Results

For mean ratings (Figure 3), the main effect of Tonal Relatedness was significant, F(1,19) = 6.41, MSE = .044, p < .05. Mean ratings were higher for less-related targets than for related targets, indicating that target tones were judged as being more out-of-tune when they functioned as subdominants than when they functioned as tonics. The main effect of Pitch Deviation was significant, F(4,76) = 111.24, MSE = .20, p < .0001. Deviations of -35 cents were judged more out-of-tune than deviations of +35 cents, than deviations of -17 cents, than of +17 cents, and in-tune tones received the lowest values. This hierarchy in ratings was significant by post-hoc Tukey HSDs (ps < .01). The interaction between Tonal Relatedness and Pitch Deviation was not significant (F < 1).

For area scores (Figure 4), the main effect of Tonal Relatedness was significant, F(1,19) = 5.05, MSE = .015, p < .05. Performance was better for related targets than for less related targets, indicating that in-tune targets were more accurately discriminated from out-of-tune targets when they functioned as tonics than when they functioned as subdominants. The main effect of Pitch Discrimination was significant, F(4,57) = 78.0, MSE = .0074, p < .0001. Discrimination was better for 0/-35 than for 0/35, than for 0/-17, and than for 0/17 cents. This hierarchy in ratings was significant by post-hoc Tukey HSDs (ps < .01). The interaction between Tonal Relatedness and Pitch Deviation was not significant (p = .10).

For hits and false alarms (Figure 5), the main effect of Tonal Relatedness was significant, F(1,19) = 5.60, MSE = .009, p < .05, indicating that participants answered “in-tune” more often when targets were related tonics. The main effect of Pitch Deviation was significant, F(4,76) = 55.3, MSE = .035, p < .0001. Post- hoc Tukey HSDs showed significantly higher percentages of “in-tune” responses for in-tune targets than for +17 cents targets, and for +17 cents targets than for -17 cents targets (ps < .001). The interaction between Tonal Relatedness and Pitch Deviation was not significant (p = .12).

Figure 3. Mean ratings of Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch deviation (no pitch deviation = in-tune, degrees of pitch deviation +/-17, +/-35 cents). On the rating scale, 0 represents “in-tune” responses and 1 to 3 represent “out-of-tune” responses (from 1: “slightly out-of-tune” to 3: “very out-of-tune”). Error bars represent standard errors.
Figure 3. Mean ratings of Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch deviation (no pitch deviation = in-tune, degrees of pitch deviation +/-17, +/-35 cents). On the rating scale, 0 represents “in-tune” responses and 1 to 3 represent “out-of-tune” responses (from 1: “slightly out-of-tune” to 3: “very out-of-tune”). Error bars represent standard errors.
Figure 4. Areas under the ROCs for Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch discrimination (0/+17, 0/-17, 0/+35, 0/-35 cents). Chance level is at 0.5. Error bars represent standard errors.
Figure 4. Areas under the ROCs for Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch discrimination (0/+17, 0/-17, 0/+35, 0/-35 cents). Chance level is at 0.5. Error bars represent standard errors.
Figure 5. Hits and False alarms in Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch deviation (in-tune, +/-17, +/-35 cents). Hits are the percentages of “in-tune” responses in the in-tune condition. False alarms are the percentages of “in-tune” responses in the +/-17 and +/-35-cents conditions. Error bars represent standard errors.
Figure 5. Hits and False alarms in Experiment 1 presented as a function of tonal relatedness (related tonic: i, less-related subdominant: iv) and pitch deviation (in-tune, +/-17, +/-35 cents). Hits are the percentages of “in-tune” responses in the in-tune condition. False alarms are the percentages of “in-tune” responses in the +/-17 and +/-35-cents conditions. Error bars represent standard errors.