Social Perception of Forecasters: People See Forecasts of Future Outcomes as Cues to Forecasters’ Desires,Attitudes,and Identity

Abstract

While people’s forecasts of future outcomes are often guided by their preferences (“desirability bias”), it has not been explored yet whether people infer others’ preferences from their forecasts. Across 3 experiments and overall 30 judgments, forecasters who thought that a particular future outcome was likely (vs. unlikely) were perceived as having a stronger preference for this outcome. Individuals were more likely to infer preferences from forecasts in the presence of cues facilitating internal attributions and in case of outcomes characterized by an actual positive empirical association between desirability and likelihood judgments. Finally, making future forecasts inconsistent (vs. consistent) with one’s stated preferences made observers doubt forecasters’ expressed preferences and identity. Overall, these findings suggest that social observers tend to interpret future forecasts as cues to others’ identity, values, and attitudes.

Keywords

forecasts desirability bias person perception social inferences lay dispositionism

Every day, millions of people make forecasts about the future, including stock prices, fashion trends, outcomes of sporting events, and political elections. Sometimes, people’s forecasts just reflect their desires (Krizan & Windschitl, 2009). Other times, they do not. For example, people often make predictions that are inconsistent with their preferred outcomes out of defensive pessimism (Norem & Cantor, 1986) to avoid tempting fate (Risen & Gilovich, 2008) or in a hope that their forecasts will affect the involved actors’ behavior and prevent the unwanted outcome. Are social observers sensitive to these forecasting motivations or do they believe that others’ forecasts are merely a reflection of their preferences? The present research was designed to answer this question. Specifically, three studies examined whether future forecasts serve as a basis for inferences about forecasters’ desires, preferences, and, ultimately, identity.

Preference–Expectation Link in Social Inferences

People tend to see positive or desirable events as more likely than negative or undesirable events, a phenomenon referred to as the desirability bias (or preference–expectation link; Krizan & Windschitl, 2007; Weinstein, 1980). For example, people overestimate the likelihood that positive, rather than negative, events will happen to them (Lench & Ditto, 2008; Weinstein, 1980). Individuals’ preferences for a certain event have also been shown to affect their judgment of this event’s likelihood in the domain of politics and sports—the finding that goes back to the 1932 presidential election, in which the majority of Roosevelt (vs. Hoover) supporters believed Roosevelt (vs. Hoover) would win the election (Hayes, 1936). Since then, the desirability bias has been documented with respect to different social, political, and sporting events (Krizan, Miller, & Johar, 2010; Massey, Simmons, & Armor, 2011; Simmons & Massey, 2012).

As people’s predictions of future outcomes seem to reflect their preferences, do people infer others’ preferences from their predictions? People routinely engage in inferring others’ mental states—intentions, goals, and beliefs—from their behavior and do so spontaneously and automatically (Hassin, Bargh, & Uleman, 2002; Malle & Holbrook, 2012; Uleman, Saribay, & Gonzalez, 2008; Van Overwalle, Van Duynslaeger, Coomans, & Timmermans, 2012). In making social inferences, people often tend to overestimate internal factors—attributes of the individual—and discount external factors—attributes of the situation (i.e., lay dispositionism and fundamental attribution error; Ross, 1977; Ross & Nisbett, 1991). For example, people are more likely to explain criminal activities by perpetrators’ bad character than by situational constraints (Kubota et al., 2014) and even tend to attribute actors’ behavior in movies to their personality (Tal-Or & Papirman, 2007).

At the same time, even when social observers on average overestimate the role of internal factors in explaining others’ behavior, they rarely ignore situational forces altogether (Reeder, Monroe, & Pryor, 2008). Similarly, people are less likely to make internal attributions when cues to internal factors are weak or ambiguous. For example, a line of studies on the fundamental attribution error showed that an attitudinal essay did not serve as a basis for inferences about the author’s underlying attitudes when it was of poor quality, as a poor quality essay represents a rather weak cue to its author’s actually held attitude (Jones, Worchel, Goethals, & Grumet, 1971; Miller & Rorer, 1982).

In the present research, I propose that people’s proneness to consider outcomes’ desirability when judging their likelihood and tendency to discount situational constraints when explaining others’ behavior might shape their interpretation of other people’s future forecasts. Specifically, individuals can preferentially attribute others’ predictions of future outcomes to internal factors, such as their desires and preferences. As a result, they might end up drawing conclusions about targets’ preferences for future outcomes based on targets’ likelihood judgment of these outcomes. For example, a person who makes an optimistic election forecast for a certain party might be seen as having a stronger preference for this party than a person who makes a more pessimistic forecast.

