DOES PAST EXPERIENCE SHARPEN OR BLUNT THE FEELING OF EMPATHY? A REVIEW
When it comes to who to trust to understand your plight, should you select someone with similar experience or someone without?
An exclaimer: I only read the limitation part of the paper under review after I was well done writing my critical review. Upon reading this part of the paper, I realized the authors were fully aware of many of the critical shortcomings I pointed out as well as the constraints these place on their conclusions. This preface is important in case any of the original authors happens to come across it and a few paragraphs in, starts thinking: “but we stated this in the paper”! True. I simply found out too late. And the highly multilayered findings of the study are, unsurprisingly, being reduced to simplistic presentations on social-educational platforms such as LinkedIn. However, there are a few other crucial points I highlighted that weren't captured anywhere in the paper. So, the review is still very much worth reading if you're interested in fine critical details. But more importantly, it is written for people who, skimming through the abstract, might come to the simplistic conclusion that people who have had similar experiences as us would be less empathic toward us. The paper packs infinitely more nuance than that, and I personally came to a different conclusion after my review.
Background
This review was wholly motivated by a LinkedIn post (https://www.linkedin.com/posts/activity-7073947037519523840-EFilutm_source=share&utm_medium=member_android) that has generated an incredible amount of reactions, mostly positive, incredulous, and approving. The poster, Roos van Duijnhoven, a behavioural scientist according to her profile, teased her reader with this arresting scenario:
“Imagine you have just become a new parent. You're overwhelmed and stressed, which shows in your work performance. You decide you want to work from home part-time. Whom will you ask? The supervisor who also had kids early in their career or the supervisor who didn't?”
She then went on to suggest one would be better served selecting the latter, referencing a 2015 study (https://psycnet.apa.org/doiLanding?doi=10.1037%2Fpspi0000012) to back this up. The study reportedly showed that when it comes to empathy, individuals with similar experiences as us are hampered by two things: first, the “inability to recall the impact of emotional distress” associated with the experience, and second, the hubris of a veteran.
The poster acknowledged the “counterintuitive” logic of this study's conclusion. Almost all the commenters remarked this counterintuitiveness (they are not alone as this was also confirmed in one of the five studies that constituted the paper). And so did I. In fact, I was struck very hard by this conclusion that I immediately decided to check out the source material in painstaking details. I was inclined to suspect that the poster either misrepresented the core gist of the study or that something was wrong in the conceptualization and design of the study. The reported findings was asking us to change a fundamental assumption about human kindness; I had to be sure it was based on a solid empirical foundation before kowtowing and propagating this new gospel.
First, let me summarize the basic conclusions I came to after reading the research paper by Ruttan, et al., (2015):
It is true that participants who “previously endured a distressing event” were less likely to take kindly to another person going through the same event.
It is not true that participants who were STILL enduring the distressing event were less empathic. In fact, the reverse was the case.
You see, the key difference boils down to whether the experience was in the remote past, in the immediate past, in the present, or in the future - this is what I call the time continuum factor. In other words, it's about whether the experience has occured, is occuring, or is still anticipated. The researchers did an imperfect but good enough job capturing this time factor. The study is, however, unintentionally misleading in one important sense: when we consider the significant conceptual shortcomings underpinning this paper (as we would shortly), then the primary conclusion of the paper - “that participants who had previously endured an emotionally distressing event more harshly evaluated another person’s failure to endure a similar distressing event” - would become even more complicated than it already was in the study. In other words, one may still be better guided to hold on to the old time heuristic that nudges us to seek affinity with, and understanding from, people who have shared our experiences, but with greater sense of what to look out for - thanks to this paper!
The Studies
Before I delve into this review, it would be very helpful to explicitly reproduce the general hypothesis which serves as the foundation for all the five studies in the paper:
"Compared with participants who never endured the event, we expected that participants who previously endured the event would form less positive evaluations, feel less compassion and more contempt, and be less willing to help an individual who failed to endure a similar distressing event."
