Joost Leunissen

Who they are

Integrative data analysis is a different operation from meta-analysis and the difference matters for what the result can say. Rather than pooling the summary statistics each study reported, it pools the raw participant-level data from those studies and analyses them together as one dataset. That removes a layer of dependence on how each original team chose to summarise its own results, and it allows questions to be asked that no individual study was designed to answer.

The question here was the hedonic character of nostalgia, which sounds narrow and is not. The literature had largely treated nostalgia as a positive emotion with some wistfulness attached, and the applied claims built on it, that nostalgia supplies meaning, continuity and social connection, tend to be phrased as though it were straightforwardly good to feel. Across 41 experiments, what inductions reliably moved was positive affect and ambivalent affect. Negative affect did not significantly change.

Ambivalence is therefore not a side note. It is part of the signature of the emotion, and a description of nostalgia that leaves it out has described something simpler than the thing that was measured.

Key papers

Leunissen, J., Wildschut, T., Sedikides, C., & Routledge, C. (2021). "The Hedonic Character of Nostalgia: An Integrative Data Analysis." First author. Forty-one experiments and 4,659 participants, pooled at the level of the raw data rather than of the reported summaries. Nostalgia inductions raised positive affect and ambivalent affect, and did not significantly alter negative affect.

Frankenbach, J., Wildschut, T., Juhl, J., & Sedikides, C. (2021). Meta-analysis of nostalgia's functions and affective consequences. The companion pooling from the same year, on what nostalgia does rather than how it feels: robust main effects on self-esteem, optimism, meaning in life, self-continuity and social connectedness across 116 functions studies and 39 affect studies, 3,556 participants.

Key ideas

Nostalgia is ambivalent by construction, not positive with a wistful edge. The ambivalent component moves with the positive one.

Negative affect does not shift. That is a real finding and it constrains how the emotion should be described.

Integrative data analysis pools raw participant data rather than reported summaries, which is why this result can say something the individual studies could not.

Credit the first author. This work is habitually attributed to the senior names on the research programme rather than to the person who led the paper.

---