These suggestions were tested in three experiments. Experiment 1 examined whether a target’s likelihood judgment of a new data protection law being enacted can serve as a basis for inferences about his or her support for this law and political identity. Experiment 2 tested the generalizability of this effect across forecasts of economic, political, and social trends as well as personal events, and explored its boundary conditions. Experiment 3 examined whether group members who make unfavorable (vs. favorable) predictions regarding their group’s future are attributed the respective preferences and are perceived as having a weaker group identity and commitment. Study materials and data of all three experiments are publicly available at the project’s Open Science Framework page: https://osf.io/vsj82/?view_only=14af4aca1e8d49339313aea553bf65c6

Study 1

Study 1 explored whether social observers tend to make inferences about people’s preferences for future outcomes based on their likelihood judgment of these outcomes in the context of the Referendum on the new “Big Brother surveillance law” in the Netherlands. The referendum was set to take place in mid-March 2018, and the study was conducted several months before this date. Participants read about a political forecaster who predicted that the new law will versus will not enter into force and were asked to estimate the forecaster’s support for the new law and political ideology. I expected that a forecaster who made an optimistic (vs. pessimistic) prediction about the law’s passing chances would be attributed a stronger support for this law. In addition, consistent with the differences between left-wing and right-wing ideologies in the support of surveillance policies (Cohrs, Kielmann, Maes, & Moschner, 2005), a more optimistic (vs. pessimistic) forecaster will be perceived as espousing a more right-wing ideology.

Method

Participants

Participants were first-year psychology students at a large Dutch university who participated in the study for course credits. The sample size was determined in advance by giving potential participants a 2-week period to fill in the survey. Based on the subject pool size, I expected to recruit at least 300 participants in 2 weeks, resulting in 80% power to detect a small (e.g., d = 0.30) effect (here and throughout the article: independent sample t test, α = .05, two-tailed). Three hundred thirty nine participants completed the study. Forty-three did not pass an attention check question (see below) and were removed. The final sample consisted of 296 individuals (M_age = 19.92, SD_age = 2.20, 20.9% male).

Procedure

Participants read a brief paragraph introducing the Referendum on the new “Big Brother surveillance law.” The planned law will grant security services the power to more closely monitor private individuals’ online behavior. Participants learned that if the law is rejected by more than half of the votes cast, its entry into force might be suspended. Next, participants read about Paul van den Bos, a legal expert and political forecaster. In the “negative forecast” condition, participants learned that Paul “expected more than 50% of the votes to reject the Big Brother surveillance law, so that it will not enter into force.” On the opposite, in the “positive forecast” condition, participants learned that Paul “expected less than 50% of the votes to reject the Big Brother surveillance law, so that it will enter into force.” Next, participants indicated whether they thought that Paul himself supported or opposed the new legislation (1 = opposed, 9 = supported), wanted the law to enter into force or to be suspended (1 = be suspended, 9 = enter into force), and how he would vote himself (1 = against the law, 9 = in favor of the law). Participants’ responses to these three questions were combined into a scale of perceived law preference (Cronbach’s α = .91). Afterward, participants indicated Paul’s perceived political ideology (1 = left, 9 = right). At the end, participants stated whether they themselves were in favor or against the law (1 = against the law, 9 = in favor of the law) and whether they planned to participate in the referendum (1 = very unlikely, 9 = very likely). As an attention check, participants indicated whether Paul predicted more or less than 50% of the votes cast to reject the law and responded to sociodemographic questions.

Results and Discussion

Means, standard deviations (SDs), and correlations among all variables are shown in Supplemental Table S1. Participants believed that a political expert who predicted that the new law would enter into force had a stronger preference for this law (M = 5.76, SD = 1.81) than a political expert who predicted that the new law would be suspended (M = 3.64, SD = 1.48), t (294) = 11.10, p < .001, d = 1.30.

The forecast also served as a basis for inferences about the forecaster’s political ideology in general. The expert who was optimistic about the surveillance law’s prospects was seen as tending toward the right-wing end of the ideological continuum (M = 5.39, SD = 1.94) than the expert who expressed less optimism regarding this law’s prospects (M = 4.57, SD = 1.66), t (294) = 3.92, p < .001, d = 0.46.