The authors further prefaced study 4 specifically with this statement: "we have proposed that enduring an emotionally distressing event can lead to negative evaluations of individuals who fail to endure a similar event", which is the core assumption driving the entire study.
It's very important that you keep the above statement of hypothesis in mind as you plough through what follows.
A General Statement on Design
The paper consisted of five separate but temporally progressive and conceptually connected studies, with each successive study meant to address the empirical limitation of the previous one. It was an artful empirico-logical edifice. Study 1 was a field experiment in which the researchers manipulated the variable of interest based on existing conditions in the experimental setting. There was no true random assignment or a true control group. Study 2 was a true experiment in that there was both laboratory manipulation of the independent variable, random assignment of participants to groups, and the inclusion of a control group (it was the only true experiment and I thought its potential to unknot some knotty aspect of the research question was not sufficiently milked). Study 3 and 4 shared the same design features with study one. Study 5 was more of a systematic survey study with post hoc sorting into two groups. This is, however, not a critique of methodology.
Going back to the statement of hypothesis quoted above, there are two fundamental problems in the conceptualization and operationalization of one of the theoretical constructs - experience - underlying the body of hypotheses from study 1 through 4:
First, is the problem of a failure to differentiate between types of experience (whether complex or simple, performance-based or experiential, and whether continuous or terminal), and
Second, is the problem of a failure to differentiate between dimensions of experience (whether at the objective or at the subjective level).
I believe these are the reasons behind the authors' observation of a disagreement between “past research” and their own with regards to the claim that "prior experience increases sympathy" (and the same criticism could be applied to those previous studies as well). It is an artificial and needless disagreement that could easily be reconciled with a reworked conceptual framework. These two fundamental conceptual shortcomings underwrote the ensuing counterintuitive and head-scratching conclusion the authors came to: a conclusion that is true only with enough caveats that had to be iterated in a complex sequence of five studies when one or two could have been enough.
Study 1
The first instalment in the series of studies started with this stated goal:
to examine participants’ emotional reactions toward an individual who failed to complete a polar plunge—an event that requires participants to enter an icy body of water— either before (never-endured condition) or after (previously endured condition) they completed a plunge themselves.
However, what this study, in my opinion, succeeded in demonstrating was not strictly the effect of experiential familiarity on empathy (as intended), but rather the differential effect of experiential success or failure at a specific task on the associated feeling of self-efficacy (a sense of competence emboldens us to be more critical of others). As a matter of fact, the causal variable in study one has the defect of not qualifying to be regarded as an “experience" in the strictest sense of the term, but a task or an exercise. This is no doubt a technical point, but a highly relevant one in the broader real world context where lessons of this studies are expected to be applied in decision-making.
There's a fundamental difference between an experience that is performance-salient and one that is emotion-salient. In the former, it is immediately clear what constitutes success or failure, and the burden of performance and outcome is almost entirely on the subject. This isn't true of the latter type of experience an example of which is loss of a close relative. In the real world, people seek empathy less for the former kind of experience than for the latter type. These limitations were, however, later addressed by study 4 (more on it later).
A big clarification: The separation of experiences into performance and emotion salience should not be taken as a clean and clear dichotomy, but as a relative composition in which one predominates over the other.
The participants in study 1 who were less judgmental towards the fictional character who failed at the same task did so because they exaggerated the perceived difficulty of the task owing to the fact that they've not successfully been through it themselves. In other words, they haven't yet had the opportunity to boost their sense of self-efficacy, and thus probably operating at baseline levels of self-esteem. Hence, a sense of reluctance and a feeling of unworthiness to judge another person who failed at the same task, keeping in mind that they too may fail at it. They had no past record of personal success to reference yet and we're thus more likely to exaggerate the perceived difficulty of the task. (This logic would later be explored and confirmed in study 3 where the authors concluded that: "Participants who previously endured unemployment were less compassionate because they thought unemployment was less difficult to overcome.") In other words, personal success makes you overconfident of your own ability and less considerate of failure by others (while personal failure achieves the opposite).
