Find the pertinent statistics in the report. The statistics you will need in the sample size estimation are the intraclass correlation, the sample variance and the sum of squares. These statistics are usually included in the original report, but if you cannot find them immediately, check other analyses in the paper, such as ANOVA (ANOVA analyses must use the sample variance and sum of squares, and thus reports of ANOVA almost always include these statistics).
Multiply the sample variance by the intraclass correlation. Call the resulting value "x."
Divide x by the sum of squares. Call the resulting value "y."
Take the inverse of y. In other words, divide 1 by y. The result is representative of the sample size for the intraclass correlation study.
Round the result. The result you found through the calculation is likely a decimal, which does not make sense in terms of sample size (samples must be integers). This rounded number is the estimate for the sample size used in the original study.