Participants’ own attitudes toward the law and their willingness to take part in the referendum were not affected by the experimental manipulation (all ps < .32).

To summarize, a political forecaster who made an optimistic (vs. pessimistic) prediction of a new law’s passing chances was perceived as supporting this law more and endorsing the respective political ideology. Hence, these results provide first evidence of social observers’ tendency to make inferences about people’s preferences based on the information about their future forecasts.

Study 2

Study 1 has shown that people tend to infer others’ political preferences and ideology from their likelihood judgment of future political events. Study 2 sought to extend this effect to a broader range of forecasts and explore its variability across the forecasts of different outcomes. I examined two potential sources of this hypothesized between-outcomes variability.

First, I assumed the tendency to infer others’ preferences from their forecasts to be at least partially grounded in individuals’ own tendency to see desirable events as more likely to occur than undesirable events. Therefore, I expected the forecast–preference link in the perception of others to be reflected in the empirical forecast–preference association, that is, in actual, empirical associations between individuals’ judgments of outcomes’ likelihood and desirability. To examine this possibility, I tested whether the empirical forecast–preference association with respect to an outcome (i.e., the association between participants’ judgments of desirability and likelihood of an outcome) moderated the effect of a target’s forecast of this outcome on participants’ perception of this target’s preference for this outcome.

Second, previous research has shown that people’s tendency to make internal attributions collapses when cues to internal factors are weak (Miller & Rorer, 1982). Therefore, I explored whether individuals are more likely to infer preferences from forecasts, when the information about a target’s forecast represents a strong (vs. weak) cue to his or her preferences. Future events and outcomes naturally differ in the degree to which they allow such inferences. Some outcomes are clearly desirable (e.g., finding a cure for cancer) or clearly undesirable (e.g., natural catastrophes), whereas others are of mixed desirability, that is, desirable for some people but not for others (e.g., success of a certain political party or future fashion trends). I expected participants to be especially likely to infer forecasters’ preferences from their predictions in case of mixed desirability outcomes (e.g., outcomes of political elections), due to a great deal of ambiguity with respect to whether such outcomes are desirable for any particular forecaster or not. Under such circumstances, the information about a target’s forecast might represent a cue to this target’s preferences. In contrast, I expected the preference attribution effect to be weaker or even vanish completely for forecasts of unambiguously desirable or undesirable outcomes (e.g., finding a cure for cancer), as such outcomes imply very little ambiguity with respect to whether any particular forecaster finds them desirable or not, turning the information about his or her forecast into a rather weak cue to his or her preferences.

Method

Participants

Participants were recruited on Amazon Mechanical Turk (MTurk). Study 1 yielded a large (d = 1.30) effect size. However, as I expected this effect to get smaller for some (e.g., highly desirable or undesirable) outcomes, to be able to detect smaller effects (e.g., d = 0.40) with 80% power, 201 individuals were recruited for this study. Three failed an attention check question (that requested them to select a particular answer instead of answering the question), resulting in a final sample of 198 individuals (M_age = 37.73, SD_age = 13.58, 52% male).

Procedure

Participants read brief statements about 28 hypothetical peoples’ predictions about the future.¹ Half of the participants read that the target thought that an event will happen, and the other half read that the target did not think that an event will happen. For example, “Jessica thinks (vs. does not think) that it will rain tomorrow” or “James thinks (vs. does not think) that scientists will find a cure for cancer soon.” Different names were used for each of 28 forecasters (the complete list of predictions is provided in Supplementary Materials). Whether the target predicted an event or not was manipulated between subjects. That is, for each participant, every forecaster was described as either thinking that the event in question will (vs. will not) happen. As a measure of perceived outcome preference, participants estimated whether the target wanted the respective event to happen or not (1 = not at all to 9 = a lot).

Participants were additionally asked whether they themselves thought each of the 28 events will happen or not (likelihood judgment) and whether they wanted the respective events to happen (desirability judgment). Both questions were answered on a 9-point scale (1 = not at all, 9 = a lot). Participants’ desirability judgments were averaged for each outcome and used as an indicator of outcome desirability: The higher the aggregate score, the more desirable the respective outcome was considered on average. To measure the empirical forecast–preference association, I computed a correlation between participants’ desirability and likelihood judgment of each outcome. The correlation ranged between r = −.17, p < .05, and r = .72, p < .001 (average r = .29). The higher the value of the correlation, the stronger the empirical forecast–preference association for each particular outcome.

