Pretest

20 participants (instrumental instruction = 0 to 10 years, mean = 2.8  3.3 years, median: 1.5) were asked to rate the degree of completion of each melody on an 8-point scale (from 1 = poorly completed, to 8 = completed). None of them participated in the main experiment. Completion ratings (averaged over the melody set) were analyzed with an ANOVA with Tonal Relatedness (Related / Less-related) as within-participant factor: melodies ending on tonally related targets were judged as more complete (5.47) than melodies ending on less-related targets (4.51) (F 1 (1,19) = 26.66; MSE = .35; p < .001) (Table 1). An additional ANOVA on completion judgments with melodies as random variable confirmed the main effect of Tonal relatedness (F 2 (1,11) = 23.94; MSE = .22; p < .001).

To study a possible influence of musical expertise, participants were separated into two groups: 8 moderately experienced participants (instrumental instruction = 3 to 10 years, mean = 6.3  2.4 years, median = 5.0) and 12 less-experienced participants (instrumental instruction = 0 to 2 years, mean = 0.5  0.8, median = 0). Musical Expertise, added to the ANOVA design as a between-participants factor, interacted with Tonal Relatedness ( F(1,18) = 10.18; MSE = .24; p < .01): the tonal relatedness effect was more pronounced for the moderately experienced participants (F(1,18) = 41.59; p < .001), but still significant for the less-experienced ones (F(1,18) = 8.14; p < .05) (Table 1).

Table 1. Completion judgments for the pretest in Experiment 1 presented as a function of tonal relatedness (related/less-related) and musical experience (moderately-experienced/less-experienced). Between-participants standard errors are in brackets.
Table 1. Completion judgments for the pretest in Experiment 1 presented as a function of tonal relatedness (related/less-related) and musical experience (moderately-experienced/less-experienced). Between-participants standard errors are in brackets.