Introduction: A number of studies investigating interests toward sport activities, especially among the youth, have reported that interests vary with regard to demographic characteristics (sex, age, residence), type and location of school, the local environment (community), lifestyle, season, parental support, socio-economic status, exposure to the media and numerous other factors. Research conducted among adolescents about their participation in sports and interest in sport in Croatia and on the global level have shown a strong significance of gender stereotyping. The advancement in gender equality in this area is hindered by the social constructs of femininity and masculinity according to which sport is often linked with “masculine” characteristics such as body strength, resistance and speed, as well as exceptional competitive spirit. The basic objective of this research is to determine the interest in sport expressed by third and fourth grade secondary school students attending three- and four-year programmes with regard to their experience in practicing sports. In line with this objective, four directive hypotheses have been set: Hypothesis 1. H1: There will be a statistically significant difference in the students’ factors of interest in sport regarding their sport experience. Hypothesis 2. H1: There will be a statistically significant difference in the students’ factors of interest in sport regarding the secondary school programme. Hypothesis 3. H1: There will be a statistically significant difference in the students’ factors of interest in sport regarding their knowledge of sports. Hypothesis 4. H1: There will be a statistically significant difference in the students’ factors of interest in sport regarding their parents’ sport experience. Methods: The main research was preceded by a pre-research done on a sample of 131 students (59.4% male and 40.6% female) attending the fourth year of the full-time study and the first year of the part-time specialist study at the Faculty of Kinesiology of the University of Zagreb in the academic year 2015/2016. Its aim was to assess the newly designed test on knowledge of sports. Considering the obtained assessment results, the original version of the questionnaire about knowledge of sports underwent four minor changes in questions number 1, 4, 7 and 30. The research included 1,184 students (male) attending third and fourth grades, as well as one fifth grade, of three- and four-year and one five-year secondary school programme in the Istria County. The answers to the set hypotheses were variables offered to examinees in the form of a questionnaire (paper – pencil), more precisely the general data questionnaire, the knowledge of sports questionnaire (Benassi and Bosnar, 2017), the PS2 questionnaire about interests in sports (Prot and Bosnar, 1999) and the scale of parents/guardians’ sport achievement. In the first step the arithmetic mean, standard deviation, minimal and maximal value of results and distribution of results were determined. The normality of distribution of the total results of the variable knowledge of sports was established by the Kolmogorov-Smirnov test (K-S test). The latent structure of interests in sport was determined by the principal components analysis (PCA) with the promax transformation of the principle axes. The canonical discriminant analysis was used for groups defined by the hypothesis. The multiple regression analysis was applied to explain the criterion variable of the knowledge of sports test. Results: The assessment of the measuring characteristics of the test on knowledge of sports showed satisfactory characteristics, especially in the version comprising 27 items (AS=16,4; SD=4,5; K-S test=0,06215, p˂0,01), which indicates that a scale so formed can successfully identify and discriminate those having vast knowledge of sports and those with poor knowledge of sports. The highest average grade in the PS2 questionnaire of interests in sports was noted for car racing and motorcycling (3.67), whereas the lowest was obtained for rhythmic gymnastics (1.54). The principal components analysis singled out five promax factors which jointly explain 47.92% of the total variability. Based on the sports which have been singled out, the promax factors are interpreted as outdoor sports, martial arts and strength sports, “basic” sports and sports with a distinguished aesthetic component, ball and racket sports and precision sports. To narrow down the variables to five significant factors resembles former studies’ results and represents a good base for further research. Differences in the factors of interests for groups which are defined by sports experience, secondary school programme, knowledge of sports and parents’ experience of sports have been determined by the canonical discriminant analysis. In groups defined by sports experience, the analysis found two statistically significant discriminant functions. The structure of the first discriminant function is mostly contributed by ball and racket sports and precision sports, but the athletes’ interests are more directed toward ball and racket sports, whereas non-athletes are more inclined toward precision sports. The structure of the second discriminant function is mostly contributed by martial arts, strength sports, basic sports and sports with a distinguished aesthetic component which athletes practising other sports show more interest in. Non-athletes and football players show the least interest in these sports. Group centroids have shown that the first discriminant function differentiates the non-athletes’ group and the group of athletes practicing other sports, especially football players, while the other discriminant function differentiates the group of athletes practising other sports and the group of non-athletes and football players. Considering the secondary school programme, the canonical discriminant analysis established one statistically significant discriminant function. The structure of the discriminant function is mostly contributed by ball and racket sports, and precision sports with a positive sign, as well as martial arts, strength sports and outdoor sports with a negative sign. The position of group centroids reveals that the discriminant function makes a significant differentiation of the group of students attending the three-year secondary school programme and students attending the four-year (and five-year) secondary school programme. The main difference in the differentiation of groups is made by martial arts, strength sports, and ball and racket sports. To answer the third hypothesis, the canonical discriminant analysis found a statistically significant discriminant function whose structure is mostly contributed by ball and racket sports, the only ones with a positive sign. The locations of group centroids on the discriminant function show a clear distinction among groups, especially among pupils with good or poor knowledge of sports. According to the values of group centroids it can be concluded that the group of students having good knowledge of sports assessed ball and racket sports higher, whereas those students having poorer knowledge of sports made a completely different assessment. The canonical discriminant analysis found one statistically significant discriminant analysis for the last, fourth hypothesis, and for both experiences (mother’s and father’s). The structure of the first discriminant function for the mother’s experience is mostly contributed by students with the highest interest in ball and racket sports and students with the highest interest in basic sports and sports with a distinguished aesthetic component, both with a positive sign. The structure of the first discriminant function for the father’s experience is mostly contributed by students with the highest interest in ball and racket sports with a negative sign. Group centroids show that the first discriminant function (mother’s experience) differentiates the group of students whose mothers did not practice any sport and the group of students whose mothers were members of a national team. The position of group centroids for the group of recreational players, school and club athletes, and competitors on the state or interstate level is close to the null value. Regarding the fathers’ experience, group centroids show that the first discriminant function differentiates the group of students whose fathers also did not practice any sport on any level and the group of students whose fathers competed on the state and interstate level. The position of centroids for recreational, school and club athletes, competitors on the regional level and national representatives is close to the centre, both on the positive and negative side of the discriminant function. The group of students whose mothers were national representatives is mostly discriminated from the group of students whose mothers did not do any sports by ball and racket sports, basic sports and sports with a distinguished aesthetic component. The group of students whose fathers were non-athletes is mostly discriminated from the group of students whose fathers competed on the state and interstate level by ball and racket sports. The multiple regression analysis established a significant correlation between the total result of the knowledge of sports and four out of five promax factors, ball and racket sports making the basic contribution to this multiple correlation, while outdoor sports do not have a significant value. Based on the results obtained by the last two analysis it can be concluded that all the four set hypotheses are completely confirmed. Conclusion: The obtained results reveal a wide range of sports for which examinees indicate their preference, i.e. they indicate the sports they would like to practice. Our secondary school students’ sports interests are primarily directed toward “male” and partially “neutral” sports confirming, among other things, the high significance of gender stereotyping in the field of sports. The clearly defined promax factors show that secondary school leavers in both three- and four-year programmes have a pretty stable structure of sports interests. The main advantage of this research is that the results of interests in sports have been collected and analysed on a representative sample obtained in the area of a whole local government unit. Therefore, the results obtained in this research can be generalised to the whole youth population in Croatia. The next step in the construction of a test on knowledge of sports should include the formation of an extended scale. Indeed, it is difficult to construe any general knowledge test with a small number of questions, which has also been confirmed by other research whose tests consisted of a hundred or more items. The higher the number of items, the higher the reliability. More precisely, to obtain a coefficient of 0.80, the scale should be expanded to 35 questions, whereas for alpha coefficient values of 0.85, the scale should consist of 49 items. Moreover, it is recommended to expand the Prot’s PS2 interest in sports scale to new sports which have primarily appeared in the last 20 years.