What the Evidence Says—and Why It Conflicts

This chapter examines the strongest quantitative evidence about boredom and physiological arousal: what was pooled, what the reported estimates mean, and why the average cannot serve as a definition of every boredom episode. The practical judgment is not whether to believe the meta-analysis or the reports of restless boredom. It is how to hold an average lower-arousal pattern together with the study conditions and measurement choices that make the pattern conditional.

What was pooled, and what the estimates mean

The evidence comes from a systematic review and three-level random-effects meta-analysis of research on the relationship between boredom and arousal. The review searched Web of Science, PsycInfo, PubMed, and ProQuest Dissertations & Theses Global for quantitative studies involving human, non-clinical participants. The inclusion question was specific: could a study provide an effect size, or enough information to calculate one, for the synchronous relationship between a person’s boredom and arousal? This is not a survey of everything ever called boring. It is a synthesis of studies that measured both parts of the relationship in a usable way.

The abstract reports 214 effect sizes from 72 unique samples, totaling 6,570 participants. Those effect sizes came from 56 studies. Seventeen effect sizes were reported as non-significant in their original papers without numerical information. Because the main analysis needed numerical effect sizes, it proceeded with 197 effects nested in 69 unique samples from 54 studies, while the authors also ran sensitivity analyses that included the 17 null effects. The abstract’s 75 correlational effects and 122 experimental effects add up to those 197 effects in the main analysis. The distinction matters: the headline total describes the evidence collected, while the primary estimates describe the numerically analyzable set.

There is one counting inconsistency in the published text. The abstract and the results report 72 unique samples, while one sentence in the discussion refers to 75. The methods and results use the 72-sample count in the structure of the analysis, so that is the count used here. This does not overturn the main direction of the findings, but it is a reminder that even a strong synthesis deserves careful reading of denominators. A number is part of an empirical claim, not decorative precision.

The review combined two kinds of evidence that answer related but different questions. In correlational studies, the researchers asked whether people who reported more intense boredom also tended to report lower arousal. The aggregate correlation was r = −0.13, with a 95% confidence interval from −0.22 to −0.05. The negative direction means that greater boredom was associated with lower arousal on average. The magnitude was small: it describes a tendency across observations, not a sharp separation between bored and non-bored people.

In experimental studies, researchers compared arousal in a boredom condition with arousal in a control or comparison condition. The aggregate standardized mean difference was d = −0.40, with a 95% confidence interval from −0.59 to −0.22. In plain language, experimentally induced boredom was associated with lower average arousal than the states used as comparisons. Because the comparison conditions included a mixture of positive, negative, high-arousal, low-arousal, and neutral states, the estimate says that boredom was lower relative to that varied collection.