Empathy in society: not in the deep downward spiral we thought it was?
“I saw all the mirrors on earth and none of them reflected me” – Jorge Luis Borges
There is a concern, an idea, that in the way we live our lives today, we’re losing our empathic abilities, that our societies are becoming less empathetic.
Part of the reason this strand of thought arises is because we increasingly live our lives online. And there is a concern that the anonymity and distance that online interactions allow may make it easier to ignore others’ feelings; that the nature of social media platforms may encourage narcissism to flourish; and that online interactions may not allow signals that are important in understanding and honing empathy.
But what do we even mean when we talk about empathy and what do we know about its fall or rise in societies?
When psychologists talk about empathy, they generally divide the construct into distinct components, cognitive and affective.
The cognitive aspect refers to how well an individual can perceive and understand the emotions of another. And the affective aspect refers to the capacity to respond with an appropriate emotion to another’s mental state; it requires feeling the same emotion as another person.
While there are some neuroscientific and physiological methods, self-reports are by a distance the most common way empathy is measured. And the most commonly used self-reporting measurement scale is the Interpersonal Reactivity Index (IRI). This asks respondents to rate where they are on an agree-disagree scale with statements like: ‘I sometimes try to understand my friends better by imagining how things look from their perspective’ (cognitive empathy); ‘I often have tender, concerned feelings for people less fortunate than me’ (affective empathy).
So what do we know about the state of empathy across different societies or over time? Well, not very much. At least part of the reason is a measurement related constraint. The nine self-reporting empathy measurement scales we observe being used most commonly are all quite long (ranging from 16 to 64 questions). Unlike a concept like interpersonal trust (often measured through a single question), repeated or large-scale empathy measurements would be time-consuming, costly exercises.
So when The Washington Post, Bloomberg, the American Psychological Association, CBC, Zurich and a range of others talk about the decline in empathy what is this based on? A single study, a compelling meta-analysis by the leading empathy researcher Sara Konrath and her colleagues at Michigan. This covers 72 samples of American college students who completed dispositional empathy tests using the same IRI scale between 1979 and 2009. Now what is happening to empathy in American college students may not necessarily reflect what is happening in wider American society (and certainly doesn’t tell us what is happening in other places around the world), but on a concept where measurement and therefore data is sparse this approach offers a glimpse of how things might be changing. What Konrath and her co-authors find is that both empathic concern (reflecting the emotional component of empathy) and perspective taking (reflecting the cognitive component of empathy) declined sharply (and consistently) in American college students from 1979 to 2009.
This made us curious about what had happened to empathy in American college students in the 2010s, especially since this last decade has seen a rapid increase in the amount of time people spend living online; so we extended the meta-analysis through a rapid review.
What we find is a drastic recent reversal of the downward trend observed over the previous decades in both empathic concern and perspective taking. For both aspects of empathy, scores for American college students over this last decade are at or close to the highest they have been in the forty years this extended meta-analysis covers.
What does this mean about the relationship between social media and empathy? Well we can’t confidently say because none of this is a causal exercise which connects the two.
We started though with a number of ways social media might have a negative relationship with empathy. How might it instead positively affect empathy? Well social media might be beneficial in the psychosocial development of adolescents; where social media encourages interpersonal depth in exchanges it might help build empathy and empathic capabilities; online technology can facilitate purely positive (e.g. charitable giving) or other prosocial behaviour with empathy at its root because of the reach it provides; and, social media might enhance sociability when deep offline engagement is otherwise difficult to attain.
A final caveat on all this. Recent meta-analytic results show that self-reported cognitive empathy scores account for just 1% of the variance in behavioral cognitive empathy assessments, indicating people may not have the insight needed to validly gauge their own empathic abilities (or at least their own cognitive empathic abilities). Perhaps it is time to rethink the very notion of how empathy is measured.
Note: We updated the Konrath et al study through a rapid review of papers with empirical results that would broadly match the inclusion criteria used in the original study. We identified papers where empathy was measured: in the decade starting 2010, for US college students, using the Empathic Concern or Perspective Taking subscales of the IRI.
We searched Google Scholar for articles published in 2010-2020 using two search methods: articles that cited the original Konrath et al (2011) paper, and using the search terms “empathic concern”, “interpersonal reactivity index”, US/ America + students. In both cases we limited our search to the first 150 papers identified (so a total of 300 studies). We identified 18 studies with relevant results that we included.
We caution that this is a rapid review and not a comprehensive study.