To determine trends in reporting test reliability, 88 articles addressing 188 instruments in 1980, 81 articles covering 205 instruments in 1985, and 67 articles assessing 195 instruments in 1990 in the "Journal of Counseling Psychology" were reviewed. Articles were examined for the way in which reliability was discussed and reported, and were grouped into the following categories: (1) calculation of reliability estimates for the sample; (2) calculations of the reliability estimates for the sample when the instrument was developed or modified by the researcher; (3) calculations of inter-rater reliability; (4) general acknowledgment of an instrument's reliability without any specific estimated mentioned; and (5) failure to report any kind of reliability estimates. As expected, during the last decade, the general discussion of reliability appeared to increase. In 1980 reliability reports were not given for 55 percent (104) of the instruments used, while in 1990, only 24 percent (47) did not report this information. While authors' tendencies to discuss reliability estimates are growing, the reliance on past estimates appears to be strong. Analysis suggests that fluid views of reliability appear to be increasing, but static views seem to be the prevalent perspective. An upward trend is seen in the calculation of sample reliability. It is hoped that the trend in reporting estimates of reliability for each sample will increase the clarity of findings by limiting confounding factors. (SLD)