CursesAndMagic
Badges: 0
Rep:
?
#1
Report Thread starter 5 years ago
#1
I was just wondering when you should use 0.001 or 0.05 or 0.5 for significance, I have a correlation coefficient of 0.493 with a probability of 0.320, would this be a significant correlation?
0
reply
seraphelle
Badges: 2
Rep:
?
#2
Report 5 years ago
#2
Which coefficient are you using? Typically Pearsons R is used, is that what you have selected?
Correlations are measured -1 to 1. With a coefficient score of 0.493 I would say that was a positive correlation, but regardless of which confidence interval you use, your result is not significant.

EG:

There was a positive correlation found between the two variables explored, however, this result was not found to be significant at the .01 level (r = 0.493, n = whatever your number was, p = 0.320).

Hope that helps.
0
reply
iammichealjackson
Badges: 18
Rep:
?
#3
Report 5 years ago
#3
(Original post by CursesAndMagic)
I was just wondering when you should use 0.001 or 0.05 or 0.5 for significance, I have a correlation coefficient of 0.493 with a probability of 0.320, would this be a significant correlation?
Firstly, .5 is never used to indicate a "significant" result.

0.05 in psychology is the usual level of significance. In the case that your running loads of different hypothesis tests within the same study, then you use a Bonferonni correction so that the level of significance lower (e.g., 0.001) to take into account the fact that if your running 100 different tests the chance of any one of them being false is much higher when using the 0.05 significance level.

Sometimes, p values that are slightly larger (.20-.05 range) are called "marginally" significant, which indicates that you cannot conclusively say that its significant but its still an interesting (see https://mchankins.wordpress.com/2013...significant-2/) result.

The problem with your analysis is that you found a large effect (r=.4) however you probably have a fairly small sample, which means that only very large effects will be significant. Its best to report the r value as well as the confidence interval around it to reflect this fact.
1
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

Should the school day be extended to help students catch up?

Yes (94)
27.89%
No (243)
72.11%

Watched Threads

View All