Also, I'd argue, contrary to what the authors thought, that the group in study 1 that “never endured” the ordeal of a icy plunge, were experientially closer to the fictional character being judged than the group that successfully endured the experience. This is because the former group's experience of never previously or successfully taking a plunge more closely matched the psychological condition of someone who did and failed than of someone who did and succeeded: there is a shorter psychological distance between have-failed and might-fail than between have-failed and have-succeeded, such that the experience of those who have succeeded at a task, at the subjective level, is actually farther from that of the person with a record of failure at the same task. (Again, the failure to account for the subjective-objective dimension of experience throughout the five studies is the second fundamental flaw of the paper in general.) Similarity of objective experiences is neither a necessary nor sufficient criterion for similarity of subjective reality. And when it comes to empathy, similarity of subjective experiences is more key.
Again, the authors appeared to have confused the simplistic and linear nature of the performance-salient ‘experience’ that study 1 examined with the more complex and messy emotion-salient experiences such as grief, trauma, bullying, whose effects are, in reality, multilayered and continuous. Although, they seemed to adequately correct for this in study 3 and 4 which looked at the experiences of unemployment and bullying respectively. And If there's one thing that studies 2, 3, & 4 interestingly went on to demonstrate quite robustly and consistently, it was that an ongoing experiencing of an event is the most conducive to empathy and compassion. The continuity-in-time of the salient experience (whether performance or emotion) in terms of occurrence, effect, and impact, rather than the mere categorical fact of "previous experiencing" is what is critical in empathic resonance and compassionate response. This was confirmed in study 3 which incorporated the control variable of "length of time of unemployment", leading the authors to observe that "with longer times since last employment predicting increased compassion for the individual." Meaning, the longer the participants' experience of unemployment, the stronger their feeling of compassion for someone going through similar experience, with those still in the middle of the experience showing the strongest feeling of compassion. To put this in words that further mirror the point I'm trying to make: participants who were most unsuccessful in quickly resolving their unemployment situation were more sympathetic compared to those who were quicker in getting themselves out it. Study 4 beautifully demonstrated this subplot about success, continuity and complexity (of experience), as it focused on bullying as the variable of causal interest. But more on this later.
It'd have been nice if study 1 made a distinction between those who successfully took the plunge and those who tried but failed, but it didn't. Actually, the researchers also didn't take into consideration the fact that those who had tried in the past and failed could not be captured by the field study precisely because they were probably least likely to show interest in the same activity in future and, therefore, not likely to show up at the centre where participants for study 1 were enlisted. This fact wasn't captured either in study 2 (which was done in a controlled setting), but slightly accounted for in study 3 (but the two experiences, in study 1&3, are not equivalent in terms of their most salient property).
An aside: When we go through a largely performance-salient experience (say writing an exam or learning to drive) rather than a prolonged embedded life situation with strong emotion salience, it makes sense that we feel a sense of injustice when placed side-by-side with someone trying to avoid or circumvent the same process, because we rightly feel that no one subsidized the experience for us even though we would have loved that very much. The feeling of contempt and annoyance would therefore naturally be aroused, engendering psychological distance rather than affinity. This probably accounted for the relatively bigger effect sizes in studies 1&2 compared to others because they're also the strongest in performance salience.
Finally on study 1, we need to keep in mind that the difference in mean compassion between the two experimental groups was by no means absolute though statistically significant. There were those in the “previously taken a plunge” condition that were indeed still very compassionate and those in the “never taken a plunge” condition who were less compassionate. This is okay and should be expected in every quantitative study that is looking at aggregated mean. But in this particular study, the small (but just significant) mean difference can be accounted for in terms of other variables that are critical in empathic response but that weren’t captured in the study, e.g emotional intelligence and self-esteem. (More on these later, as well.)
