Redefining Pitch Difficulties in Sight-Singing

The pilot study was conducted with 60 volunteer participants from three faculties of Atatürk University offering professional music education. Participants were informed that recordings would be taken, and consent was obtained via Participant Information and Consent Forms, which also collected demographic data. Each participant was assigned a number. The test lasted approximately 5–6 minutes per participant. Due to technical issues, three recordings were excluded, leaving 57 for analysis. Of these, 31 were female and 26 male; 29 were Conservatory students, 19 from the Faculty of Education, and 9 from the Faculty of Fine Arts; 14 were first-year, 27 second-year, 6 third-year, and 10 fourth-year students; 15 were graduates of Fine Arts High Schools, while 42 were from other schools. Evaluation was performed by two researchers, with correct responses scored as 1 and incorrect as 0. Inter-rater reliability was found to be excellent (Kappa ≈ 0.987).
To establish face and content validity, expert opinions and group differences were examined. Since Fine Arts High School graduates had received four additional years of music education, their mean scores were expected to be significantly higher. Group means were calculated as 38.87 and 31.57. Due to unequal group sizes, the non-parametric Mann-Whitney U test was applied. The significance value of 0.036 indicated a meaningful difference in favor of Fine Arts High School graduates (p < 0.05), demonstrating high content validity. Reliability was assessed using KR-20, an alternative to Cronbach’s Alpha for achievement tests. The KR-20 value of 0.888 indicated high internal consistency and reliability of the test. The main study was conducted with 104 volunteer participants from the same faculties. Of these, 55 were female and 49 male; 59 were Conservatory students, 28 from the Faculty of Education, and 17 from the Faculty of Fine Arts; 36 were first-year, 19 second-year, 25 third-year, and 24 fourth-year students; 33 were graduates of Fine Arts High Schools, while 71 were from other schools. The same procedure as in the pilot study was followed. Unison intervals (e.g., C–C, D–D) were retained to ensure pitch accuracy but excluded from calculations due to their lack of discriminative value, leaving 56 measures for analysis.
Item difficulty indices (pj) were calculated using a 50% upper and 50% lower subgroup division due to the limited sample size. Interval difficulty levels were compared according to the variables “second note,” “interval type,” “first note,” and “interval direction.” Intervals with the second note C were treated as a separate category and excluded from subsequent comparisons. Sixteen categories were created and ranked by pj averages, grouped into five levels from easiest (Level 1) to most difficult (Level 5). This ranking was then verified using the Rasch model, with logit difficulty parameters (bj) calculated on the raw data and compared to pj averages.