The order in which participants judged the targets versus indicated their own preferences and likelihood judgments was counterbalanced. As the order did not interact with the experimental condition (b = 0.30, p = .14), I’m not considering it in the main analyses.

Results

On average across the outcomes, targets who considered a particular outcome as likely were attributed a stronger preference for this outcome (M = 4.92, SD = 2.53) compared to targets who thought this outcome to be unlikely (M = 3.99, SD = 2.29). The difference between the conditions reached 0.93 points on average (d = .52, 95% confidence interval, CI: [.23, .80]), ranging between −0.51 (d = -0.24, 95% CI [−.52, .05]) and 2.45 (d = 1.36, 95% CI [1.05, 1.67]) points, depending on the outcome (see Figure 1).

Figure 1.

Effect of targets’ forecast on targets’ perceived preferences for predicted outcomes, Study 2.

To account for the random sampling of both participants and outcomes, I examined the effect of forecast on perceived preference in a multilevel regression analysis (Judd, Westfall, & Kenny, 2012). The model included a random intercept at the level of participants and outcomes. The experimental condition was effect-coded (1 = predicted, −1 = did not predict) and modeled as random at the level of outcomes. Supporting the descriptive results reported above, these analyses showed that targets who considered a particular outcome as likely were perceived as having a stronger preference for this outcome compared to targets who considered it to be unlikely (b = 0.93, p < .001; see Model 1; Table 1).

Table 1.

Multilevel Regression Analysis on Perceived Preferences, Study 2.

	Model 1		Model 2		Model 3		Model 4
Predictor	b	SE	b	SE	b	SE	b	SE
Fixed effects
Intercept	3.99***	0.27	3.99***	0.28	3.94***	0.12	4.03***	0.13
Condition (1 = predicted, −1 = did not predict)	0.93***	0.18	0.93***	0.13	1.49***	0.18	1.18***	0.17
Empirical forecast-preference association	—	—	−0.46	1.01	—	—	−0.45	0.34
Condition × Empirical Forecast–Preference Association	—	—	2.25***	0.39	—	—	1.61***	0.43
Outcome desirability	—	—	—	—	0.62***	0.05	0.64***	0.06
Outcome desirability²	—	—	—	—	0.01	0.01	−0.01	0.02
Condition × Outcome Desirability	—	—	—	—	0.29***	0.08	0.21**	0.07
Condition × Outcome Desirability²	—	—	—	—	−0.12***	0.02	−0.06*	0.02
Random effects (participants)	Variance	SD	Variance	SD	Variance	SD	Variance	SD
Intercept	0.42	0.65	0.43	0.66	0.70	0.84	0.67	0.82
Empirical forecast-preference association	—	—	1.09	1.05	—	—	1.62	1.27
Outcome desirability	—	—	—	—	0.19	0.43	0.17	0.42
Outcome desirability²	—	—	—	—	0.004	0.06	0.002	0.05
Random effects (outcomes)	Variance	SD	Variance	SD	Variance	SD	Variance	SD
Intercept	1.93	1.34	1.99	1.41	0.08	0.28	0.08	0.27
Condition (1 = predicted, −1 = did not predict)	0.56	0.75	0.20	0.44	0.20	0.44	0.10	0.32

Note. Outcome desirability and empirical forecast-preference association were grand-mean centered.

***p < .001. **p < .01. *p < .05.

Next, I examined whether this effect was moderated by the empirical forecast–preference association (Model 2). The interaction effect between the experimental condition and the empirical forecast–preference association was significant (b = 2.25, p < .001). Figure 2 shows the pattern of this interaction by plotting the effect of targets’ forecast on perceived preference for each outcome as a function of the empirical forecast–preference association. This figure shows that individuals are more likely to infer preferences from forecasts of outcomes characterized by a stronger (vs. weaker) actual empirical association between desirability and likelihood judgments (r = .80, N = 28, p < .001).

Figure 2.

Perceived forecast–preference association (effect of forecast on perceived preferences, Cohen’s d) as a function of the empirical forecast–preference association (correlation between participants’ desirability and likelihood judgment for each outcome), Study 2.

To explore whether the effect of the experimental condition on perceived preferences was stronger for mixed desirability outcomes than for both highly desirable and highly undesirable outcomes, in the next step, I regressed perceived outcome preference on the experimental condition, linear and quadratic terms of outcome desirability and interactions of the experimental condition with the linear and the quadratic term of outcome desirability (Model 3). The effect of the condition was qualified by a significant interaction with outcome desirability (b = 0.29, p < .001) and with its quadratic term (b = −0.12, p < .001).

