Abstract: What changes have taken place in Chinese college students’ mental health since the resumption of college entrance examination in 1977? Many researchers tried to answer this question by using the Symptom Checklist 90 (SCL-90). However, these researchers drew different conclusions: someone found that college students have significantly decreased in mental health since the 1980s and they were a high risk group, while others found that college students’ mental health was good in general and has become better since then. Generally speaking, the traditional meta-analysis is an available method to solve these controversies; it can obtain a general conclusion by re-analyzing studies published before. However, it is a pity that these meta-analysis studies still arrived at conflicting conclusions about college students’ mental health, mainly due to their neglect of publication year effect and using different norms. Given this neglect, Twenge (2000, 2001, 2004, 2011, etc.) introduced a special meta-analysis named cross-temporal meta-analysis, which has been applied to examine the change in mean scores on psychological measures over time.
Using the cross-temporal meta-analysis, the present study examined the changes of Chinese college students’ scores on the SCL-90 in the past 25 years (1986~2010). Two hundred and thirty-seven samples of college students (N = 318972) were included in the data. The analysis not only included the description of the nine factorial mean scores of the SCL-90 changing over time as a whole, but also included the changes of different groups separately.
The results showed that: (1) Correlations between SCL-90 mean scores and year of data collection were negative, and the year could explain 4%~36% of the total variation of the nine factorial mean scores. The SCL-90 factorial mean scores have decreased 1%~13% from 1986 to 2010, with factorial mean scores of interpersonal sensitivity, depression and hostility decreasing more significantly. These results suggested that Chinese college students’ psychological problems have decreased and their mental health has been improving gradually over the past 25 years. (2) The correlation between year and college students’ mental health was different from group to group. First, year of data collection accounted for more variations for post-freshmen than freshmen, suggesting that the increase of college students’ mental health was mainly due to post-freshmen. Second, year of data collection accounted for more variations for key than non-key universities, suggesting that the improvement of college students’ mental health was mainly due to those from key universities. Third, although most of male students’ factorial mean scores of the SCL-90 were lower than that of female students, the changes over time of male students were higher than that of female students, that is, male students improved more rapidly than female students in mental health. Fourth, though college students from urban areas have slightly lower scores than college students from rural areas on the dimensions of the SCL-90, the changes over time of college students from urban areas were partly higher than college students from rural areas, suggesting that the mental health of college students from urban areas improved more rapidly than college students from rural areas.

Key words: SCL-90, mental health, college students, cross-temporal meta-analysis, meta-analysis, social change