Study 2
Again, the only experimental study among the whole lot failed to actually match the true psychological condition of the fictional character (who attempted but failed to complete a strenuous memory test) being evaluated by the study participants (who took the same test). Instead, participants were either those who had previously completed it, just completed it, or have never completed it. No group to represent those who attempted but failed to complete the task. Study 2 also had the same conceptual shortcoming as study 1 of being a performance-salient experience, which should be more accurately regarded as a task/exercise/aptitude rather than an experience with complex emotional salience.
I don't understand the logic behind the researchers' prediction that those who endured the pain of the mental exercise a week ago would be less inclined to look favourably on the failure of the fictional character than those who had just recently experienced the mental pain. The implied logic does not seem consistent with the authors' central claim that previous experience reduces the capacity for leniency in evaluation of others going through similar experience (although this lapse was indirectly accounted for in study 3 through the introduction of the moderating variable of duration). Hence, in the spirit of logical continuity, one would expect that those who never had to endure the test should be more generous in their rating of the fictional character's competence (one of the two DVs in study 2) than those who previously did and those who recently did (since prior exposure to similar experience fosters callousness). However, while this was true with respect to those who previously endured the task, it was exactly the opposite with respect to those who recently did. Those who recently did out-nice both those who never did and those who previously did. And the margin between the competence ratings of the group that recently took the test and the group that never did is larger (>7) than that between the group that never did and the one that previously did (< 6). This pattern was again repeated in the other DV measure (willingness to rehire). This is not in any way consistent with the core assumption of the paper: that we are more likely to judge others harshly and less sensitive to their plight If we had “previously endured the same event” as them. But the authors summarily declared that "the results of study 2 support our prediction." They wrote:
"Participants in the previously endured condition were less willing to rehire and provided less favorable evaluations of the employee who struggled with the test than did those in the never-endured and currently enduring conditions."
True, but they failed to take this logic further in that, if their hypothesis implied an increasing tendency towards less consideration in judgement as experiential familiarity with task/event increases, then the “currently enduring group” should be next in line in terms of less favourable evaluation. But this was not the case at all. Rather the next in line for harshness of judgment was the group unfamiliar with the task ahead of the currently familiar group. The closest the authors came to cognizing and explaining this inconsistency was this hypothesis-pushing statement which would be better to quote in its entirety:
"Participants in the previously endured condition were less willing to rehire and provided less favorable evaluations of the employee who struggled with the test than did those in the never-endured and currently enduring conditions. It is important that this difference persisted when including the currently enduring condition because it rules out an information account for this effect; participants in the currently enduring condition completed the same test as did the participants who previously endured the test (and thus possessed similar knowledge about the test), but only the former made their evaluations while actively experiencing the associated mental fatigue."
They were, however, completely and curiously silent about the reverse finding in the evaluative attitude of the currently enduring condition compared with the never endured condition.
The question then is why did the authors fail to notice or address this contra result embedded within one that's pro their hypothesis? Perhaps they were blindsided by the excitement of having their hypotheses apparently confirmed across all studies so much that they failed to pick up the one conspicuous exception that undermines the entire edifice.
As to why this loophole in hypothesis testing occured at all in their second study, my guess is that the second group (the one in which participants just freshly completed the memory task) combined the best qualities of the other two conditions: completion and recency. Like the group that completed the test one week ago, they were fully aware of the difficulty at a visceral and phenomenological level, but unlike that group, this experience hasn't faded yet in memory and thus less susceptible to recall bias where one is likely to appreciate less how really difficult the task is (hence judging the failed employee more harshly) and to increasingly see their own performance in a more positive light forgetting how much they also struggled (also leading to a harsher judgment stance). Interestingly, this was the basis for the third study. In fact, I was surprised to see a shocking similarity between my own attempt above to account for this part of their result in study 2 that didn't hold up to logical scrutiny and the authors' reasoning for their hypothesis in study 3. Here is study 3’s statement of hypothesis:
"We predict that the combination of a constrained memory for the impact of past emotional distress and knowledge of one’s own completion of the event makes the event seem easier to overcome, which, in turn, makes struggling with the event appear more blameworthy."
