The purpose of this study was to develop the Attitude Scale for Condom Use of Japanese University Students (ASCU-J) and to examine whether the Knowledge,Attitude,and Behavior Model (KAB model) predict condom use for Japanese university students.

　The sample data were collected in January 2006 for our first survey and January 2007 for our second survey.In the first survey,349 Japanese university students (male＝175,female＝174,M＝20.15) were asked to answer the questionnaire that was composed of the STDs Knowledge Scale (STDKS) (Kihara,et al,2000) ,and the Attitude Scale for Condom Use (ASCU) ,which was translated from Sexual Risks Scale (Dana D.et al.,1997) in English to Japanese.

　Stepwise exploratory factor analysis (SEFA) and reliability analysis were conducted to explore the factor structure of the ASCU and to confirm the reliability of a factor on this scale.In order to examine whether the STDKS and the ASCU predict the condom use of Japanese university students,covariance structure analysis (CSA) was conducted.

　In the second survey,253 Japanese university students (male＝122,female＝131,M＝20.08) were asked to answer the ASCU-J with 7 items that was developed from SEFA.In order to confirm the structural validity of this scale,confirmation factor analysis (CFA) was conducted on these 7 items.

　The results of the study showed that SEFA identified a one-factor solution with 7 items.The factor demonstrated acceptable internal consistency with Cronbach's alpha reliability value (0.87) .The result of CFA showed that the structural validity of this scale satisfied the high level of statistical requirements.CSA identified the high level of statistical requirements of the KAB model.That model showed that path correlation from the STDKS to the ASCU-J was not statistically significant whereas path correlations from the STDKS to the Condom Use and from the ASCU-J to the Condom Use were statistically significant.

　The major findings of this study support that the Japanese university student version of the ASCU-J satisfied statistical reliability and structural validity.Furthermore,the examination of the KAB model partially supported prior findings that knowledge and attitude modify behavior.On the other hand,the knowledge was negatively associated with the behavior and did not associate with the attitude in the KAB model.This result indicates that the lecture style of health education about STDs has a limitation to modify behaviors.Future studies need to examine educational programs that effectively modify the attitudes for condom use positively.