Overestimating the valuations and preferences of others. Jung, Minah H. Moon, Alice Nelson, Leif D. Journal of Experimental Psychology: General. Nov 2019.
https://psycnet.apa.org/record/2019-69146-001
Abstract: People often make judgments about their own and others’ valuations and preferences. Across 12 studies (N = 17,594), we find a robust bias in these judgments such that people overestimate the valuations and preferences of others. This overestimation arises because, when making predictions about others, people rely on their intuitive core representation of the experience (e.g., is the experience generally positive?) in lieu of a more complex representation that might also include countervailing aspects (e.g., is any of the experience negative?). We first demonstrate that the overestimation bias is pervasive for a wide range of positive (Studies 1–5) and negative experiences (Study 6). Furthermore, the bias is not merely an artifact of how preferences are measured (Study 7). Consistent with judgments based on core representations, the bias significantly reduces when the core representation is uniformly positive (Studies 8A–8B). Such judgments lead to a paradox in how people see others trade off between valuation and utility (Studies 9A–9B). Specifically, relative to themselves, people believe that an identically paying other will get more enjoyment from the same experience, but paradoxically, that an identically enjoying other will pay more for the same experience. Finally, consistent with a core representation explanation, explicitly prompting people to consider the entire distribution of others’ preferences significantly reduced or eliminated the bias (Study 10). These findings suggest that social judgments of others’ preferences are not only largely biased, but they also ignore how others make trade-offs between evaluative metrics.
General Discussion
People are sometimes called upon to assess the preferences of others, assessments which
we find to be prone to persistent biases. Across several studies, we find that across various
measures of valuation and utility (i.e., WTP (willingness to pay), enjoyment, and willingness-to-wait), people believe
that others have more intense experiences than they themselves do (Studies 1-8). We propose
that this overestimation of others stems from a narrow focus on the primary dimension of the
option being evaluated (e.g., a trip to Rio De Janeiro is generally thought to be positive, shaving
your head is generally thought to be negative). But this only involves estimations of others. Selfassessments are further informed by the subtle vagaries of personal preferences, reducing the
total preference intensity (e.g., Rio De Janeiro is encumbered by its hard-to-pronounce local
language, and a shaved head is buoyed by the opportunity it affords for a novel scalp tattoo).
Thus, personal evaluations are more moderate than are the estimates of the evaluations of others.
This intuition is strong enough that it is applied even when the target of comparison is
explicitly similar to the self (Studies 9A-9B). When asked to evaluate someone with an identical
WTP, people think that person will anticipate more enjoyment; and when evaluating someone
with identical anticipated enjoyment, people think that person will have a higher WTP. In
combination, people can demonstrate the paradoxical belief that others are willing to pay more
for the same level of enjoyment (when asked about someone identical in enjoyment) or that
others are willing to pay less for the same level of enjoyment (when asked about someone
identical in WTP). Finally, explicitly prompting people to think about the full distribution of
others’ possible valuations significantly interrupted the intuitive process of overestimation based
on the core representation of objects being considered (Study 10).
Relation to Previous Research
Why do people have such persistent judgmental errors when estimating the evaluations of
others? After all, people are not blind to the evaluations of others. People frequently observe the
choices of others, and at least occasionally, are told something about the preferences which led to
those choices. Research indicates that judgments about values and preferences are often
inherently automatic (Chaiken & Trope, 1999; Kahneman & Frederick, 2002; Kahneman, 2003;
Simmons & Nelson, 2006, 2018; Sloman, 1996). Understanding the trade-offs between
evaluative metrics (e.g., a longer wait versus a higher price), however, is more complicated (e.g.,
Tversky, Sattath, & Slovic, 1988). When reporting their own evaluations, people have the
benefits of each metric being accessible and generally reliable, and consequently, evaluative
trade-offs are more likely to be consistent. When predicting the evaluations of others, on the
other hand, people do not have the same basis of knowledge. Without knowledge of how other
people trade off between evaluative metrics, people appear to ignore them altogether.
Consequently, they use intuitive but incomplete heuristics that people experience things more
intensely, which can be misapplied in the case of similar others (i.e., those who would like a
good as much as they would or would pay as much for a good than they would).
