Mood Judgments and Memory for Tunes: A Special Case of Levels of Processing?

ABSTRACT

Although levels of processing (LOP) effects are well-established in memory research, beneficial effects of “deep” orienting tasks have rarely been reported in studies of tune memory. Our prior work implicated mood judgments as one candidate for a beneficial orienting task. The current series explored both the robustness and potential explanations for that enhancement. In four experiments, we varied type of processing tasks (including mood and other putatively deep/conceptual and shallow/perceptual tasks) and the familiarity of the tunes in a recognition paradigm, which included “remember/know” judgments. Experiment 1, with low-familiarity tunes, revealed a LOP effect for two conceptual (mood, continuation) over two perceptual (contour tracing, note counting) tasks in “remember” scores. This effect was most pronounced for the mood task. In Experiment 2a (high-familiarity tunes) and 2b (low-familiarity tunes), we found superiority of distinctiveness- and categorization-based orienting tasks over the control task of loudness ratings, but again for “remember” scores only; the first two tasks were equivalent. Finally, in Experiment 3, we asked participants to compare pairs of low-familiarity tunes on mood, distinctiveness, or length (control task). The mood task led to better memory than length judgments, and the distinctiveness task was marginally superior to length judgments. All four experiments revealed LOP effects only in “remember” scores. Mood judgment was the most consistently effective orienting task. Results are discussed in relation to models of memory, including how some tasks offer particularly effective “affordances” in some domains.

KEYWORDS:

• Levels-of-processing
• episodic memory
• music
• emotion
• distinctiveness

Acknowledgments

We thank Turgut Coşkun, Ezgi Mamus, and Ezgi Bozkurt in Experiments 1 and 2, and Nia Tannis in Experiment 3, for their meticulous data collection and data coding.

Disclosure statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. We report no financial interest or benefit from direct applications of this research.

Open Practices

The data and materials used in all experiments are available upon request from any of the authors. None of the experiments were preregistered.

Notes

1. d’ is a recognition sensitivity measure that is calculated by transforming hit and false alarm rates to z scores and then subtracting the false alarm z from the hit z rate (see Stanislaw & Todorov, 1999).

2. As mentioned in the General Statistical Analysis Protocol at the end of the Introduction, all post hoc tests were Bonferroni tests with an overall alpha-value set at .05.

3. For G dˈ scores we also see some fluctuation across OTs due to the counting task. In Experiment 1 of our 2011 study (Mungan et al., 2011) we saw a similar fluctuation.

4. When collapsing the two experimental OTs, a 2 (distinctive/relational vs. neutral) x 2 (RK) ANOVA yielded a significant OT effect only, F (1, 35) = 14.36, p = .001, ηp2 = .29. A marginal OT effect also occurred in a one-way ANOVA on R dˈ values, F (1, 35) = 3.82, p = .06, ηp2 = .10.

5. Note that this was not quite so in Experiment 1 where, descriptively speaking, R/K false alarm rates were overall lower compared to Experiment 2a, and K type false alarms exceeded R type ones (Fig. 2).

6. When collapsing the two experimental OTs (distinctive and relational processing), there was no other effect but a cross-over interaction between OT and response type, F(1, 35) = 4.60, p = .04, ηp2 = .12, where the experimental tasks led to more R than K sensitivity, and vice versa for the neutral task. Again, similar to Experiment 2a, there also was a marginal OT effect on R dˈ once the two experimental tasks were collapsed in comparison to the neutral task, F (1, 35) = 3.57, p = .07, ηp2 = .09.

7. The largest more/less dˈ difference was observed for the mood-setting task (dˈ difference between “more” and “less” of .49), followed by the neutral task (dˈ difference between “more” and “less” of .24). The distinctiveness task created no difference (dˈ difference between “more” and “less” of +.02). Since the 3(OT) x 2(processing judgment) analysis was not of primary interest, we refrained from doing additional statistical testing on the obtained interaction effect.