The pattern of this interaction is shown in Figure 3. Participants were more likely to infer targets’ preferences from their forecasts of outcomes of mixed desirability (e.g., outcomes of political elections) than of outcomes of low (e.g., a rise in income inequality) or high (e.g., eradication of international terrorism) desirability. Finally, as a robustness check, I entered all predictors and interaction effects in one model: Both interaction effects remained significant (Model 4), suggesting that the two tested moderators are independent of each other.

Figure 3.

Perceived forecast-preference association (effect of forecast on perceived preferences, Cohen’s d) as a function of average outcome desirability, Study 2.

As shown in Figure 3, participants were least likely to infer targets’ preferences from their predictions of the outcomes that most people find highly desirable or highly undesirable. To make sure that this pattern is not due to floor or ceiling effects, I conducted two additional analyses. First, I replicated the present findings using Tobit regression—a method recommended for outcome variables with floor and ceiling effects (Twisk & Rijmen, 2009; Supplemental Table S1). Second, if the quadratic effect of outcome desirability was simply a result of a lack of variation in the perceived preferences for highly desirable and undesirable outcomes, accounting for the extent of between-outcomes variation in perceived preferences should reduce this effect to nonsignificance. Therefore, I computed the SD of perceived preferences for each outcome and included it in the analyses as a control variable (Supplemental Table S2). Both additional analyses replicated the results presented above, suggesting that the potential floor and ceiling effects are unlikely to explain the findings.

Discussion

Study 2 showed that individuals tend to infer others’ preferences from their likelihood judgment of a range of different outcomes, including personal events and economic, political, and social trends. Although this effect emerged with respect to most predictions sampled in this study, some outcomes were affected more than others. Additional analyses showed that the perceived forecast–preference associations were related to actual empirical associations between forecasts and preferences. Specifically, participants were more likely to infer preferences from forecasts of outcomes characterized by a stronger (vs. weaker) actual empirical association between judgments of outcomes’ desirability and likelihood.

Also, consistent with the literature on lay dispositionism (e.g., Miller & Rorer, 1982), participants’ tendency to infer forecasters’ preferences from their predictions was moderated by the strength of the cues to internal attributions. Specifically, the forecast–preference link was stronger for mixed desirability outcomes than for desirable or undesirable outcomes, as the former involved much more ambiguity regarding any particular forecaster’s preferences than the latter, turning the information about his or her prediction into a relatively strong (vs. weak) cue to his or her preferences.

Study 3

While the results of Studies 1 and 2 showed that people tend to attribute others’ future forecasts to their preferences, real-life forecasters often make predictions that don’t reflect their preferences at all. For example, Sosnik (2017), a prominent member of the Democratic Party, has recently predicted Donald Trump’s reelection in 2020 in the Washington Post opinions section. People make preference-inconsistent forecasts for multiple reasons, including defensive pessimism (Norem & Cantor, 1986) or fear of tempting fate (Risen & Gilovich, 2008). Can expressing such preference-inconsistent forecasts make social observers doubt the forecasters’ preferences and group identification? For example, will a target person who made a pessimistic (vs. optimistic) forecast of a party’s winning chances in the upcoming elections be attributed a weaker desire for this party’s victory, even when being explicitly described as this party’s supporter? Study 3 was designed to answer this question.

Method

Participants

As the instructions included explicit information about the target’s preferences, I expected the effect to be smaller than in Study 1 (d = 1.30) and Study 2 (on average across outcomes, d = 0.52) and set the minimum sample size at N = 200 (which would be enough to detect an effect of d = 0.40 with 80% power). Two hundred twenty individuals were recruited on MTurk. Twenty-eight did not pass an attention check question (see below) and were removed, resulting in a final sample of 192 individuals (M_age = 35.69, SD_age = 11.26, 63.0% male).

Procedure

Participants read about Jack who “lives in a small Western European country and supports a certain political party—party AC.” All participants also learned that “Jack approves of his party’s program.” In the “predicted success” condition, they further read that Jack estimated his party’s “chances of winning the upcoming elections as very good.” In the “predicted failure” condition, they read that Jack estimated his party’s “chances of winning the upcoming elections as very slim.” To measure perceived preferences, participants indicated how much they thought Jack wanted his party to win the elections. They also rated Jack’s endorsement of his party’s political program, his party’s values, his loyalty to this party, his support of his party, and his identification with his party. Participants’ responses were combined into a scale of perceived party identification (Cronbach’s α = .94). All responses were given on a scale from 1 (not at all) to 9 (very much). Participants also indicated Jack’s willingness to vote for his party and to participate in the elections altogether (1 = very unlikely, 9 = very likely) and were asked to indicate whether Jack estimated his favorite party’s chances as very good or very slim (used as an attention check) and responded to sociodemographic questions.

