Given that the evidence for cultural variation in moral psychology is at best weak, and given that the theoretical argument for moral reframing is flawed, why does moral reframing seem to work? Some evidence suggests that it may work in part because moral reframing makes an argument appear to you to fit better with your moral psychology (Wolsko, 2017). Perhaps another part of the answer is that moral reframing provides cues to the source of a message, and people are more influenced by sources they perceive as sharing their political identity (Fielding, Hornsey, Thai, & Toh, 2020). And perhaps a further part of the answer is that moral reframing can modulate how fluently people with different political identities can parse a message, and people are more influences by messages they can parse more fluently. But these speculations about how moral reframing works have yet to be tested directly, and are unlikely to be the whole story. The puzzle remains.
This recording is also available on stream (no ads; search enabled).
Why does moral reframing work?
According to Feinberg & Willer (2019, p. 4), the ‘primary explanation’ is that moral reframing of an argument influences how well the argument matches (their term) a person’s moral psychology.
But both the evidence and the theoretical basis for this view faces objections (as we saw in The Argument and Some Objections). These objections do not imply that the ‘primary explanation’ is wrong, only that we do not know that it is true. This motivates considering alternative possibilities.
We will consider three candidate explanations: perceived match, source and fluency.1
Wolsko (2017) provides evidence for the hypothesis that moral reframing works in part because it influences how well an the argument matches a person’s moral psychology. Their approach does not rely directly on Moral Foundations Theory and neatly avoids the objections to Feinberg & Willer (2013)’s position considered earlier (see The Argument and Some Objections for these objections).
Wolsko (2017, p. Experiment 1) directly measured how participants’ perceived the match between their values and the values in the message:
‘Immediately after reading the moral framing manipulation, participants […] were asked to indicate their level of agreement with a 5-item measure of salient value similarity, including: “The message above contains values that are important to me,” “The message above comes from someone who thinks in a similar way as me,” and “I share similar values with those that are presented in the message above.”’ (Wolsko, 2017, p. 287).
Like Feinberg & Willer (2013, p. Study 3), Wolsko (2017, p. Experiment 1) found that moral reframing caused an increase in conservatives’ proenvironmental attitudes and a decrease in liberals’.2 Importantly, this effect was mediated by the degree to which participants’ perceived the match between their values and the values in the message. They conclude
‘it is a perceived shift in the personal moral relevance of the message which increases the persuasiveness of these environmental appeals’ (Wolsko, 2017, p. 289).
One limit of this study is that it does not involve any manipulation of the source of the message and so cannot distinguish the degree to which a message is perceived to match participants’ values from the degree to which participants identify with the source of the message.
Perhaps moral reframing is effective in part because it provides cues to the source of a message, and people are more influenced by sources they perceive as sharing their political identity.
Hurst & Stern (2020) provide indirect support for this idea in a study on attitudes to reducing use of fossil fuels. They manipulated both the content and the source of a message. When the content matched emphasised all five foundations to match socially conservative moral foundations but was identified as originating from a liberal source, it rarely made a difference to conservative participants’ environmental attitudes.
Fielding et al. (2020) manipulated only the source of a message and measured the influence of reading the message on participants’ support for carbon tax. They found a significant effect of message source. This is evidence that people are more influenced by sources they perceive as sharing their political identity. (Schuldt, Pearson, Romero-Canyas, & Larson-Konar (2017) provide further, less direct evidence along these lines.)
They offer a bold conjecture on the basis of these results:
‘it is possible that the values framing in past studies worked because it provided conservatives with information about the source of the message: when messages aligned with conservative values, Republicans [conservatives] filled in the gaps and simply presumed that the message came from a Republican source’ (Fielding et al., 2020, p. 196).
While we do not have evidence sufficient to accept it, this conjecture does underline the importance of distinguishing the effects of perceived match and source in explaining why moral reframing works.
Fluency is important for judgements in a range of domains, including familiarity (e.g. Whittlesea, 1993; Scott & Dienes, 2008), agency (e.g. Sidarus, Vuorre, & Haggard, 2017), and surprise (e.g. Reisenzein, 2000). Most importantly for us, the perceived fluency with which you process a message can influence how likely you are to hold it true (e.g. Unkelbach, 2007). This is thought to be why repeating a message can make people more likely to believe it.3
Kidwell et al.’s conjecture is therefore coherent. If framing a message in a way that fits a person’s moral psychology can increase the fluency with which they process it, this could explain why moral reframing works.
We have seen that Perceived Match, Source and Fluency provide at least three candidate explanations for why moral reframing works. None rely directly on Moral Foundations Theory, and each avoids the objections considered in The Argument and Some Objections.
The candidate explanations are not exclusive: perhaps moral reframing works by way of multiple distinct processes.
To my knowledge, we lack insufficient evidence to conclude that any the explanations we have considered is correct. Perhaps none of them are. And even if they are all correct, these candidate explanations need not be the whole story. Maybe other processes are also needed to explain the success of moral reframing.
These authors did not find that moral reframing could eliminate the contrast between groups on climate scepticism. (‘The one prominent exception to the moral framing effects observed in the present experiments was on climate change skepticism in Experiment 2. While the common ingroup condition was effective in decreasing skepticism overall, attitudes remained strongly and consistently polarized across conditions’ (Wolsko, 2017, p. 293).) ↩