What does it mean if my Pearson r correlation is not significant?

#1
Hi there,

So ive got a sample size of 400 Uni students, measuring a correlation between stress and exercise levels in minutes, and I received a non-significant results. Can anybody help explain why this is the case? Im having trouble interpreting it.
 

Miner

TS Contributor
#2
It could mean several things. First, there is simply not enough evidence that there is a relationship between stress and exercise level. This lack of evidence could be caused by 1) no relationship exists; 2) the relationship exists but is too weak to detect with the sample size; 3) the measurement system is too noisy and obscures the relationship. Second, there is a relationship, but that relationship is not a linear relationship. Start by plotting your data. If there appears to be a non-linear relationship, try a Spearman's rho correlation instead of Pearson's r. The Spearman's rho can detect a monotonically curved relationship, which Pearson's r cannot.
 

hlsmith

Not a robit
#3
Also, how is stress level collected and formatted?


I strongly agree with Miner's recommendation of plotting your data. Start with a scatterplot.