Results and Discussion

The descriptive statistics can be found in Supplemental Table S2. Participants ascribed Jack a stronger desire for his party to win the elections and believed that Jack identified himself with his party more if Jack thought that the party’s chances of winning were very good (desire: M = 7.99, SD = 0.99; identity: M = 7.87, SD = 1.00) versus very slim, desire: M = 7.16, SD = 1.69, t (156) = 4.14, p < .001, d = 0.60; identity: M = 7.31, SD = 1.28, t (190) = 3.33, p = .001, d = 0.49.

Participants estimated Jack’s likelihood of voting for his party as higher if Jack thought that his party’s chances of winning were very good (M = 8.59, SD = 0.78) versus very slim (M = 7.84, SD = 1.64), t (137) = 4.09, p < .001, d = 0.55. Finally, participants also believed that Jack was more likely to participate in the elections altogether, when he thought that his party was likely (vs. unlikely) to win (M = 8.38, SD = 0.97 vs. M = 7.59, SD = 1.71), t (152) = 3.95, p < .001, d = 0.59.

To conclude, the mere act of forecasting an outcome shaped observers’ perceptions of a forecaster’s preferences, even when these preferences were explicitly stated in the instructions. Participants attributed a political party’s supporter a weaker desire for his party’s victory and a weaker identification and support for his party in general, if he made a pessimistic forecast of his party’s winning chances in the upcoming elections. In other words, making predictions consistent versus inconsistent with one’s stated preferences can have downstream consequences for social perception, making others doubt one’s expressed preferences and identity.

General Discussion

The present studies showed that people use the information about others’ forecasts of future outcomes to draw inferences about their preferences. Across the forecasts of 30 different outcomes, forecasters who described a particular future outcome as very likely were perceived as desiring this outcome more than forecasters who described it as very unlikely.

It is important to note that although on average across outcomes individuals showed a significant tendency to infer others’ preferences from their future forecasts, there was substantial between-outcome variation. While individuals readily inferred forecasters’ desires from some predictions (e.g., outcomes of sporting contests, elections, new product success, or weather forecasts), they were reluctant to do that in other cases (e.g., rise in income inequality, eradication of international terrorism). Consistent with previous research of lay dispositionism (e.g., Miller & Rorer, 1982), individuals’ tendency to infer preferences from future forecasts was substantially stronger for mixed desirability outcomes than for outcomes that most participants considered desirable or undesirable. In addition, the between-outcome variation in the effect of forecast on perceived preferences nearly perfectly corresponded to the between-outcome variation in the empirical forecast–preference associations. Participants were more likely to believe that forecasters’ preferences are reflected in their forecasts in case of outcomes characterized by a strong (vs. weak) empirical association between preferences and forecasts.

Although these results suggest that the desirability bias in individuals’ own judgment and their perception of others show a great deal of similarity, they also have one important difference. Specifically, research on defensive pessimism and “bracing for loss” showed that sometimes individuals make predictions that contradict their preferences, demonstrating a reversal of the desirability bias (Sweeny, Carroll, & Shepperd, 2006). Indeed, in Study 2, such a reversal was detected with respect to predictions of the eradication of international terrorism and a rise in income inequality: The more participants desired these outcomes, the less likely they considered them. In contrast, a reversal of the preference–expectation link never happened (or at least, never reached significance) when asked to judge others’ preferences based on their predictions (see Figure 1).

