Looking Behind and Looking Ahead: Personality Differences in Counterfactual and Prefactual Thinking

Introduction

Individuals often allow their thoughts to drift from the present moment to the past and the future. When they reflect on the past, they not only recall things that happened but may imagine how they could have happened differently. This imagination of alternatives to reality is known as counterfactual thinking (CFT). CFT often takes the form of if only …  thoughts about what might have been, for instance, a student might imagine if only I had studied harder, I could have passed my exams. Similarly, people reflect on the future, known as prefactual thinking (PFT), imagining different possible outcomes, for instance if I work hard for my exam next month, I will pass where there is a causal link between an antecedent and consequent (Epstude, Scholl, & Roese, 2016). These types of thought can impact on our well-being and ongoing behavior. There is evidence of individual differences in the propensity for individuals to spontaneously engage in CFT (Bacon, Walsh, & Martin, 2013), and Markman and Miller (2006) have shown that counterfactuals may serve different functions for different people. However, although we know these differences exist, we currently know little about what might determine them. This study is concerned with whether individual differences in the Big Five personality traits (neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness) can account for the tendency toward thinking counterfactually and prefactually.

Mental simulation occurs automatically and is a pervasive feature of human thought (Byrne, 2016). Although simulating the future and past may share similar processes (Schacter, Benoit, DeBrigard, & Spuznar, 2015), they can have different consequences (De Brigard, Szpunar, & Schacter, 2013) and functions (Ferrante, Girotto, Straga, & Walsh, 2013). Simulating future possibilities plays a key role in planning and goal setting (Epstude et al., 2016) by allowing individuals to consider the consequences of different actions and events. CFT has also been linked to future planning (Epstude & Roese, 2008) but compared with future thoughts, it is more likely to focus on uncontrollable events (Ferrante et al., 2013; Mercier et al., 2017) and in doing so, may also allow individuals to excuse past failures (Byrne, 2016). The CFT that is most functional is that which is self-focused and which concerns how events might have turned out better (termed upward CFT). Actions simulated in such thoughts are inherently controllable and retrospective control perceptions can enhance functions such as adaptive preparation, anticipatory planning, and behavior change (Epstude & Roese, 2008; Nasco & Marsh, 1999; Roese & Morrison, 2009).

CFT and PFT may also have different consequences for affect. When people imagine alternatives to the past, they often compare what happened to what might have been. Thinking about how past events could have been better can therefore lead people to feel worse about how things are (Roese, 1997). This has been associated with negative emotions generally (Bacon et al., 2013) and specifically feelings such as regret (Roese & Summerville, 2005; Seta & Seta, 2013), blame (Alicke, Buckingham, Zell, & Davis, 2008), guilt and shame (Niedenthal, Tangney, & Gavanski, 1994), potentially impacting on how well people cope with traumatic events (Bhushan & Kumar, 2012; Davis & Lehman, 1995). Importantly, these relationships hold even after controlling for more general ruminations about the past suggesting that the tendency to engage in CFT is independent of rumination and has important consequences. CFT can also be triggered by low mood (Roese & Hur, 1997) and hence a self-perpetuating cycle of CFT and low mood may result (Roese et al., 2009). In contrast, when people think ahead, they can imagine a future that is still possible. Probably for this reason, imagining a better future is more likely to lead to positive mood (Epstude et al., 2016; Ruby, Smallwood, Engen, & Singer, 2013).

Little research to date has focused on individual differences in the tendency to think counterfactually or prefactually. Understanding these differences in important given their potential impact on behavior and mood. In this study, we focus on the Big Five traits, arguably the most widely researched model of personality, yet we know little about the influence of the five traits on CFT and no research has previously considered them in the context of PFT. One early study (Kasimatis & Wells, 1995) found no significant association between CFT and any of the Big Five. However, they did observe positive associations between CFT and low self-esteem, negative affectivity, and depression. It is therefore surprising that they report no association with neuroticism, a trait typified by negative emotions such as these. The Big Five model proposes that each trait is underpinned by six subfacets (specific aspects) and that analyses which include scores on these can provide a more fine-grained picture of an individual’s trait profile. This study extends that of Kasimatis and Wells (1995) by including the Big Five subfacets in addition to the main superordinate trait scores.

There is virtually no research into individual differences in the tendency to prefactual thought, and what exists is equivocal. Sanna (1998) suggested that optimists have a greater tendency to think retrospectively about how things could have been worse, while defensive pessimists tend to look to the future, but expect the worst. Conversely, however, Smallwood and O’Connor (2011) have shown that a negative mood induction increases thinking about the past, whereas a positive mood induction increased thinking to the future. One study has attempted to examine the relationship between Big Five traits and general forms of thinking in the past, present, or future (Fortunato & Furey, 2009). In that study, future thinking correlated positively with extraversion and openness, whereas past thinking correlated positively with neuroticism and negatively with extraversion. The latter results were explained by the general association between extraversion and cheerfulness and positive mood. However, this study used an entirely psychometric method, asking participants to complete a self-report questionnaire on the type of thinking they tend to employ, and past and future thinking were not operationally defined. In this study, we specifically examine CFT and PFT using a CFT research paradigm which collects data on participants’ spontaneous thoughts in relation to a given event (Bacon et al., 2013; Briazu, Walsh, Deeprose, & Ganis, 2017; McEleney & Byrne, 2006). To date, there is only one study of individual differences using this method (Bacon et al., 2013) which further indicated that CFT was associated with negative emotions.

Overall, there is consistent evidence that CFT is associated with low mood and Fortunato and Furey (2009) reported that thinking about the past is negatively related to extraversion (a trait generally typified by positive attitudes and mood, Costa & McCrae, 2006), and positively associated with neuroticism. Individuals high in neuroticism tend to experience negative affect such as sadness, anxiety, and depression, as well as fear, embarrassment, anger, and disgust (Costa & McCrae, 2006; Widiger, 2009). These can result in poor coping, impulsivity, and a range of negative-life outcomes (Lahey, 2009). As such, although Kasimatis and Wells (1995) reported no relationship between CFT and trait neuroticism, we proposed that association may be observed if the subfacets of this trait (particularly depression, anxiety, and angry hostility) are examined. Conversely, individuals who think ahead and plan for better future outcomes are assumed to be those with generally more positive and goal-oriented dispositions. Accordingly, Fortunato and Furey (2009) reported associations between thinking about the future and both extraversion and openness to experience, a trait associated with interest in new experiences and ideas (Costa & McCrae, 2006). We made the following a priori predictions:

Prediction 1: CFT will be positively associated with neuroticism subfacets depression, anxiety, and angry hostility.

Methods

Results

The mean number of spontaneous counterfactuals observed was 1.23 (SD = 1.06), and the mean number of prefactual thoughts was .73 (SD = 1.10). CFT and PFT were not significantly associated (r = .03, p = .56). Table 1 presents correlations between these measures, the Big Five superordinate traits, and their subfacets. In terms of the superordinate traits, neither CFT nor PFT showed any significant associations with either openness or conscientiousness. As predicted, however, CFT was positively and PFT was negatively associated with neuroticism. In addition, we observed a positive correlation between extraversion and PFT but not CFT. Both forms of thinking were associated with agreeableness, for CFT, the relationship was negative, while for PFT, it was positive.

OPEN IN VIEWER

Table 1.

Correlations Between CFT, PFT, and the Big Five Traits and Subfacets.

Trait/subfacet	CFT	PFT
Neuroticism	.14*	−.15*
Extraversion	.05	.15*
Openness	.01	.04
Agreeableness	−.15*	.13*
Conscientiousness	.02	.06
N1 anxiety	.15*	−.10
N2 angry hostility	.21*	−.11*
N3 depression	.05	−.13*
N4 self-consciousness	.09	−.12*
N5 impulsiveness	.04	−.10
N6 vulnerability	.09	−.12*
E1 warmth	−.01	.16*
E2 gregariousness	.03	.08
E3 assertiveness	.10	.12*
E4 activity	.09	.07
E5 excitement seeking	−.02	.09
E6 positive emotions	.03	.13*
O1 fantasy	.10*	.002
O2 aesthetics	−.06	.003
O3 feelings	.04	−.03
O4 actions	−.03	.05
O5 ideas	−.01	.06
O6 values	.04	.08
A1 trust	−.15*	.06
A2 straightforwardness	−.17*	.16*
A3 altruism	−.02	.12*
A4 compliance	−.08	.07
A5 modesty	−.02	.06
A6 tender-mindedness	−.11*	.02
C1 competence	.04	.06
C2 order	.01	−.002
C3 dutifulness	−.03	.02
C4 achievement striving	.05	.02
C5 self-discipline	−.01	.07
C6 deliberation	−.08	.04

Note. CFT = counterfactual thinking; PFT = prefactual thinking.

In terms of the facet correlations in Table 1, we can see that CFT is positively associated with N1 and N2 (anxiety and anger respectively), while PFT shows negative associations with all neuroticism facets except N1 (anxiety) and N5 (impulsiveness) where the association does not reach significance. For agreeableness, Table 2 indicates negative associations between CFT and A1 (trust), A2 (straightforwardness), and A6 (tender-mindedness), while PFT shows positive associations between agreeableness facets A2 (straightforwardness) and A3 (altruism). PFT also presents a positive correlations with extraversion facets E1 (warmth), E3 (assertiveness), and E6 (positive emotions).

OPEN IN VIEWER

Table 2.

Results of Regression Analyses on CFT and PFT.

	St. β	t	p	95% confidence interval	Adj. R2
Counterfactual thinking
1
N	.19	3.14	.002	[.003, .01]
E	.11	1.74	.08	[−.001, .01]
O	.001	0.02	.98	[−.01, .01]
A	−.15	−2.72	.01	[−.02, −.003]
C	.11	1.89	.06	[−.002, .01]	.04
2
N2	.18	3.32	.001	[.01, .05]
A2	−.15	−2.65	.01	[−.05, −.01]	.06
Prefactual thinking
1
N	−.11	−1.70	.09	[−.01, .001]
E	.11	1.84	.07	[−.003, .01]
O	−.001	−0.02	.99	[−.01, .01]
A	.12	2.09	.03	[−.004, .01]
C	−.003	−0.04	.97	[−.01, .01]	.03
2
N3	.12	2.09	.04	[.002, .06]
E1	.15	2.75	.01	[.01, .06]
A2	−.13	−2.25	.03	[−.04, −.003]	.05

Note. In each case, Analysis 1 tests the Big Five superordinate traits and Analysis 2 tests the subfacets of the five traits.

The results of regression analyses are shown in Table 2. We first examined the effects of the Big Five superordinate traits on CFT and PFT in turn. These accounted for 3% variance in CFT and 5% in PFT. In both cases, significant independent effects of agreeableness were indicated, a negative effect on CFT and a positive one for PFT. A positive effect of neuroticism on CFT was observed and a negative effect on PFT. Extraversion shared variance with PFT but showed no effects on CFT. We conducted further regressions with the 30 individual facet scores. In this case, due to the relatively large number of independent variables, we used a forward stepwise method and the results are also shown in Table 2. Variance in CFT was accounted for by facets N2 (anger; positive) and A2 (straightforwardness; negative). In PFT, variance was explained by three facets, E1 (warmth; positive), N3 (depression; negative), and A2 (straightforwardness; positive).

Discussion

This study is the first to examine the Big Five personality traits as potential explanations for individual differences in the tendency toward both CFT and PFT. In sum, our results suggest that contrasting levels of neuroticism and agreeableness can differentiate between a tendency toward one or other form of thought. PFT was also associated with higher levels of extraversion.

Neuroticism was positively associated with levels of CFT but negatively with PFT. This falls in line with our predictions and with previous research which has shown that CFT is related to low mood states and PFT with more positive ones. Examination of the facet scores allowed for a more fine-grained analysis. This revealed that scores on facets N1 (anxiety) and N2 (angry hostility) were both associated with higher levels of CFT though only the latter explained variance in CFT independently of the other facets. This facet (N2, angry hostility) refers to the tendency to experience anger and hostility but also related states such as bitterness and frustration. We can imagine how looking back at a negative outcome might trigger such feelings. Individuals high in neuroticism are motivated to avoid potential threats, and other research has shown that people with a high level of prevention focus (risk-averse and maintain the status quo in order to remain safe) focus on past experiences, rather than thinking about future goals (e.g., Higgins, 1998). This further supports the idea that negative emotionality is closely associated with dwelling on the past. Moreover, there is evidence that counterfactual-related regret for actions taken can trigger what have been termed hot emotions, including anger, frustration, and embarrassment (Gilovich, Medvec, & Kahneman, 1998; Kedia & Hilton, 2011). In addition, individuals who score highly on N2 (angry hostility) tend to be lower in agreeableness (Costa & McCrae, 2006) which was also associated with CFT in the present data. Overall, the link between neuroticism and CFT is more complex than a mutual association with negative affect and, given the poor life outcomes associated with neuroticism, is worthy of further study. It may be that excessive CFT, together with associated rumination, may help to explain the mechanisms by which neuroticism and outcomes are linked.

The finding that low agreeableness was associated with a higher rate of CFT is an interesting result that we had not predicted. Costa and McCrae (2006) describe agreeableness as a dimension of interpersonal tendencies whereby high scores indicate altruism and a positive view of human nature (though also possible overdependency and gullibility) while low scores reflect antagonism, cynicism and a tendency toward competition rather than cooperation, fitting given the established association between CFT and negative mood. Examination of the agreeableness facet data revealed that CFT was negatively correlated with facets A1 (trust), A2 (straightforwardness), and A6 (tender-mindedness), though only A2, shared a significant amount of independent variance with CFT in regression, reflecting a lack of straightforwardness in dealings with others, a willingness to be manipulative and disingenuous. Interestingly, recent research by Briazu et al. (2017) has shown that individuals with a tendency to lie also seem to generate more counterfactual thoughts. Like CFT, some forms of lying require imagination of an alternative to past events. In direct contrast, agreeableness scores were positively associated with PFT.

In addition, PFT was positively associated with extraversion as predicted. Extraverts are sociable, outgoing, and talkative, and they like excitement and stimulation and tend to be optimistic and cheerful. The facet-level correlations indicated this effect was driven by facets E1 (warmth), E2 (gregariousness), E3 (assertiveness), and E6 (positive emotions). However, only E1 (warmth) was shown to share independent variance with PFT in our regression analysis. We also anticipated that PFT would be related to openness to experience and the possibility of a range of potential future outcomes. Our results here suggest these characteristics are more typical of those captured by the trait of extraversion, rather than openness to experience as we had originally anticipated.

In terms of the positive associations between CFT and neuroticism and between PFT and extraversion, our results reflect those of Fortunato and Furey’s (2009) psychometric study. Their other finding that openness is related to future thinking was not observed. However, their thinking questionnaire included items such as People think of me as a visionary and I am known for invention/innovation. These items may reflect creativity as much as past or future thinking and, given that creativity and intellectual curiosity are intrinsic to openness, it is perhaps unsurprising the two were related in that study. We did not predict a link between openness and CFT and did not observe one in terms of the superordinate trait. However, we did observe a modest, though significant, positive association between facet O1 (openness to fantasy) and CFT. In retrospect, however, this might have been predicted given that Bacon et al. (2013) reported an association between CFT (using the same measure) and fantasy proneness, a trait associated with an active fantasy life and imagination. In addition, the cognitive need for closure is found to be negatively associated with thinking of alternative outcomes to events (Hirt, Kardes, & Markman, 2004) and also with openness (Stalder, 2007). This evidence only suggests indirect evidence for a CFT-openness association, however, may be worthy of further study. Stalder (2007) highlights how need for closure can be differentiated into two subfactors which inversely correlate with openness. How these factors are associated with CFT and PFT might form a useful focus for further research.

The one trait which did not feature at all in our analysis was conscientiousness. Previous work has not suggested any link between this and CFT or PFT, though we might hypothesize that conscientious individuals (those who are organized, determined, and strive to achieve) may wish to learn from past mistakes and set goals for future improvement. However, neither superordinate trait scores not those for any of the subfacets of conscientiousness showed any relationship with the tendency toward CFT or PFT. Again, a more fine-grained analysis that is possible here may reveal some useful associations.

In the above analysis, we have tended to assume that personality drives thinking and, because our data are correlational, we need to consider potential inverse effects. While the Big Five traits are generally considered to be fairly inherent, it may be possible that CFT or PFT can precede emotions that we associate with certain traits. For instance, it has been suggested that CFT can trigger negative moods, such as when thoughts about how things might have turned out better generate high levels of regret (Roese & Morrison, 2009; Roese, Park, Smallman, & Gibson, 2008). Roese et al. (2008) state that the correlational nature of their data precludes a definitive assignment of causation and that CFT may be part of a cyclical process of negative affect and problem-focused cognition (Roese et al., 2008). As noted earlier, thinking about the future has also been shown to precede positive mood, even if thoughts are themselves negative (Ruby et al., 2013). Nevertheless, our findings are novel and informative and suggest some fundamental individual differences in how personality may impact across the two forms of thinking. Exactly how specific traits function in this way will be a useful focus of for future research.

Finally, our findings may have implications for the functional theory of CFT. For instance, retrospective control perceptions can enhance functions such as adaptive preparation, anticipatory planning, and behavior change (Epstude & Roese, 2008; Nasco & Marsh, 1999; Roese & Morrison, 2009), and both control perceptions and positive imagery are known to be protective of mental health and helpful in coping after stressful events (Taylor, Kemeny, Reed, Bower, & Gruenewald, 2000). Markman and Miller (2006) showed that CFT depleted perceptions of outcome control in depressed patients, suggesting that CFT may be less functional with increasing depressive symptomology. Potentially, this may be also the case with high neuroticism and present further motivation for research such as this into the role of personality traits. Future research might usefully explore more specifically the ways in which neuroticism and agreeableness impact on functional counterfactual outcomes.

Our findings raise the question as to why Kasimatis and Wells (1995) reported no significant associations between the Big Five and CFT. They reported upward counterfactuals (those we focus on here) as well as downward counterfactuals (those where people simulate worse potential outcomes), however, used different methodology in that they included scenarios with positive and neutral as well as negative outcomes. They state that the tendency to think counterfactually was consistent across time and scenario type (i.e., people who generate a lot of counterfactuals in one scenario, also do in others) and report data reflecting only this general tendency, rather than for the different scenario types. It is probable that averaging the number of counterfactuals generated across all scenarios may have masked some effects. Furthermore, they do not discuss subfacets of the Big Five traits. A future study could examine the relationship between the Big Five, their subfacets, and both counterfactuals and prefactuals generated across scenarios with different outcomes.

The study is not without some limitations, particularly the use of a single scenario with a negative outcome, which does not readily allow for the examination of downward counterfactuals or the content of those counterfactual or prefactual thoughts. However, we wanted to directly compare the tendency toward CFT and PFT and the latter does not tend to be observed in a downward form—people do not plan ahead for worse outcomes. Nevertheless, we recognize the need for further studies to extend this line of enquiry, and we have suggested some in this discussion. A related point is that according to the functional approach, CFT as well as PFT works to prepare people for future events (Epstude & Roese, 2008; Smallman & Roese, 2009) and as such we would expect an association between CFT and PFT which was not observed in this study. It is likely that such an association would be observed were we to study the overlapping content of counterfactual and prefactual thoughts and whether the Big Five traits mediate the relationship. The content of counterfactuals has been found to vary between focus on past and future, for instance, Choi and Markman (2018) showed that reflection on events from the perspective of understanding the past (vs. preparing for the future) tended to elicit more subtractive CFT (I wish I had not … ) than additive CFT (I wish I had … ). These trends might also vary between individuals as a function of personality traits. Our scenario does not provide rich enough data for such an analysis, and we recommend that future studies incorporate methods which capture more detail of the focus and direction of thoughts generated.

In conclusion, this is the first study to identify a clear relationship between CFT, PFT, and Big Five traits. Given the ubiquity of the Big Five in individual differences research, understanding their correlates is important for the ongoing development of personality theory. Moreover, neuroticism is known to predict an array of negative-life outcomes for health and well-being (Lahey, 2009; Widiger, 2009). CFT in particular may be one of the mechanisms which influence these outcomes and further work to examine this possibility is desirable. That PFT is associated with consistent personality traits characterized by positive mood and interpersonal tendencies is an important new finding and one which might usefully inform interventions to change unfavorable cognitions and develop more positive methods of thought.

Appendix

The scenario and instructions presented to participants. After reading the scenario, participants were given 5 minutes to write their diary entry.