Abstract
CONTEXT: Several language variations of YANS have been published. There is a rationale in grouping languages where one variation is usable for several countries. The people of four ex-Yugoslavian countries do speak practically one language whatever its present name. AIM: To make a Serbian version of YANS which would be usable in Serbia, Croatia, Bosnia and Herzegovina and Montenegro. SETTINGS AND DESIGN: Translation and test-retest survey in a secondary school. MATERIAL AND METHODS: The translation process was performed according to the relevant guidelines through a standard procedure: English-Serbian-English-Consensus (people fluent in English)-Pilot assessment (20 students). The survey was performed in a Belgrade high school. Test YANS was completed by 244 students (response rate 98,8%, males 44%). Retest YANS was carried out among 60 randomly chosen students from the primary sample (response rate 96.7%, males 58%). In the statistical analysis we used the Olsen's model of classification as well as previous validation of YANS. A reliability measure for analyzing survey items was Cronbach's alpha. Determination of the mean differences between test and retest with respect to normal distribution of data was performed with the Student's paired t-test. RESULTS: Factor analysis between items grouped into four factors showed no significant association, except for a weak but negative one between two specific factors. The internal reliability (Cronbach's alpha) was 0.721 and assessed as acceptable. The test-retest comparison did not reveal any significant differences. Mean overall YANS score was 2.76, which is higher than on testing in Sweden, very similar to the Brazilian one, and lower than in Belgium and China. CONCLUSION: Serbian version of YANS is a valid and reliable research instrument. It may also be used in Croatia, Bosnia and Herzegovina and Montenegro.