Previous research in judgment and decision-making documents abundant evidence that
people do not always hold stable preferences but construct them on the spot when they are
making decisions (Bettman, Luce, & Payne, 1998; Fischhoff, 2013; Payne, Bettman, & Johnson,
1992; Slovic, 1995). If preferences are partially constructed for the self, they might be entirely
constructed when judging others. Studies 9A-9B demonstrate that while people’s own valuation
of a good remains stable, their beliefs about others’ valuation of the same good reverse
depending on how they are asked about others’ valuation. More specifically, people believed that
others derived simultaneously more and less utility from the same goods than they did.
The overestimation bias we document also offers a new approach to understanding the
endowment effect and why selling prices tend to exceed buying prices. Previous research has
largely focused on a “pain-of-losing” account for this phenomenon, which proposes that people
feel significantly more pain when selling their good than others feel when acquiring the same
good (Kahneman & Tversky, 1979; Thaler, 1980; Tversky & Kahneman, 1991). Another
explanation more recently put forth by Weaver and Frederick (2012) hypothesizes that instead
sellers and buyers use different reference prices. Sellers typically focus on market prices in
determining their selling price, whereas buyers typically focus on their own valuation. Because
market prices tend to be higher than people’s valuations (Kahneman, Knetsch, & Thaler, 1991)
and both parties are averse to bad deals, selling prices tend to exceed buying prices. Our
overestimation bias account suggests that in addition to these explanations, people’s expectation
that others derive more value from goods might also contribute to a discrepancy in buying and
selling prices. In particular, sellers may believe that buyers would value the good more than they
themselves would, leading them to set higher selling prices.
Alternative accounts for the overestimation bias
This paper reports 12 experiments showing the existence, robustness, and consequence of
the overestimation bias. We also conducted a handful of additional investigations to try to
understand the forces that may moderate the expression of our effects. Though these studies do
not authoritatively answer why people overestimate others’ valuation, in combination they may
provide some hints. We review two of those investigations, and report them in further detail in
the Supplemental Materials.
> Others with Extreme Preferences.
Study 9A introduced the matching paradigm as a
strong tactic for controlling how people generate an exemplar when estimating the evaluation of
others. An alternative approach, we thought, might be to simply heighten the salience of some
comparison others who are more or less positive about the same stimulus. If people are
spontaneously thinking of an enthusiastic consumer, then forcing them to consider the behavior
of an unenthusiastic consumer might change their estimate. We examined this exceptional other
account in two additional studies described in detail in the Supplemental Materials.
First, in Study S8 (N = 807), we recruited people who self-identified as having extreme
preferences to investigate whether the overestimation bias would persist. Specifically, we
recruited self-identified fans of Star Wars movies and asked them to estimate either: (a) the
average Star Wars fan’s or (b) the average US person’s evaluations of a Star Wars product.
Though these Star Wars fans rationally understood that the average US person’s evaluation of a
Star Wars product would be less extreme than their own, their overestimation emerged when
considering the average Star Wars fan, assuming that the average Star Wars fan would evaluate
the product more positively than they themselves would.
Second, in Study S9 (N = 1214), we used the match paradigm in Study 9A with an
additional factor (Other). In addition to examining how people view identical others (i.e., those
matched on either enjoyment or WTP), we explored people’s estimations for either: (a) a person
who had greater preference for the product (i.e., would pay $5 more than they would for the
product [Higher WTP Other] or would enjoy the product 5 units more than they would [Higher
Enjoyment Other]), or (b) a person who had lesser preference for the product (i.e., would pay $5
less than they would for the product [Lower WTP Other] or would enjoy the product 5 units less
than they would [Lower Enjoyment Other]). By our reasoning, it is possible that explicitly
considering a less enthusiastic consumer would disrupt people’s intuitions for their preferences,
thereby eliminating overestimation. We first replicated the paradoxical results of Study 9A when
people considered identical others: People assumed both that those matched on enjoyment would
pay more for the product than they would, but also that those matched on WTP would enjoy the
product more than they would. But importantly, people asked to consider lower enjoyment others
(i.e., those who would enjoy the product 5 units less than they would) rationally assumed that
those others would pay less for the product than they would, and people asked to consider lower
WTP others (i.e., those who would pay $5 less than they would for the product) rationally
assumed that those others would enjoy the product less than they would.
