Dear Abdur Rauf,

Generally speaking, the interpretation of correlation coefficient values is the same across disciplines. In point of fact, a number of methods are available to  evaluate the significance of the correlation. In the book "Research in Education" Best and Kahn (2006) enumerate some useful methods. One quick and easy way to interpret correlations is the crude method. As such,  r values larger than 0.6 are significant.Naturally, larger magnitudes approaching 1 are better. Another method of interpreting r is the use of variance of the measure we want to predict. This method is made possible by squaring r . This value helps the researcher to predict the percentage of variance not explained by the predictor variable ; that is- the part of relationship magnitude due to other factors including sampling error.

Best regards,

R. Biria 

@Reza Biria, I download the book you mentioned. This really helpful for me. But here is one problem. On the page number 372 of this  book, the correlation coefficient values are divided into five categories i.e. 

.oo to .20

.20 to .40

.40 to .60

.60 to .80

.80 to 1.00.

Let my correlation coefficient values are 

r= [.00  .40  .50  .20 .80  .90  .30 .60].

Now if I want to use the first category,

r>=.00 & r=.20 & r

Hello Rauf,

Just use the values on the right as your yard stick. Anything above 0.6 is meaningful.

Best of lick,

R. Biria

Dear All

I think all the correlation factors must be considered. Because the significance depends also on the data size which will make for example 0.5 significant. Dear abdur, you need to read some book around correlation carefully to understand it and learn how to interrupt your data.

All the best.

Ammar 

Yes, I know that the significance of correlation can be tested by the null hypothesis. According to this hypothesis, a correlation can be considered significant if the p-value is less than the level of significance. What I am looking for is to find the percentage of occurrence of the different range of correlation coefficients. This means that I need to consider all the significant correlations, not only those higher than 0.6. But I don't know how to select the different ranges as shown in my first reply. thanks

Let I have diffrent correlation coefficient values like, r= [.00 .40 .50 .20 .80 .90 .30 .60]. 

is this meaningful to  find its mean value i.e 0.4625.?