The Student Room Group
Uni students - Complete our survey >>
Uni students - Complete our survey >>
Uni students - Complete our survey >>

Stats interpretation - Correlations

Basically, I've got a set of data (2 variables) and I've run a Spearmans rank correlation analysis on them.

This has come up with a Spearmans rank of -0.44 but a two tail P of 0.06.

Does this mean that there is a non-significant (as P>0.05) weak inverse correlation between the two or am I interpreting it incorrectly?

Thanks.
Original post by aphex.twin
Basically, I've got a set of data (2 variables) and I've run a Spearmans rank correlation analysis on them.

This has come up with a Spearmans rank of -0.44 but a two tail P of 0.06.

Does this mean that there is a non-significant (as P>0.05) weak inverse correlation between the two or am I interpreting it incorrectly?

Thanks.

Assuming that you are meant to use a 5% significance level, then yes you are right. (I think. It's been ages since I did this!)
Reply 2
Avatar for aphex.twin
aphex.twin
OP
Yeah, I'm going for a 5% significance level.

If the P value is > 0.05 what is the interpretation of the test?

Does it mean that there is no correlation between the two variables? (even though the SRCC is -0.44)

Thanks
(edited 13 years ago)
Original post by aphex.twin
Yeah, I'm going for a 5% significance level.

If the P value is > 0.05 what is the interpretation of the test?

Does it mean that there is no correlation between the two variables? (even though the SRCC is -0.44)

Thanks

It means you reject the hypothesis that there is a correlation (i.e. You conclude that there is no significant correlation. You have NOT proven that there isn't correlation; you have just failed to give enough evidence that there is a correlation!)


Think about it this way: Suppose I gave you a coin and told you to flip it 10 times then decide whether or not it's biased. You flip it and record 6 heads and 4 tails.
Would you conclude that the coin is biased based on this? No, because there's not enough evidence. But does this mean for sure that the coin isn't biased? Again no; that's just the best educated guess that you can make with the given information.

So similarly in this question, can you conclude that there is a significant correlation? No, because there is a lack of evidence.
Does this mean for sure that there isn't any correlation? No, that's just the best guess you can make.

Quick Reply

Related discussions

Latest

Trending

Trending

Articles for you