However, this prediction is already unintentionally anticipated and confirmed in their second study, but in a way that went against the grain of their core assumption.
Study 3 and 4
Study 3 and 4 seemed the most attuned to testing the core assumption driving the entire study. But there's one tiny but crucial difference between the groups in study 2 and 3: whereas in study 2, the group designated as "currently enduring" actually completed the endurance memory task, but in study 3, the "currently enduring" group was still immersed in the experience that constituted the distressing event. This is partly as a result of the difference in the setting: laboratory vs field, as well as in event: task vs experience. These may seem minor distinctions, but the ramifications, as already explained, are huge.
Also, it's my opinion that the moderating variable - 'perceived difficulty' - would have been better suited to study 2 than 3, because it is somewhat objectively less meaningful to measure the perceived difficulty of a situation as compared to a task. The notion of 'difficulty' itself presupposes as well as connotes a performative experience as well as some level of control over outcome, neither of which is salient enough in the condition (unemployment) that study 3 manipulated naturally. (Of course, this would make sense as a subjective measure, but the study design was not sensitive to this distinction, as already stated.) On the contrary, it'd have been very apposite to talk about “perceived difficulty” in a memory task which was the artificially manipulated variable in study 2. Hence, it'd have been more meaningful to measure the 'severity' of the event in study 3 rather than its 'perceived difficulty'. I know that this problem was perhaps partly sidestepped by the subjective nature of the measured difficulty, but the nature of the referent also considerably widens the subjectivity variance reported by each participant. This could have been significantly minimized by a more objective and unified experience that study 2 explored.
So far, we must note that study 1 & 2 were more of mental than emotional experience contrary to the authors' stated belief that they were inducing "emotionally distressing event". As already stated, they are both high in performance salience. Study 3 and 4 were closer to actually inducing or, more accurately, exploiting a preexisting emotionally distressing experience, most especially study 4 which, interestingly and unsurprisingly, also happened to score some major exceptions to the general trend of results observed in the studies that preceded it.
Study 4, which chose bullying as its IV, was the most emotionally salient variable naturally manipulated in all the four studies. As a result, and far more than the other three studies before it, it also adequately captured the property of continuity of effect of the event "previously endured". This was reflected in the reported outcomes in two ways: First, it is the only study in which a demographic variable had a significant effect: "gender was a significant predictor of compassion, with females reporting higher levels of compassion for the bullied student." [It could be argued that one of the reasons why gender effect wasn't seen in any of the other 3 studies was because of the insufficiency of the chosen IVs in emotional salience.]
Compared to other three preceding studies, study 4 also recorded the lowest mean differences between or among the various experimental groups in each study. In fact, if the criterion for significance was raised to 0.1 most of the hypotheses in study 4 wouldn't pass muster, but majority of those in other three studies would still manage to. What does this mean? It means that if the two already highlighted conceptual shortcomings at the core of the theoretical framework of this study were addressed, all the other three studies would be more similar to study 4 in effect sizes. And it'd be easily demonstrated that, in fact, existing literatures and theories that posit that past experiences enhance the feeling of compassion actually still holds true, even when someone failed to endure.
Summary
What the authors demonstrated through their studies, as conceptualized and operationalized, is that SUCCESSFULLY enduring (or resolving) difficult past experiences or tasks hardens the subject's mind against leniency in evaluating someone else who FAILED at the task/experience (for reasons the authors already did well to speculate on). This is true enough as the study more than adequately demonstrated. But it shouldn't be generalized into the broad claim that past experiences means less compassion because the study was not conceptually nuanced and experimentally sensitive enough to make this broader claim (as Roos van Duijnhoven innocently did). And the reason this is so is already explained in terms of the study's failure to make two key conceptual distinctions, that is; failure to differentiate types of experiences (performance vs emotional salience) and failure to differentiate dimensions of experience (objective vs subjective). While the authors' case rest on the caveat: fictional characters “who failed to endure”, the other side of this relationship (a character who endured successfully) also merit equal foregrounding. Study 4, however, did well to test this other side of the equation.