This is consistent with a common (Chambers, Epley, Savitsky, & Windschitl, 2008), although contested (Vazire & Mehl, 2008), idea that people might be better at introspection than at understanding others. For example, individuals don’t forget to take into account situational factors when explaining their own behavior but routinely underestimate the role of the situation and overestimate the role of personal characteristics when explaining the behavior of others (Nisbett, Caputo, Legant, & Marecek, 1973; but see Malle, 2006). Indeed, the present results showed that, in the domain of politics, a party’s supporter who expressed pessimism regarding his party’s success was perceived as having a weaker party identification and being less likely to support it by a vote than a more optimistic supporter of this party. In other words, individuals appear to doubt the political identity of a party’s proponent who expresses doubt in his party’s winning chances. This finding might be particularly important, as sometimes people make pessimistic forecasts for the outcomes they deem particularly desirable (Shepperd, Findley-Klein, Kwavnick, Walker, & Perez, 2000). More generally, these findings also hint at a possibility that individuals’ desires and preferences can serve as a signal to their identity, values, attitudes, and even future behaviors. Hence, people might use others’ forecasts of future outcomes not only to infer forecasters’ preferences but other aspects of identity as well.

The present findings contribute to several research areas. First, they extend the long-standing literature on the desirability bias by showing that people don’t only let their preferences guide their future forecasts but also use the information about other people’s forecasts to make judgments about their preferences. Second, recent advice-taking literature has shown that people generally prefer optimistic to pessimistic forecasters (Stavrova & Evans, 2018). The present findings suggest that people’s tendency to infer others’ preferences from their forecasts might underlie this effect and thus contribute to a small but rapidly growing forecasting and advice-taking literature (Bonaccio & Dalal, 2006).

Turning to practical implications, the inferences people make from others’ predictions might have downstream consequences for whom they befriend and maintain a professional relationship with. For example, a political advisor making a negative election forecast or a business analyst making a pessimistic earnings forecast might be considered disloyal and be replaced. Finally, if forecasters are aware of the inferences their clients make from their predictions, they might avoid communicating pessimistic forecasts, which might damage organizational performance in the long run (Morrison & Milliken, 2000).

To conclude, trying to foresee the future is a tough task, and it’s not surprising that people often let their preferences guide their predictions. The present research showed that social observers might be well aware of this and use the information about others’ forecasts to draw inferences about their preferences. In other words, how people see the future might have implications for how they are seen by others.

Supplemental Material

Supplemental Material, SPPS789607_suppl_mat - Social Perception of Forecasters: People See Forecasts of Future Outcomes as Cues to Forecasters’ Desires, Attitudes, and Identity

Supplemental Material, SPPS789607_suppl_mat for Social Perception of Forecasters: People See Forecasts of Future Outcomes as Cues to Forecasters’ Desires, Attitudes, and Identity by Olga Stavrova in Social Psychological and Personality Science

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

The supplemental material is available in the online version of the article.

Note

References

Bonaccio

Dalal

R. S.

(2006). Advice taking and decision-making: An integrative literature review, and implications for the organizational sciences. Organizational Behavior and Human Decision Processes, 101, 127–151.

Chambers

J. R.

Epley

Savitsky

Windschitl

P. D.

(2008). Knowing too much: Using private knowledge to predict how one is viewed by others. Psychological Science, 19, 542–548.

Cohrs

J. C.

Kielmann

Maes

Moschner

(2005). Effects of right-wing authoritarianism and threat from terrorism on restriction of civil liberties. Analyses of Social Issues and Public Policy (ASAP), 5, 263–276.

Hassin

R. R.

Bargh

J. A.

Uleman

J. S.

(2002). Spontaneous causal inferences. Journal of Experimental Social Psychology, 38, 515–522.

Hayes

S. P.

Jr . (1936). The predictive ability of voters. The Journal of Social Psychology, 7, 183–191.

Jones

E. E.

Worchel

Goethals

G. R.

Grumet

J. F.

(1971). Prior expectancy and behavioral extremity as determinants of attitude attribution. Journal of Experimental Social Psychology, 7, 59–80.

Judd

C. M.

Westfall

Kenny

D. A.

(2012). Treating stimuli as a random factor in social psychology: A new and comprehensive solution to a pervasive but largely ignored problem. Journal of Personality and Social Psychology, 103, 54–69.

Krizan

Miller

J. C.

Johar

(2010).Wishful thinking in the 2008 U.S. presidential election. Psychological Science, 21, 140–146.

Krizan

Windschitl

P. D.

(2007). The influence of outcome desirability on optimism. Psychological Bulletin, 133, 95–121.

10.

Krizan

Windschitl

P. D.

(2009). Wishful thinking about the future: Does desire impact optimism? Social and Personality Psychology Compass, 3, 227–243.

11.

Kubota

J. T.

Mojdehbakhsh

Raio

Brosch

Uleman

J. S.

Phelps

E. A.

(2014). Stressing the person: Legal and everyday person attributions under stress. Biological Psychology, 103, 117–124.

