Charles Rotter
European Psychologist published a paper by Clara Kühner, Gudrun Reindl and Hannes Zacher of Leipzig University on October 23, the product of five survey waves, two attention checks, four linear mixed-effects models and nearly fifteen months between submission and acceptance, on the question of how cheerful climate change makes German workers feel.
From the abstract:
We collected data from N = 1,313 employed adults in Germany across five measurement waves with 3-month time lags. Results of mixed-effects models revealed that, at the between-person level, negative climate emotions (i.e., fear, guilt, and sadness) were positively and moderately related to different dimensions of PEB. […] At the within-person level, climate change–related fear was positively, albeit weakly, related to PEB, whereas climate change–related joviality was negatively and weakly related to PEB.
https://doi.org/10.1027/1016-9040/a000614
PEB is pro-environmental behaviour. Climate change–related joviality is exactly what it says on the label, and it is a variable now.
The participants came from Norstat, an online panel company. Of the 1,931 eligible respondents who completed the first wave, 618, nearly a third, were thrown out over the five waves as “careless responders” for failing attention-check questions, which left 1,313. By the fifth wave 628 were still answering.
Here is the instrument:
The instruction was “Please indicate the extent to which climate change evokes the following feelings in you,” followed by 60 adjectives representing 11 distinct climate emotions. Response options ranged from 1 (=not at all) to five (=very much).
https://doi.org/10.1027/1016-9040/a000614
Sixty adjectives, every three months, for a year. The list came from a standard mood checklist, the PANAS-X, with “climate change” bolted onto the front, so somewhere in Germany a panel member got to the joviality items, one of which is “energetic,” and was asked to rate how energetic climate change makes her feel. Then she did it again three months later, and again after that, in the interest of capturing the fluctuations.
The paper reports the results with a straight face:
mean values for climate emotions ranged between 1.19 (joviality) and 1.83 (attentiveness), suggesting rather low average levels of climate emotions in our sample.
https://doi.org/10.1027/1016-9040/a000614
On a scale where 1 means “not at all,” the highest of the 11 averages was 1.83. Germany’s workers, asked four times a year for their feelings about climate change, answered, on average, that they didn’t have many, and the study went ahead and measured those for a year anyway.
The other side of the ledger, the green behaviour, came out of the same questionnaire. Nobody checked a meter or a grocery receipt, the survey just asked things like “How often do you cut down on heating or air conditioning to limit energy use?” and “During the past three months, how often did you eat beef/pork/poultry?” Membership in an environmental group was a yes or no, so the authors scored no as 1 and yes as 5 and averaged it in with the how-often questions. The temperature of the laundry went on a 3-point scale and was stretched to fit the 5-point one.
The authors know what this is:
the mixed response format may decrease reliability and self-report ratings of PEB may be biased due to social desirability and retrospective recall errors (Kormos & Gifford, 2014).
https://doi.org/10.1027/1016-9040/a000614
Social desirability is the psychologist’s term for people telling the survey what sounds good. The paper they cite, a 2014 meta-analysis by Kormos and Gifford of the University of Victoria, compared what people said they did against what meters and observers recorded and found that 79% of the variation in the self-reports was left unexplained by the actual behaviour.
So the experiment is this. A person rates a feeling from 1 to 5, rates her own habits from 1 to 5 from memory, in the same sitting, and the two numbers are set against each other as cause and effect.
Then comes the lab equipment. A mixed-effects model, stripped of the vocabulary, splits each person’s answers in two. One part is how that person usually answers compared with everyone else, which the authors call between-person. The other part is how this particular quarter’s answer differs from her own usual, which they call within-person. The within-person part is the study’s selling point, since it is supposed to show what happens when somebody feels a bit more afraid of climate change than she normally does.
Here is what it found, from Table 3. A person who reported one point more climate fear than her own usual also reported 0.03 points more energy conservation and 0.04 points more of what the paper calls citizenship behaviour, such as talking to others about the environment. One point more climate joviality than usual came with 0.06 points less citizenship.
Three hundredths of a point on a five-point scale.
That was three significant results out of 44 within-person tests, 11 emotions against four kinds of behaviour, at the usual p < .05 bar, where chance alone would be expected to turn up two or so. The authors conclude, correctly, that the swings in emotion “were not consistently associated with PEB.”
None of these numbers was measured in any sense a chemist or an engineer would recognize, just volunteered by people working through a questionnaire, many of them about feelings the respondents said they mostly didn’t have. Statistical software doesn’t know that. It will happily separate a woman’s “usual” climate joviality from this quarter’s climate joviality to two decimal places, when both are just her guess at how a news topic makes her feel on the day the email arrived, and it will set that against her guess at how often she turned the heat down, which she knows has a right answer. Run 4,903 of those guesses through a model with a separate baseline for every respondent and out come coefficients to two decimal places and p-values to three, every one of them carrying the full authority of the method and none of it borrowed from anything that was weighed, counted or metered.
The between-person results got a theory. People who generally said climate change made them jovial reported less energy conservation and more meat, but more citizenship and more car-pooling, public transport and cycling. The paper puts it this way:
individuals who feel more enthusiastic and energetic when thinking about climate change conserve fewer resources and consume more meat, but they also engage in more citizenship PEB and use sustainable transportation options more frequently.
https://doi.org/10.1027/1016-9040/a000614
And explains the meat:
The negative associations between joviality and conservation PEB and reduced meat consumption, respectively, could be further explained by the potential mismatch between the hedonically positive nature of joviality and the self-restrictive nature of these behaviors.
https://doi.org/10.1027/1016-9040/a000614
Translated, people who enjoy themselves don’t enjoy going without. It took a five-wave panel, a 60-adjective inventory and “the theoretical lens of motivational direction (Elliot et al., 2013)” to get there, from a joviality score that averaged 1.19 out of 5.
Climate change–related shyness, measured with items such as “timid” and averaging 1.22, came with more meat eating too, and that got a theory as well:
One possible explanation is that reducing meat consumption often requires individuals to express climate-related concerns in social contexts, for example by refusing offered food or explaining dietary choices. […] Individuals who feel shy about their climate-related concerns may be less likely to engage in such visible behavioral changes.
https://doi.org/10.1027/1016-9040/a000614
The timid don’t refuse the schnitzel, per a feeling about climate change that averaged barely above “not at all.”
The models, by the paper’s own Table 3, explain between 6.5% and 20.4% of the variation in the behaviour scores, with all 11 emotions in at once. The authors also draw a practical lesson from the gap between the stable and the quarter-to-quarter results:
interventions aiming to foster PEB may benefit from targeting stable emotional tendencies rather than transient emotional states.
https://doi.org/10.1027/1016-9040/a000614
The data came from a larger collection of 10 survey waves between August 2022 and June 2024, and the paper notes, “So far, 15 studies using data from this data collection have been published,” to which this one adds another.
What would an intervention targeting someone’s stable tendency toward climate joviality look like?
And who is going to tell the 618 people who failed the attention check that they may have been the most accurate respondents in the study?
=====
Quotes and figures from Kühner, Reindl and Zacher 2026, https://doi.org/10.1027/1016-9040/a000614, used under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Chart drawn from the paper’s Table 2.
