Do mood swings matter more than average mood for learning a skill?
How good or bad people felt on average predicted their attention and game performance much better than how much their emotions fluctuated.
Source
The relative importance of affect variability and mean levels of affect in predicting complex task performance
Study at a glance
- Design
- Other — Within-lab repeated-measures correlational study: 14 game sessions with an unannounced complexity increase after session 7; relative-importance and dominance analyses
- N
- N=253 · Undergraduates retained from 288 recruited, after excluding those with technical problems, not following instructions or responding carelessly
- Population
- US psychology undergraduates aged 18-30 learning the first-person shooter Unreal Tournament 2004
- Outcome
- Self-reported off-task attention and game performance (kills, deaths and rank) during skill acquisition and after the task change
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
Undergraduates learned a fast first-person shooter game over 14 sessions, rating 16 emotions and their off-task attention after every pair of games. Halfway through, the game suddenly became harder (more opponents, tougher bots, a bigger map) to force adaptation. The researchers computed average levels of each kind of emotion (positive/negative, activating/deactivating) and several indices of emotional variability, then used relative-importance and dominance analyses to see which predicted attention and performance.
What they found
Performance followed a learning curve, dropped sharply at the task change and recovered slowly, while off-task attention rose steadily across sessions. Average negative deactivating emotion (feeling discouraged, bored, fatigued) was the dominant predictor of off-task attention, and average positive activating emotion (enthusiastic, excited, happy) was the dominant predictor of performance, both before and after the change. Variability indices were weaker; spin and pulse never reached the importance threshold, and only fluctuation in negative emotions mattered a little.
The limits
What it doesn't show
The sample was young college students in a lab playing one perceptual-motor game for lottery incentives, so the results may not generalise to work or other skills. Emotions were measured at the same time as the outcomes, so the study cannot show that emotions or their variability cause changes in attention or performance. Off-task attention was self-reported, and participants may misjudge or under-report mind-wandering; the authors suggest eye-tracking or EEG instead. Prior game experience and emotion-regulation ability were not accounted for.
Key terms
- Affect variability
- Stable individual differences in how much a person's emotions fluctuate over time, measured from repeated mood ratings.
- Affect flux
- The within-person standard deviation of one emotion dimension across repeated measurements.
- Affect spin
- How much a person's emotional state jumps around the valence-arousal circle, i.e. variability in the kind of emotion felt.
- Relative importance analysis
- A statistical method that splits the variance explained by correlated predictors to show how much each contributes.
- Complete dominance
- When one predictor adds more unique variance than every other predictor in every possible combination of regression models.
- Adaptive performance
- Maintaining or regaining performance after task demands change unexpectedly.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What did participants do repeatedly in this study?
Common questions
Does this mean emotional ups and downs are irrelevant to performance?
Not entirely: fluctuation in negative emotions reached the importance threshold for attention and performance, but it was consistently much weaker than simply how positive or negative people felt on average.
Why was the game made harder halfway through?
To separate learning a new skill (acquisition) from coping with unexpected change (adaptation) and see whether emotions predicted the two phases differently; the pattern turned out to be very similar.
Why use relative importance analysis rather than ordinary regression?
Mean levels and variability indices are all computed from the same ratings and correlate strongly, so ordinary regression weights are hard to interpret; relative importance divides up shared variance fairly.
More on Emotion