Together, the results from these two supplemental studies bolster our finding in Studies
9A-9B that the bias cannot be fully explained by the salience of others with extreme preferences
or the extremity of one’s own preferences. When people explicitly consider others who are less
positive towards a product, people display rational responses. However, when considering
average others or those who should have similar preferences, the overestimation bias persists.
Although people often inaccurately predict others’ preferences, they are more likely to be
accurate about the relative difference between their own and others’ preferences of certain
experiences. For instance, if a parent were asked how much they and others would like their
child’s drawing, they would no doubt recognize that their liking of the drawing would be greater
than that of others. Or if people are explicitly told that someone likes a product less than they
themselves do, this also appears to disrupt the reliance on intuitive core representations. The
results in Study 10 are consistent with this logic: People can more accurately predict others’
preferences when they are explicitly prompted to consider others whose preferences are not
consistent with the core representation of a stimulus. Therefore, people are capable of
understanding others’ preferences, but they do not spontaneously consider and integrate the
entire distribution of possible preferences unless they are explicitly compelled to do so.
The combination of the above points, does highlight an interesting parallel account; one
that we can articulate with some clarity, but one that our present findings can neither perfectly
rule out, nor perfectly rule in.10 In our theorizing, the prospective visitor to Rio De Janeiro
forecasts a positive experience encumbered by a small number of idiosyncratic negative
experiences. That person, when judging others, starts with the core representation of the
experience (i.e., that a visit to Rio De Janeiro is enjoyable), and that confidence in that initial
intuition means that they do not adjust from there. Accordingly, whereas personal assessments of
Rio De Janeiro are somewhat middling, others are perceived to be more positive. The alternative
account focuses not on the mixture of experience within an individual that moves a high rating to
a lower rating, but rather the mixture of experiences across people that produces many positive
evaluations, but also some idiosyncratically individually low ratings for generally positive
stimuli. Consider again the person evaluating the trip to Rio De Janeiro. On average, that person
is probably positive (say, an 85 on a 101-point scale), but some people might be quite negative
(perhaps they are actively avoiding irritating in-laws back in Brazil), and give an extremely low
evaluation of the potential visit. The average person and the negative person both have equal
weight in the overall true average, but they might not have equal weight in how people form their
perceptions of the average other. In essence, it may be the case that people are fully capable of
integrating both the core representation of a prospect with the more unusual negative features;
they accurately recognize that most people think that Ipanema beach is beautiful, and they
accurately recognize that most people are nevertheless bothered by the risk for potential theft,
but they fail to capture that for some people the latter factor is so significant that it overwhelms
the former. That is, they accurately perceive the experiences of others, but they do not consider
all of those experiences when estimating the average experience.
There is merit to this account. First, even for very positive stimuli, there are always a
number of participants whose valuations are quite low. Consider, just as an example, the density
plot for WTP for a movie ticket from Study 1. The mean is not low, but the distribution is hardly
normal, and a sizable fraction of participants (26.4%) say that they are willing to pay $0. Perhaps
it is exactly that segment of the population that people are failing to identify when constructing
their averages. The distributions generated by participants in Study 10 partially challenge the
extreme version of that possibility. Participants generated very accurate representations of other
people’s WTP (albeit less accurate for other people’s enjoyment), but still showed the overall
bias. Still, it may be the case that people are capable of bringing the full and accurate
representation of the distribution to mind, but they do not do so unless prompted.
For now, we remain agnostic. It could be the case that the core representation of a
positive product or experience creates an intuition that biases predictions about everyone
upward, or it may be the case that creates a mental sample that is biased by selecting out those
people who are uncharacteristically negative. Some of our data seems more consistent with one,
but none of it is so consistent to eliminate the possibility of the other. We think that future
research can hopefully untangle those (and we hope that we are the researchers who do so).
Posts