This shows that the phenomenon under study operates in a seesaw manner rather than as a one-way swing. If people who had previously endured were harsh in judging those who failed to endure, it's not just because they had previously endured the ordeal, but more precisely because they endured it 'successfully’. This relationship will reverse if they didn't endure successfully and the person they're judging managed to endure successfully. This point is very critical because without it, the paper's core findings would be very misleading.
Factors of Empathy
So, what does this elegant paper actually reveal about empathy and its manifestation?
Whether you're currently going through similar experience or whether you have already gone through it and how long ago you did - temporal salience - matters a great deal in your empathic response. As a matter of fact, I think this is the biggest difference maker in empathic response that the study demonstrated without realizing it. They took a dichotomous or trichotomous view of time instead of operationalizing it as a continuum. The longer the distance in time you had the experience, the less likely you are to recall the difficulty and distress associated with it and more likely to exaggerate how well you handled the situation in your own case. But there are other factors;
Whether you failed in the face of that experience and/or you had to be bailed out or assisted: those who did are more likely to be more sympathetic and less self-congratulatory. So, using van Duijnhoven's example, what matters most isn't whether the boss also had a baby early in their career, but how recently they did and whether they handled it badly and/or received help and the amount of help they got. If they did, you'd still better be served by choosing them over the ‘inexperienced' boss.
Endurance versus entitlement: this is an interesting one with some deep theological flavour. Study 4 is perhaps the most nuanced and intriguing. People admire those who suffer and bear their suffering valiantly. On the other hand, they tend to despise those who complain, bail out, retaliate, or demand to be assisted. In other words, even the empirically unsympathetic group (those who endured similar experience) was equally inclined to sympathize with those who quietly bore their bullying ordeal. I think we all do. The question is why?
The emotional salience of the experience and its continuity also matter. Actually, if the emotional salience is strong enough, continuity of impact is guaranteed. This principle means that the experience is less likely to quickly fade from memory and thus guarding against the error of biased recall. So, the boss whose experience with childbirth and maternity care was less harrowing is more likely to have had an easier time navigating this period and to thus think it was not that difficult.
None of the control or matched groups in all the studies actually strictly fits the descriptor: “never endured”. The only exceptions may be studies 3 & 4 where the control groups were not even exposed to the experience being examined at all (but the possibility for vicarious experience must be high). In studies 1 & 2, they were exposed to the experience but didn't go through it. And that little exposure is key in kindling empathic response. In the real world, those who have never gone through an experience are even more likely to underestimate the difficulty and how much help or misjudge the type of help you need. Depending on their starting bias, they may easily invoke examples of people who had the same experience and did well or people who did terribly. With such boss, your odds of getting sympathy is about 50-50, but your odds are slightly higher with one who has similar experience. However, if you chose the wrong type of experienced boss, your luck would be worse than if you chose the inexperienced one.
And, lastly, traits such as emotional intelligence, self-esteem, critical cognitive orientation, metacognitive ability, and religiousity, are likely to combine in ways that trump such variables as whether you have had similar experience or not. This would account for why someone who never had similar experience could be more considerate and someone who had similar experience could be less considerate. For individuals high in emotional intelligence, self-esteem and philosophical reasoning, past experience wouldn't matter in their capacity to feel for others and gauge the amount and breath of support they'd need. This last factor is especially important with regard to the authors’last paragraph in their discussion of the sociopolitical implication of their study, where they asked and commented on the question: “How do people from marginalized groups respond to members of their community once they have achieved success?”
Factors 1&3 were directly tested and confirmed in the study, while factor 4 was indirectly confirmed. Factors 2&5 were not examined, while factor 6 transcends the scope of the study and wasn't looked into at all.