its more appropriate to post in this forum, that's why I'm posting again.

How do I test the relationship between Govt. Spending on Police and the Number of robberies (annually) I have 51 data points.

I've done Pearson's, I tried Chi, but I ran into problems and that's it!
 

Karabiner

TS Contributor
#2
Looks like a time series analysis. Or are the data
from different countries/states? In either case,
the crude number of robberies and crude amout of
spending on police could be misleading, if factors
such as varying population sizes or e.g. inflation (in
case of a time series analysis over 51 years) were
not taken into account.

Kind regards

K.
 
#4
What problems did you run into?

51 points is not enough for most methods. Regression for instance generally requires at least a hundred (as does ANOVA I believe) and time series generally needs several years of data at least if you are concerned with cycles and seasons.

If you had enough data running OLS regression with robberies as your dependent variable would be fine (assuming its interval which I assume is true).
 
#5
What problems did you run into?

51 points is not enough for most methods. Regression for instance generally requires at least a hundred (as does ANOVA I believe) and time series generally needs several years of data at least if you are concerned with cycles and seasons.

If you had enough data running OLS regression with robberies as your dependent variable would be fine (assuming its interval which I assume is true).
I'm looking into OLS regression, but what do you mean by 'assuming its interval' ?