12.

Lench

H. C.

Ditto

P. H.

(2008). Automatic optimism: Biased use of base rate information for positive and negative events. Journal of Experimental Social Psychology, 44, 631–639.

13.

Malle

B. F.

(2006). The actor-observer asymmetry in attribution: A (surprising) meta-analysis. Psychological Bulletin, 132, 895–919.

14.

Malle

B. F.

Holbrook

(2012). Is there a hierarchy of social inferences? The likelihood and speed of inferring intentionality, mind, and personality. Journal of Personality and Social Psychology, 102, 661–684.

15.

Massey

Simmons

J. P.

Armor

D. A.

(2011). Hope over experience: Desirability and the persistence of optimism. Psychological Science, 22, 274–281.

16.

Miller

A. G.

Rorer

L. G.

(1982). Toward an understanding of the fundamental attribution error: Essay diagnosticity in the attitude attribution paradigm. Journal of Research in Personality, 16, 41–59.

17.

Morrison

E. W.

Milliken

F. J.

(2000). Organizational silence: A barrier to change and development in a pluralistic world. The Academy of Management Review, 25, 706–725.

18.

Nisbett

R. E.

Caputo

Legant

Marecek

(1973). Behavior as seen by the actor and as seen by the observer. Journal of Personality and Social Psychology, 27, 154–164.

19.

Norem

J. K.

Cantor

(1986). Defensive pessimism: Harnessing anxiety as motivation. Journal of Personality and Social Psychology, 51, 1208–1217.

20.

Reeder

G. D.

Monroe

A. E.

Pryor

J. B.

(2008). Impressions of Milgram’s obedient teachers: Situational cues inform inferences about motives and traits. Journal of Personality and Social Psychology, 95, 1–17.

21.

Risen

J. L.

Gilovich

(2008). Why people are reluctant to tempt fate. Journal of Personality and Social Psychology, 95, 293–307.

22.

Ross

(1977). The intuitive psychologist and his short-comings: Distortions in the attribution process. In Berkowitz

(Ed.), Advances in experimental social psychology (Vol. 10, pp. 173–220). New York, NY: Academic press.

23.

Ross

Nisbett

R. E.

(1991). The person and the situation: Perspectives of social psychology. New York, NY: McGraw-Hill.

24.

Shepperd

J. A.

Findley-Klein

Kwavnick

K. D.

Walker

Perez

(2000). Bracing for loss. Journal of Personality and Social Psychology, 78, 620–634.

25.

Simmons

J. P.

Massey

(2012). Is optimism real? Journal of Experimental Psychology: General, 141, 630–634.

26.

Sosnik

(2017). Trump is on track to win reelection. The Washington Post. Retrieved December 27, 2017, from https://www.washingtonpost.com/opinions/trump-is-on-track-to-win-reelection/2017/10/06/91cd2af0-aa15-11e7-850e-2bdd1236be5d_story.html?utm_term=.6f8c6d43a2a7

27.

Stavrova

Evans

A. M.

(2018). Examining the trade-off between confidence and optimism in future forecasts. Journal of Behavioral Decision Making. Advance online publication. doi:10.1002/bdm.2087

28.

Sweeny

Carroll

P. J.

Shepperd

J. A.

(2006). Is optimism always best?: Future outlooks and preparedness. Current Directions in Psychological Science, 15, 302–306.

29.

Tal-Or

Papirman

(2007). The fundamental attribution error in attributing fictional figures’ characteristics to the actors. Media Psychology, 9, 331–345.

30.

Twisk

Rijmen

(2009). Longitudinal tobit regression: A new approach to analyze outcome variables with floor or ceiling effects. Journal of Clinical Epidemiology, 62, 953–958.

31.

Uleman

J. S.

Saribay

S. A.

Gonzalez

C. M.

(2008). Spontaneous inferences, implicit impressions, and implicit theories. Annual Review of Psychology, 59, 329–360.

32.

Van Overwalle

Van Duynslaeger

Coomans

Timmermans

(2012). Spontaneous goal inferences are often inferred faster than spontaneous trait inferences. Journal of Experimental Social Psychology, 48, 13–18.

33.

Vazire

Mehl

M. R.

(2008). Knowing me, knowing you: The accuracy and unique predictive validity of self-ratings and other-ratings of daily behavior. Journal of Personality and Social Psychology, 95, 1202–1216.

34.

Weinstein

N. D.

(1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806–820.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB