Retrospective future thinking: Keeping distant personal future events mentally close

Abstract

By manipulating our basic mode of prospectively thinking about the future, the present study examines the effect of retrospective future thinking on future events imagined as if they had already taken place. Here, 142 young adults were randomly assigned to report five autobiographical future events either prospectively from the perspective of their current self or retrospectively, imagining events from the perspective of their 100-year-old self. Participants indicated the expected age of occurrence and assessed phenomenological characteristics for each event. Results suggest that a shift in future thinking perspective affects the content and temporal distribution of future events. Characteristics of prospective future events diminished with increasing distance, whereas retrospective future events included overall more spatial details and remained high on belief in occurrence, vividness, and rehearsal across life. Retrospective future thinking influences the psychological distance of autobiographical future events, allowing us to hold even distant future events psychologically close.

Keywords

Retrospective future thinking autobiographical event mental time travel cultural life script construal-level theory prospection

Introduction

Imagine your last working day before retirement. Where are you? What can you see and hear? Who is with you and how do you feel? Mentally constructing a representation of a specific autobiographical event that might take place in one’s future is referred to as episodic future simulation (Szpunar et al., 2014). It represents one mode of prospection, our ability to imagine from the perspective of the current self what might happen tomorrow, a week, or a year from today (Szpunar et al., 2014). Previous research has investigated the content, functions, phenomenological characteristics, and mechanisms characterising future thought (Schacter et al., 2017; Szpunar, 2010). However, little is known about how engaging in a perspective that deviates from our normal mode of prospective future thinking affects mental representations of future events. The present study aims to investigate whether—and if yes, how—simulating future events from a perspective as if they had already taken place might affect mental future event representations.

How one’s last working day before retirement, an event that is yet to be experienced, is mentally represented likely depends on its perceived psychological distance. The concept of psychological distance describes the subjective experience that something is near or far in time (temporal distance), space (spatial distance), self (social distance), and probability (hypotheticality) (Trope & Liberman, 2010). Construal-level theory (CLT, Trope & Liberman, 2003, 2010) proposes that representations of future events are construed on a continuum of abstraction, ranging from low-level to high-level construals, and that, as psychological distance increases across at least one of the dimensions outlined above, events will be constructed in more high-level construals (Trope & Liberman, 2010). Low-level construals include concrete and contextual information (e.g., reading a textbook) and inform us of event content, whereas high-level construals focus on general, essential characteristics in simpler, more schematic representations (e.g., gaining knowledge) and highlight the broader meaning of an event (Trope & Liberman, 2010). One form of high-level construal schematic representations are cultural life scripts. They refer to culturally shared depictions of an idealised life, including mostly positive important transitional events expected to happen in young adulthood (Berntsen & Rubin, 2004). The dimensions of psychological distancing are assumed to be interrelated and, importantly, changes in psychological distance are hypothesised to influence the structure and organisation of personal information and associated future predictions, actions, and preferences (Trope & Liberman, 2010).

In line with this, changes in psychological distance have been shown to affect mental representations of prospectively simulated future events. For example, future representations of temporally distant events are perceived as more personally significant than temporally close events (Addis et al., 2008; Berntsen & Bohn, 2010; Özbek et al., 2017) and draw more on schematised knowledge or high-level construal information (Berntsen & Bohn, 2010; Özbek et al., 2017). Furthermore, more temporally distant events are associated with low-level contextual information, such as fewer sensory details (Berntsen & Bohn, 2010; D’Argembeau & Van der Linden, 2004; Özbek et al., 2017), reduced feeling of pre-experience, and are less likely to be experienced from a first-person perspective than temporally close events (D’Argembeau & Van der Linden, 2004). Studies manipulating social distance by altering future thinking instructions from the psychologically close first-person perspective to the psychologically distant third-person perspective showed that socially distant events are associated with less pre-experience and lower clarity of people (McDermott et al., 2016) but with increased achievement motivation for the simulation of successfully completing a future task (Vasquez & Buehler, 2007).

These studies support the idea that increases in psychological distance are associated with future event representations characterised by less low-level and more high-level construal information (Trope & Liberman, 2010). However, experimentally manipulating the temporal perspective from which individuals simulate future events will provide additional evidence of the relationship between psychological distance and construal level. Here, we manipulated temporal perspective by asking participants either to imagine autobiographical future events, looking from their current self forward in time (prospective future thinking), or to imagine being 100 years old and to remember from this perspective events they would have experienced in their 100-year-old life (retrospective future thinking). Engaging in retrospective future thinking might be a useful tool in contexts that benefit from long-term planning as well as in settings focusing on the reflection of current circumstances and upcoming decisions, such as developing and maintaining positive health behaviours (e.g., Dassen et al., 2016) and economic savings plans (Macrae et al., 2017) that are likely to be beneficial in later life.

We hypothesised that retrospective future thinking serves as a form of psychological distancing, representing a psychologically more distant perspective than prospective future thinking. Following CLT (Trope & Liberman, 2010), we expected that a change in temporal perspective would affect future event representations in the following ways. First, retrospectively reported future events were expected to be dated further into the future, leading to a wider temporal distribution of events across the lifespan than prospectively reported future events (Trope & Liberman, 2010). Second, it was expected that future events reported retrospectively would be construed in more high-level terms than future events reported prospectively. Indicators of high-level construal were the reliance on schematised knowledge, such as cultural life scripts (Berntsen & Rubin, 2004) and the tendency to refer less often to specific events but more often to general schematic information (Williams et al., 2007). Highlighting the meaning of high-level construals, retrospectively constructed future events were expected to be perceived as more personally significant than prospectively constructed future events (e.g., Addis et al., 2008; Berntsen & Bohn, 2010). Furthermore, high-level construals were expected to be characterised by lower cognitive feelings (likelihood, belief in occurrence, and autonoetic experience) (e.g., D’Argembeau & Van der Linden, 2004, 2012) and fewer sensory-perceptual details (sensory details, spatial layout, vividness, and visual perspective) (e.g., Berntsen & Bohn, 2010; D’Argembeau & Van der Linden, 2004). Finally, we explored the effect of future thinking perspective on the phenomenological characteristics of emotional details (valence-then, valence-now, and intensity) and rehearsal.

In short, simulating a future event, such as one’s last day before retirement retrospectively as if it had already taken place rather than prospectively from the perspective of the current self was expected to lead to mental representations dated further into the future and construed in more high-level terms associated with heightened significance.

Method

The data presented here are part of a larger study investigating the effect of two future thinking perspectives on mental representations of personal goals and autobiographical events (Roderer et al., 2022). Here, we report data investigating the effect of future thinking perspective on the content and characteristics of autobiographical events.

Participants

Data were collected online from 142 young US participants aged between 18 and 25 years (M_age = 21.9 years, SD = 1.58). Of these, 80 were female, 60 male, and 2 preferred not to say. Most participants (93.7%) had completed at least 12 years of education. Overall, 56.3% identified as White, 14.8% as Black or African American, 9.9% as Asian, 8.5% as Hispanic, Latino or Spanish origin, 0.7% as American Indian or Alaska Native, 8.5% as biracial, and 1.4% with another ethnic background. Participants took on average 28.1 min (SD = 15.77) to complete the study and were compensated with US$3.5. The demographics for both perspective conditions are presented in Table 1.¹ The study was conducted in accordance with current professional ethical guidelines and was reviewed by the Local Research Ethics Committee. The committee found that no scientific ethical considerations spoke against the implementation of the study.

Table 1.

Demographic characteristics, completion time, and mood rating of participants in prospective condition and retrospective condition.

	Prospective		Retrospective
	M	SD	M	SD
Age	21.94	1.66	22.04	1.51
Years of education	14.56	3.22	14.41	3.06
Completion time in minutes	26.45	13.91	29.71	17.27
Mood rating	4.07	5.24	4.97	4.20
	n	%	n	%
Gender
Female	37	54.4	43	58.1
Male	29	42.6	31	41.9
Prefer not to say	2	2.9	0	0
Ethnicity
American Indian or Alaska Native	0	0	1	1.4
Asian	8	11.8	6	8.1
Black or African American	9	13.2	12	16.2
Hispanic, Latino, or Spanish origin	6	8.8	6	8.1
White	39	57.4	41	55.4
Some other origin	1	1.5	1	1.4
More than one choice	5	7.4	7	9.5

Materials

Manipulation of future thinking perspective

The experimental manipulation of future thinking perspective was based on the random assignment of participants to one of two perspective conditions. The instructions were the same as reported in the work of Roderer et al. (2022). In the prospective condition, participants (n = 68) were asked to think about life experiences that may take place in their own personal future, while participants in the retrospective condition (n = 74) were asked to imagine they are 100 years old and to look back from this perspective on their life experiences (see also Roderer & Bohn, in press). From their assigned perspective condition, participants were asked to report five life goals they have in their life (prospective condition) or had in their 100-year-old life (retrospective condition) and to rate life goal qualities on several scales. For each reported life goal, participants were asked to either imagine a specific event in their own personal future (prospective condition) or remember from the perspective of their 100-year-old self a specific event (retrospective condition) closely related to their goal. The instructions specified that the participants should imagine the event in as much detail as possible—with details about location, activity, people, objects, sensory reactions, and emotions. A specific event was defined as an event that should only happen once at a specific time and place and should last a few minutes or hours but no longer than a day (de Vito et al., 2012; Williams et al., 2007). An example of a specific event was provided, adjusted from de Vito et al. (2012). When the participants had a specific event in mind, they were asked to provide an event title or keyword, a short event description, and the expected date of occurrence. Furthermore, they were asked to assess the phenomenological characteristics of each event on several scales.

Phenomenological characteristics

The phenomenological characteristics of future events are mainly described on the three dimensions of sensory-perceptual details, cognitive feelings, and the integration with autobiographical knowledge (D’Argembeau & Van der Linden, 2012; Ernst et al., 2018). The sensory-perceptual details of future events were assessed by the items visual imagery, auditory imagery, and spatial layout (Özbek et al., 2017; Rubin et al., 2003), as well as the items vividness and visual perspective (Berntsen & Bohn, 2010). Cognitive feelings associated with future events were assessed by two items for autonoetic experience, including mental time travel and the feeling of p/reliving (D’Argembeau & Van der Linden, 2006; Rubin et al., 2003) and one item assessing the belief or likelihood of an event, depending on perspective condition. Participants in the retrospective condition rated their belief that the event really took place as remembered (Berntsen & Bohn, 2010; Rubin et al., 2003) and participants in the prospective condition rated the likelihood that the event will take place in the future (Berntsen & Bohn, 2010). The integration of future events with autobiographical knowledge was assessed by the item importance (Berntsen & Bohn, 2010) and two items from the Centrality of Event Scale measuring the extent to which a life event is central to an individual’s life story and identity (Berntsen & Rubin, 2006). In addition to the three main dimensions of phenomenological characteristics, three items assessed the emotions associated with event simulation: One item assessed emotions at the time when the event happened or is expected to happen (valence-then) (D’Argembeau & Van der Linden, 2006), one item assessed the emotions in the present moment when a person is thinking about the event (valence-now) (Berntsen & Bohn, 2010), and one item assessed the intensity of emotions (Berntsen & Bohn, 2010). Furthermore, the rehearsal of events was assessed on the two items rehearsal talk and rehearsal thought (Berntsen & Bohn, 2010; Johnson et al., 1988). Most of the items were rated on a 7-point Likert-type scale except for the two items on life story and identity, which were rated on a 5-point Likert-type scale (1 = totally disagree, 5 = totally agree) and the two items on emotional valence, which were rated from –3 (extremely negative) to +3 (extremely positive). The items assessing phenomenological characteristics of events and their corresponding scales are presented in Table 2 for the prospective condition. The wording of all items was tailored to events imagined from the perspective of the current self and events remembered from the perspective of the 100-year-old self (Özbek et al., 2017).

Table 2.

Phenomenological characteristic items and their corresponding scales in the prospective condition.

Characteristic	Description of each rating scale
1. Valence-then	If this event happens, my emotions would be (–3 = extremely negative, +3 = extremely positive)
2. Valence-now	The emotions I have when I imagine the event are (–3 = extremely negative, +3 = extremely positive)
3. Intensity	The emotions I have when I imagine the event are intense (1 = not at all, 7 = to a very high degree)
4. Vividness	This imagined event is vivid (1 = not at all, 7 = to a very high degree)
5. Visual imagery	As I imagine the event, I can see it in my mind (1 = not at all, 7 = as clearly as if it were happening now)
6. Auditory imagery	As I imagine the event, I can hear it in my mind (1 = not at all, 7 = as clearly as if it were happening now)
7. Spatial layout	As I imagine the event, I know the location of people and objects in the place where it might occur—spatial layout (1 = not at all, 7 = as clearly as if it were happening now)
8. P/reliving	As I imagine the event, I feel as though I am experiencing it (1 = not at all, 7 = completely)
9. Mental time travel	As I imagine the event, I feel that I travel forward to the time when it would happen (1 = not at all, 7 = completely)
10. Visual perspective	When I imagine the event, I primarily see what might happen from a perspective as seen through (1 = my own eyes, 7 = an observer’s eyes)
11. Likelihood^a (belief^b)	How likely is it that the imagined event will take place in your future?^a (1 = not likely at all, 7 = completely likely)I believe that the event really took place the way I remember it and I did not imagine anything or invent anything that did not take place^b (1 = 100% fantasy, 7 = 100% real)
12. Rehearsal talk	Before imagining it today, I already talked about this event (1 = not at all, 7 = very often)
13. Rehearsal thought	Before imagining it today, I already thought about this event (1 = not at all, 7 = very often)
14. Importance	The imagined event will be important to my life (1 = not at all, 7 = to a very high degree)
15. Life story	I feel that this event will become a central part of my life story (1 = totally disagree, 5 = totally agree)
16. Identity	I feel that this event will become part of my identity (1 = totally disagree, 5 = totally agree)

Likelihood as assessed in the prospective condition.

Belief as assessed in the retrospective condition.

Procedure

Participants were recruited via Amazon’s Mechanical Turk, and they completed the questionnaire online using the survey platform Qualtrics. Following informed consent, the study consisted of four parts. First, participants were asked to provide demographic data (age, gender, years of education, and ethnicity). Second, participants filled in individual difference measures, which will be reported elsewhere. This was followed by a short mood assessment. Third, participants were randomly assigned to one of the two perspective conditions and were asked to generate five personal life goals (e.g., finish college). Fourth, participants received their reported personal life goals back one by one. For each reported personal life goal, they were asked to rate their qualities on five scales. Moreover, they were asked to generate a specific event that was closely related to each of the five corresponding personal life goals (e.g., graduation party) and to rate their phenomenological characteristics on several scales. To control for the potential effect of the experimental manipulation on mood, a second mood assessment was included before debriefing. The present article focuses on the specific events and their characteristics, while results related to the life goals and their qualities are reported in Roderer et al. (2022).

Coding procedures

Content coding and overlap with cultural life scripts

The thematic content of all events and their descriptions was coded according to the 23 life goal categories obtained by Roderer et al. (2022) plus 16 additional cultural life script categories of young Americans (Rubin et al., 2009). Following Berntsen and Rubin (2004), events that did not correspond to an event category or that were mentioned by fewer than 4% of all participants were coded as other. New categories were created for events that did not correspond to an existing category but were mentioned by at least 4% of all participants. This procedure led to 28 categories, including the category other (see Table 3). Coding was conducted by the first author and an independent scorer on a data sheet blind to perspective conditions. Interrater agreement was high (Cohen’s κ = .823) and disagreements were resolved by discussion.

Table 3.

Frequencies of event categories and their estimated age at occurrence (means and standard deviations) for prospective condition and retrospective condition.

Event Category	Prospective (n = 68)				Retrospective (n = 74)
	Mentions		Age		Mentions		Age
	n	%	M	SD	n	%	M	SD
Settle on career*	47	69.12	26.28	6.53	40	54.05	29.18	6.98
College*	36	52.94	25.03	4.19	32	43.24	21.84	3.07
Buying property	26	38.24	29.65	3.71	24	32.43	30.75	11.43
Having children*	23	33.82	29.26	3.48	42	56.76	30.12	9.39
Travelling	22	32.35	30.18	8.25	19	25.68	32.00	11.14
Marriage*	19	27.94	27.79	4.43	26	35.14	27.27	3.26
Financial success	17	25.00	34.59	10.43	18	24.32	30.00	8.49
Leisure/sport activities	16	24.53	28.25	4.36	15	20.27	26.87	6.30
Creating art	16	23.53	23.50	6.66	4	5.41	23.88	4.66
Buying possessions	13	19.12	24.69	3.59	12	16.22	25.00	8.55
Fall in love/relationship*	13	19.12	24.54	3.02	10	13.51	28.60	16.95
Financial stability	11	16.18	28.95	10.75	7	9.46	32.64	10.13
Getting engaged	9	13.24	26.22	3.60	9	12.16	27.56	3.81
Life enrichment	8	11.76	35.25	14.94	14	18.92	32.93	21.67
Health	8	11.76	22.38	2.13	3	4.05	31.67	15.89
Leave home/move*	7	10.29	26.57	6.65	9	12.16	23.44	6.95
Social connection	7	10.29	24.29	5.68	9	12.16	24.00	3.43
Support others	7	10.29	23.00	8.25	6	8.11	27.50	13.69
First job*	5	7.35	23.60	3.58	4	5.41	23.25	2.63
Gain knowledge	5	7.35	28.60	5.90	6	8.11	30.17	15.72
Retirement*	4	5.88	55.00	9.13	16	21.62	61.12	11.15
Get a pet	4	5.88	27.00	2.45	6	8.11	21.00	5.40
Doctoral degree	3	4.41	27.33	2.52	3	4.05	22.00	2.65
Grandchildren*	2	2.94	71.00	19.80	4	5.41	57.50	7.23
High school*	2	2.94	18.00	.00	9	12.16	17.11	4.76
Own death*	1	1.47	76.00	.	-	-	-	-
Begin driving*	1	1.47	24.00	.	4	5.41	18.50	3.00
Other	7	10.29	26.29	6.16	17	22.97	32.88	17.37
Total N	339				368

Life goal categories overlapping with a life script category by Rubin et al. (2009) are denoted with *.

To assess the overlap of events with the cultural life script, all events coded into a category matching the cultural life script (Rubin et al., 2009) were labelled “scripted events”; all other events were labelled “non-scripted events.”

Specificity of events

As an indicator of level of construal, event descriptions were coded for specificity. An event was categorised as “specific event” if its description included an event that happened or is expected to happen at a specific time and place within a 24-hr period (Williams et al., 2007). Events that described an extended period of time (e.g., a weekend), were categorical (e.g., taking classes), referred to a life period (e.g., going to college), or a semantic statement (e.g., I have travelled multiple countries) were coded as “non-specific events” (Williams et al., 2007). Events that could not be coded to these categories were excluded (n = 1). Coding was performed by the first author and an independent scorer on a data sheet blind to perspective conditions. Both scorers were trained in specificity coding on a set of 16.3% of the data by the third author (L.A.W.) and then rated the remaining events (n = 594) independently. Events the scorers disagreed on were separated and scorers rated those again, before solving disagreement by discussion. The agreement between the two scorers on data excluding the training data was good (Cohen’s κ = .832).

Statistical analyses

Prior to data analyses, data were screened to detect participants who did not meet the inclusion criteria specified a priori, including a completed questionnaire, passed attention checks, an indicated age between 18 and 25 years, and a minimum completion time of 8 min. Of 155 participants who finished the questionnaire and passed attention checks, participants who did not meet the age criteria (n = 1), who completed the questionnaire in less than 8 min (n = 7), or who took too long to complete the questionnaire (extreme value with completion time > third quartile + 3*IQR, n = 1) were excluded. Furthermore, participants who reported inappropriate answers in open text fields (e.g., copy and paste of answers or who mentioned fewer than three of five required life goals, n = 4) were excluded. Following this procedure, data of 142 participants were analysed.

To account for the hierarchical data structure, two-stage multi-level models were constructed, including 618 future events (Level 1) nested within 138 participants (Level 2). Thirteen separate models were conducted to predict the effect of perspective conditions on the outcome variables distance from present and the phenomenological characteristics. Based on acceptable internal consistency for both perspective conditions on future events (Cronbach’s α ranged from .77 to .90), four indices for items of phenomenological characteristics were created. The items mental time travel and p/reliving were grouped to an autonoetic experience index (see Ernst et al., 2018; Lehner & D’Argembeau, 2016) and the items importance, life story, and identity were combined to a self-narrative index (see Berntsen & Bohn, 2010), taking their difference in scale points into account. The items visual imagery and auditory imagery were combined to a visual–auditory index and the items rehearsal talk and rehearsal thought were averaged to a rehearsal index. Stepwise models were utilised and built up with a maximum likelihood estimation method, starting with a random intercept model (Model 1) to assess intraclass correlation. Next, a Level 1 predictor (distance from present: continuous; grand mean centred) and a Level 2 predictor (perspective condition: prospective = 0; retrospective = 1) were added as fixed effects to the model (Model 2). Both predictors were added for all outcome variables with the exception of the outcome variables distance from present, likelihood, and belief, which only included the Level 1 predictor. A random slope for the effect on the event-varying Level 1 predictor was added to Model 2 (covariance structure: variance components [VC]), if it significantly improved the model fit. For models, including a random slope on the Level 1 predictor, cross-level interactions between the two fixed effects were included to explain the variability in the distance from present slope across participants (Model 3). Model fit was assessed hierarchically by comparing the decrease in –2 log likelihood (–2LL) when adding parameters (e.g., random slopes or interactions). Tables 4 and 5 display the model with the best model fit for each outcome variable. The results from all multi-level model analyses can be found in supplementary materials and data are available on the Open Science Framework (OSF) (https://osf.io/m6aux/).

Table 4.

Model parameters and goodness of fit of multi-level model with best model fit.

	Model 2
	Distance	Valence-then	Valence-now	Intensity	Likelihood	Belief	Autonoetic experience index
Regression coefficients (fixed effects)
Intercept	10.01*** (0.77)	2.18*** (0.11)	2.07*** (0.11)	5.65*** (0.13)	5.52*** (0.12)	5.50*** (0.14)	5.41*** (0.15)
Distance		–0.01 (0.00)	–0.00 (0.01)	–0.00 (0.01)	–0.04** (0.01)	0.01 (0.01)	0.00 (0.00)
Condition	–3.41** (1.07)	–0.08 (0.16)	0.05 (0.16)	–0.24 (0.19)			–0.16 (0.21)
Interaction
Variance components (random effects)
Residual	95.13*** (6.21)	1.07*** (0.07)	0.98*** (0.07)	1.48*** (0.10)	0.98*** (0.09)	1.34*** (0.13)	1.13*** (0.07)
Intercept	17.74*** (5.22)	0.62*** (0.11)	0.63*** (0.11)	0.83*** (0.15)	0.74*** (0.17)	1.06*** (0.24)	1.21*** (0.18)
Slope			0.00 (0.00)	0.00 (0.00)	0.00 (0.00)
Model summary
Deviance	4,651.87	1,968.15	1,946.44	2,186.88	1,027.79	1,023.95	2,069.16
Number of estimated parameters	4	5	6	6	5	4	5

Standard errors are presented in parentheses.

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

Table 5.

Model parameters and goodness of fit of multi-level model with best model fit.

	Model 2				Model 3
	Visual–auditory index	Visual perspective	Spatial layout	Self-narrative index	Vividness	Rehearsal index
Regression coefficients (fixed effects)
Intercept	5.66*** (0.12)	4.72*** (0.20)	5.51*** (0.14)	5.53*** (0.13)	5.62*** (0.13)	5.15*** (0.15)
Distance	–0.00 (0.00)	–0.00 (0.01)	–0.01^† (0.01)	–0.00 (0.01)	–0.00 (0.01)	0.02 (0.01)
Condition	–0.15 (0.17)	–0.50^† (0.29)	–0.54** (0.20)	–0.08 (0.18)	–0.21 (0.19)	–0.31 (0.22)
Interaction					–0.03* (0.01)	–0.07*** (0.02)
Variance components (random effects)
Residual	1.18*** (0.08)	1.85*** (0.12)	1.86*** (0.12)	1.10*** (0.07)	1.27*** (0.09)	1.25*** (0.09)
Intercept	0.76*** (0.13)	2.49*** (0.36)	0.94*** (0.17)	0.82*** (0.13)	0.87*** (0.14)	1.33*** (0.20)
Slope			0.00 (0.00)	0.00 (0.00)	0.00^† (0.00)	0.00^† (0.00)
Model summary
Deviance	2,038.57	2,400.40	2,318.08	2,026.99	2,114.81	2,168.60
Number of estimated parameters	5	5	6	6	7	7

Standard errors are presented in parentheses.

†

p ⩽ .10. *p < .05. **p < .01. ***p < .001.

Results

A total of 707 events were reported (three people did not report an event). First, the content of all reported events and differences between the perspective conditions on event categories are examined. Second, the focus is on events expected to happen in one’s future. Here, the life span distribution and phenomenological characteristics of future events are outlined.

Content analyses

Overall, the five most frequently mentioned event categories were settle on career (12.3%), college (9.6%), having children (9.2%), buying property (7.1%), and marriage (6.4%). Of the 28 event categories, 12 event categories overlapped with categories of the cultural life script obtained by Rubin et al. (2009). These included settle on career, college, having children, marriage, fall in love/relationship, leave home/move, first job, retirement, grandchildren, high school, own death, and begin driving. Table 3 displays the frequencies of event categories for the prospective and the retrospective condition as well as the means and standard deviations of age at expected occurrence per event category. The most frequently reported event categories in the prospective condition were settle on career, college, and buying property, and in the retrospective condition having children, settle on career, and college. Chi-square tests of independence showed significant associations between perspective conditions and whether events were mentioned in the event categories having children, χ²(1) = 4.53, p = .033, V = .08; retirement, χ²(1) = 6.44, p = .011, V = .09; high school, χ²(1) = 3.97, p = .046, V = .08; and creating art, χ²(1) = 8.47, p = .004, V = .11. These indicate that participants in the retrospective condition mentioned significantly more events related to having children, retirement, and high school, and significantly fewer events related to creating art than the participants in the prospective condition. Where appropriate, t-tests for independent samples were conducted to compare the expected age of the event categories between perspective conditions. In the prospective condition, events related to college were expected to be experienced significantly later in life, t(63.77) = 3.60, p < .001, d = .86, and events related to settle on career significantly earlier in life, t(85) = –1.99, p = .049, d = .43, than in the retrospective condition. In all other event categories, the number of mentioned events and expected age of events did not differ significantly between the two perspective conditions, ps > .065.

Characteristics of future events

For all events, the distance from the present was calculated by subtracting the participant’s current age from the expected age of the event. For example, if a participant’s current age was 25 years and he/she expected a future event to happen at the age of 30 years, the distance from the present is 5 years. In total, 89 events were dated to have already taken place in the participant’s past, while 618 events were dated to occur in the same year as the participant’s current age or in the future. These are considered future events. In the prospective condition, 323 future events were reported and 295 future events in the retrospective condition. The following analyses focus on future events.

Lifespan distribution of future events

The life span distribution of future events for both perspective conditions, including error bars for standard errors, is displayed in Figure 1. It illustrates that most future events were expected to occur between the age of 20 and 30, while only a few were dated to occur later in life with a small peak around age 60. However, the peak of future events in young adulthood was lower and the small increase of events in later life was larger in the retrospective than the prospective condition.

Figure 1.

Life span distribution of future events by expected age of occurrence.

A series of independent sample t-tests showed that participants in the retrospective condition dated their third and fifth future events (ps < .014), but not their first, second, and fourth (ps > .202), significantly further into the future than participants in the prospective condition. As illustrated in Figure 2, most future events were dated to occur within 10 years into the future, whereas only a few were dated to occur further into the future. Events mentioned as the fifth event were dated more frequently in the further distance in both conditions, forming a bump around 40 years from the present. However, this bump was more pronounced in the retrospective condition.

Figure 2.

Distribution of future events according to the distance from present for each of the five mentioned future events for the prospective condition (left panel) and retrospective condition (right panel).

Multi-level modelling indicated that perspective condition significantly predicted the distance of future events from the present, F(1, 121.07) = 10.09, p = .002 (see Table 4, Column 1). The intercept b = 10.01 symbolises that participants in the retrospective condition dated their future events on average 10.01 years into the future, whereas participants in the prospective condition expected their future events to happen on average 6.60 years into the future. Thus, future events were expected to occur 3.41 years closer to the present in the prospective condition than in the retrospective condition. These findings support the hypothesis that retrospectively reported future events are expected later in life, leading to a wider distribution across the life span, than future events reported prospectively.

Cultural life script overlap and specificity

Testing the hypothesis that retrospectively reported future events draw to a stronger extent on schematised knowledge, a Chi-square test of independence showed no association between the two perspective conditions and whether future events were scripted or not, χ²(1) = 1.86, p = .173, V = .06. This indicates that the frequency of future events corresponding to a cultural life script category was similar in the retrospective condition (52.5%) and the prospective condition (47.1%).

Besides cultural life script overlap, the event specificity provides an additional indicator for level of construal. The participants in the prospective condition reported on average a lower percentage of future events that were specific (M = 62.80, SD = 34.27) than participants in the retrospective condition (M = 68.98, SD = 33.77). This difference, –6.18, was not significant, BCa 95% CI [–17.64, 4.69], t(136) = –1.07, p = .288, d = .18. At the same time, participants described their future events on average with a similar number of words in the prospective condition (M = 40.40, SD = 25.47) and the retrospective condition, (M = 38.15, SD = 25.95; mean difference_pro-retro = 2.24, BCa 95% CI [–6.85, 11.18], t(136) = 0.51, p = .610, d = .09).

Phenomenological characteristics of future events

Multi-level modelling examined the effect of perspective condition and temporal distance on the phenomenological characteristics of future events. Contrary to our hypothesis, multi-level modelling showed that neither perspective condition nor the predictor distance from present affected the phenomenological characteristics of future events on belief, autonoetic experience index, visual–auditory index, visual perspective, self-narrative index, valence-then, valence-now, and intensity (see Tables 4 and 5).

Testing the hypothesis that changes in psychological distancing affect cognitive feelings, the effect of the predictor distance from present on likelihood in the prospective condition and belief in the retrospective condition were calculated separately to account for differences in the wording of the scales. In the prospective condition, the predictor distance from present was significantly negatively associated with likelihood, F(1, 25.37) = 12.76, p = .001, b = –.04 (see Table 4, Column 5), indicating that future events dated in the further distance from the present were rated as less likely to occur. In the retrospective condition, the predictor distance from present was not significantly associated with belief, F(1, 269.51) = 1.15, p = .285, b = .01 (see Table 4, Column 6). This indicates that the rating of belief remained high also for events dated to occur further into the future from the present self.

Focusing on the hypothesis that sensory-perceptual details are affected by perspective condition, the predictor perspective condition was significantly negatively associated with the item spatial layout, F(1, 135.49) = 7.09, p = .009, b = –.54, suggesting that participants in the prospective condition rated the spatial layout on average lower than participants in the retrospective condition (see Table 5, Column 3). Moreover, a Condition × Distance from Present interaction significantly predicted the outcome vividness, F(1, 74.74) = 4.35, p = .041 (see Table 5, Column 5). Follow-up analyses indicated that distance from present was negatively associated with vividness in the prospective condition, b = –.03, t(28.95) = –2.75, p = .010, but not in the retrospective condition, b = –.00, t(37.59) = –0.17, p = .869. This indicates that with increasing distance from the present, vividness decreased in the prospective condition and remained stable in the retrospective condition.

Exploratory multi-level modelling indicated that rehearsal index was significantly predicted by a Condition × Distance from Present interaction (see Table 5, Column 6). Follow-up analyses indicated that distance from present was negatively associated with rehearsal index in the prospective condition, b = –.08, t(39.92) = –4.27, p < .001, but not in the retrospective condition, b = .01, t(19.52) = 1.77, p = .092. This interaction suggests that with increasing distance from the present, the rehearsal ratings decreased in the prospective condition and remained stable in the retrospective condition.

Discussion

This study provides novel evidence that simulating the future retrospectively as if it has been already experienced leads to a wider temporal distribution of events across the life span compared with looking forward in time, supporting the assumption that retrospective future thinking serves as a form of psychological distancing. In line with previous research (Berntsen & Bohn, 2010; D’Argembeau & Van der Linden, 2004; Özbek et al., 2017), the characteristics of prospective events diminished the further they were in the future. However, this was not the case for retrospective events. Retrospective events were distributed more widely across the lifespan and were characterised not by high-level construal information as predicted but by low-level construal information. Participants reported high levels of belief that events would occur and imagined events vividly and spatially detailed, independent of temporal distance. One explanation for these findings might be that engaging in retrospective future thinking allows individuals to construct temporally distant events that are held as psychologically close on the dimension of hypotheticality, supporting the construction of temporally distant yet sensorial near or low-level construal events associated with high levels of belief in occurrence.

Shifting from our normal mode of prospective future thinking to a retrospective perspective affected the temporal distribution of future events. In both perspective conditions, most future events were expected in young adulthood and only few in later life. However, in the retrospective condition, future events were distributed more widely across the life span, forming a lower peak in young adulthood and an additional small bump around age 60. Complementing this wider distribution, engaging in retrospective future thinking prompted participants to think more often about events related to having children, high school, and retirement, mirroring the bump around age 60. Furthermore, college-related events were expected to happen earlier, and events related to settling on a career later in life compared with the prospective condition. In addition, retrospective future events were dated further into the future than prospective future events. The findings suggest that retrospective future thinking expands the time interval used when structuring one’s future, supporting the hypothesis that retrospective future thinking serves as a form of psychological distancing that leads to changes in temporal predictions about future events (Trope & Liberman, 2010). An alternative explanation for the obtained effects on the wider temporal distribution of events might be an anchoring effect driven by the focus on a lifetime of 100 years in the instructions in the retrospective condition compared with an unspecified interval of a whole lifetime in the prospective condition. Future research might systematically vary either the age interval specified in retrospective future thinking or highlight an equal interval in prospective future thinking to investigate the mechanisms underlying the observed effects of shifting temporal perspective (see also Roderer et al., 2022).

Contrary to our hypothesis, retrospective future thinking was not associated with higher construal level information: In both perspective conditions, around half of all events corresponded to the cultural life script, the majority of events were specific, and events were perceived as equally important. These findings are in line with previous research on prospective future thinking (Berntsen & Bohn, 2010; D’Argembeau & Mathy, 2011) and demonstrate stability in our ability to think about the future: Even when twisting the angle from which one views the future by imagining what might happen from a retrospective perspective, highly robust aspects of prospective cognition, such as event specificity and the use of the schematised cultural life script remain stable. Another interpretation of this null finding might be that personal future events are, by definition, highly concrete constructs that are construed in low-level terms. Related research has demonstrated that retrospectively reported future goals rely to a greater extent on schematised knowledge and remain important across the lifespan compared with prospectively reported goals (Roderer et al., 2022). In hierarchal models of goals, events represent subordinate, low-level actions that are related to superordinate, high-level goals (Austin & Vancouver, 1996; Carver & Scheier, 2000). Therefore, changes in temporal perspective may lead to fewer variations in event representations compared with goal representations. Similarly, instructions to simulate specific events in close association to one’s life goals might have prompted participants to think about specific and important events, minimising potential variability on these measures between perspective conditions. A broader conceptualisation of more general future events, for example, in response to word cues (Crovitz & Schiffman, 1974) might maximise the possibility of detecting differences in construal level between perspective conditions.

Phenomenological characteristics of event representations were affected by perspective condition, however, not in the hypothesised direction. Replicating the effect of temporal distance on prospective future events in previous research (e.g., Berntsen & Bohn, 2010), temporally distant events in the prospective condition were perceived as less likely than near events and compared with events in the retrospective condition as less vivid and less rehearsed the later they were expected to occur. This supports the view of CLT that low-level construal information decreases as temporal distance increases during natural future prospection (Trope & Liberman, 2010). In contrast, in the retrospective condition, belief in event occurrence, vividness, and frequency of rehearsal remained high for events more temporally distant to the current self. Furthermore, retrospective events were associated with a clearer spatial layout than prospective events. The findings contradict to some extent the hypothesis that psychologically distant temporal perspectives should lead to future event representations construed in higher-level terms, associated with more schematised knowledge, personal significance, lower specificity, and fewer cognitive feelings and sensory-perceptual details (Trope & Liberman, 2010).

While the reliance on schematised knowledge observed when thinking about important goal-related future events (across both perspective conditions) and effects of temporal distance on the characteristics of future thinking in the prospective condition provide evidence to support CLT, it is relevant to account for the findings related to the retrospective condition which seemingly contradict CLT. Within the context of CLT, it is possible that increasing temporal distance through retrospective future thinking also leads to contrasting changes in other dimensions of psychological distance, such as hypotheticality or self (Trope & Liberman, 2010). Considering retrospectively reported events as remembered from the perspective of the 100-year-old self, the belief that events have taken place is logical and that events are characterised as more detailed, vivid, and more often rehearsed replicates research contrasting past and future events (Berntsen & Bohn, 2010). For a direct comparison between perspective conditions, future studies can enhance the alignment of the instructions between perspective conditions. First, the belief in occurrence should also be assessed in the prospective condition, as in the retrospective condition the belief that an event has taken place as remembered, but not its likelihood can be assessed (Ernst & D’Argembeau, 2017). Second, the expectation (un)certainty of both conditions might be further aligned by adjusting the wording of the instructions in the prospective condition (e.g., “life experiences which will take place” instead of “life experiences which may take place”) to match the expectation certainty of the instructions given in past tense in the retrospective condition.

Alternatively, the construction of details for temporally distant events in the retrospective condition might be due to the increasing closeness to one’s hypothetical 100-year-old self. Defining the current self as the baseline to examine effects of temporal distance in both perspective conditions might have created an asymmetry between psychologically near perspective in the prospective condition (current self) and the retrospective condition (100-year-old self). The findings suggest that engaging in retrospective future thinking leads to future event representations that are characterised as less hypothetical and represented by concrete low-level construals as already having taken place. Rather than representing a psychologically more distant temporal perspective leading to high-level construals, retrospective future thinking seems to affect temporal distance but at the same time keeps even temporally distant events close on other psychological dimensions, such as hypotheticality and the self. These results do not contradict CLT but suggest a more nuanced picture of the interrelation between the different dimensions of psychological distance. Retrospective future thinking may serve as a form of psychological distancing that affects several dimensions and their interrelation by increasing temporal distance while at the same time decreasing the psychological distance of the self or hypotheticality. If replicated, such findings may challenge the general assumption of CLT that a straightforward bidirectional relation exists between dimensions of psychological distance (Trope & Liberman, 2010).

Simulating one’s future retrospectively differs from our natural mode of future thinking. Therefore, intentionally engaging in retrospective future thinking might be a valuable tool to simulate what might happen in the distant future. Prospective simulation has previously demonstrated beneficial effects on saving (Macrae et al., 2017), reducing temporal discounting promoting healthy behaviour (Dassen et al., 2016), and enhancing pro-environmental behaviour (Lee et al., 2020). Retrospective future thinking might be especially beneficial when thinking about distant future events in a psychologically close way. It could be applied to facilitating the appraisal and planning of long-term goals relevant for family planning and retirement savings, and also to evaluate upcoming decisions and consequences of taking (no) action (Valenti et al., 2011) as well as a coping strategy by setting current tasks into relation to one’s lifetime.

This study provides evidence that while the content, organisation, and phenomenology of future event simulations are robust, retrospective future thinking influences psychological distance by affecting the temporal distribution of autobiographical future events. While changes in temporal perspective did not influence the reliance on schematised knowledge, specificity, and perceived meaning, retrospective future thinking might allow us to keep distant events psychologically close on other dimensions of psychological distance by construing them as highly likely to occur and more detailed, vivid, and more often rehearsed than prospective future thought.

Supplemental Material

sj-pdf-1-qjp-10.1177_17470218221126471 – Supplemental material for Retrospective future thinking: Keeping distant personal future events mentally close

Supplemental material, sj-pdf-1-qjp-10.1177_17470218221126471 for Retrospective future thinking: Keeping distant personal future events mentally close by Ayleen Roderer, Annette Bohn and Lynn Ann Watson in Quarterly Journal of Experimental Psychology

Footnotes

Acknowledgements

The authors thank Daniel Munkholm Møller for his support with data collection, Kaare Bro Wellnitz for his guidance with statistical analysis, and Ansley Raines and Johanna Hein for their assistance with data scoring.

Author contributions

Ayleen Roderer: conceptualization, methodology, data curation, investigation, project administration, visualization, formal analysis, writing – original draft.

Annette Bohn: conceptualization, methodology, validation, writing – review & editing.

Lynn Ann Watson: conceptualization, methodology, validation, statistical analysis, writing – review & editing.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Danish National Research Foundation (DNRF89).

ORCID iD

Ayleen Roderer

Supplementary material

The supplementary material is available at qjep.sagepub.com.

Notes

References

Addis

D. R.

Wong

A. T.

Schacter

D. L.

(2008). Age-related changes in the episodic simulation of future events. Psychological Science, 19(1), 33–41. https://doi.org/10.1111/j.1467-9280.2008.02043.x

Austin

J. T.

Vancouver

J. B.

(1996). Goal constructs in psychology: Structure, process, and content. Psychological Bulletin, 120(3), 338–375. https://doi.org/10.1037/0033-2909.120.3.338

Berntsen

Bohn

(2010). Remembering and forecasting: The relation between autobiographical memory and episodic future thinking. Memory & Cognition, 38(3), 265–278. https://doi.org/10.3758/MC.38.3.265

Berntsen

Rubin

D. C.

(2004). Cultural life scripts structure recall from autobiographical memory. Memory & Cognition, 32(3), 427–442. https://doi.org/10.3758/BF03195836

Berntsen

Rubin

D. C.

(2006). The centrality of event scale: A measure of integrating a trauma into one’s identity and its relation to post-traumatic stress disorder symptoms. Behaviour Research and Therapy, 44(2), 219–231. https://doi.org/10.1016/j.brat.2005.01.009

Carver

C. S.

Scheier

M. F.

(2000). Chapter 3—On the structure of behavioral self-regulation. In Boekaerts

Pintrich

P. R.

Zeider

(Eds.), Handbook of self-regulation (pp. 41–84). Academic Press. https://doi.org/10.1016/B978-012109890-2/50032-9

Crovitz

H. F.

Schiffman

(1974). Frequency of episodic memories as a function of their age. Bulletin of the Psychonomic Society, 4(5), 517–518.

D’Argembeau

Mathy

(2011). Tracking the construction of episodic future thoughts. Journal of Experimental Psychology: General, 140(2), 258–271. https://doi.org/10.1037/a0022581

D’Argembeau

Van der Linden

(2004). Phenomenal characteristics associated with projecting oneself back into the past and forward into the future: Influence of valence and temporal distance. Consciousness and Cognition, 13(4), 844–858. https://doi.org/10.1016/j.concog.2004.07.007

10.

D’Argembeau

Van der Linden

(2006). Individual differences in the phenomenology of mental time travel: The effect of vivid visual imagery and emotion regulation strategies. Consciousness and Cognition, 15(2), 342–350. https://doi.org/10.1016/j.concog.2005.09.001

11.

D’Argembeau

Van der Linden

(2012). Predicting the phenomenology of episodic future thoughts. Consciousness and Cognition, 21(3), 1198–1206. https://doi.org/10.1016/j.concog.2012.05.004

12.

Dassen

F. C. M.

Jansen

Nederkoorn

Houben

(2016). Focus on the future: Episodic future thinking reduces discount rate and snacking. Appetite, 96, 327–332. https://doi.org/10.1016/j.appet.2015.09.032

13.

de Vito

Gamboz

Brandimonte

M. A

. (2012). What differentiates episodic future thinking from complex scene imagery? Consciousness and Cognition, 21(2), 813–823. https://doi.org/10.1016/j.concog.2012.01.013

14.

Ernst

D’Argembeau

(2017). Make it real: Belief in occurrence within episodic future thought. Memory & Cognition, 45(6), 1045–1061. https://doi.org/10.3758/s13421-017-0714-3

15.

Ernst

Philippe

F. L.

D’Argembeau

(2018). Wanting or having to: The role of goal self-concordance in episodic future thinking. Consciousness and Cognition, 66, 26–39. https://doi.org/10.1016/j.concog.2018.10.004

16.

Johnson

M. K.

Foley

M. A.

Suengas

A. G.

Raye

C. L.

(1988). Phenomenal characteristics of memories for perceived and imagined autobiographical events. Journal of Experimental Psychology: General, 117(4), 371–376.

17.

Lee

P.-S.

Sung

Y.-H.

C.-C.

L.-C.

Chiou

W.-B.

(2020). Using episodic future thinking to pre-experience climate change increases pro-environmental behavior. Environment and Behavior, 52(1), 60–81. https://doi.org/10.1177/0013916518790590

18.

Lehner

D’Argembeau

(2016). The role of personal goals in autonoetic experience when imagining future events. Consciousness and Cognition, 42, 267–276. https://doi.org/10.1016/j.concog.2016.04.002

19.

Macrae

C. N.

Mitchell

J. P.

Golubickis

N. S. P.

Sherlock

Parlongo

Simpson

O. C. M.

Christian

B. M.

(2017). Saving for your future self: The role of imaginary experiences. Self and Identity, 16(4), 384–398. https://doi.org/10.1080/15298868.2016.1264465

20.

McDermott

K. B.

Wooldridge

C. L.

Rice

H. J.

Berg

J. J.

Szpunar

K. K.

(2016). Visual perspective in remembering and episodic future thought. Quarterly Journal of Experimental Psychology, 69(2), 243–253. https://doi.org/10.1080/17470218.2015.1067237

21.

Özbek

Bohn

Berntsen

(2017). Imagining the personal past: Episodic counterfactuals compared to episodic memories and episodic future projections. Memory & Cognition, 45(3), 375–389. https://doi.org/10.3758/s13421-016-0671-2

22.

Roderer

Bohn

(in press). Retrospective future thinking as a novel method to imagine the future: Remembering autobiographical events from the perspective of the future self. Memory. https://doi.org/10.1080/09658211.2022.2120997.

23.

Roderer

Watson

L. A.

Bohn

(2022). Remembering future life goals: Retrospective future thinking affects life goal qualities. Acta Psychologica, 226, Article 103582. https://doi.org/10.1016/j.actpsy.2022.103582

24.

Rubin

D. C.

Berntsen

Hutson

(2009). The normative and the personal life: Individual differences in life scripts and life story events among USA and Danish undergraduates. Memory, 17(1), 54–68. https://doi.org/10.1080/09658210802541442

25.

Rubin

D. C.

Schrauf

R. W.

Greenberg

D. L.

(2003). Belief and recollection of autobiographical memories. Memory & Cognition, 31(6), 887–901. https://doi.org/10.3758/BF03196443

26.

Schacter

D. L.

Benoit

R. G.

Szpunar

K. K.

(2017). Episodic future thinking: Mechanisms and functions. Current Opinion in Behavioral Sciences, 17, 41–50. https://doi.org/10.1016/j.cobeha.2017.06.002

27.

Szpunar

K. K.

(2010). Episodic future thought: An emerging concept. Perspectives on Psychological Science, 5(2), 142–162. https://doi.org/10.1177/1745691610362350

28.

Szpunar

K. K.

Spreng

R. N.

Schacter

D. L.

(2014). A taxonomy of prospection: Introducing an organizational framework for future-oriented cognition. Proceedings of the National Academy of Sciences—PNAS, 111(52), 18414–18421. https://doi.org/10.1073/pnas.1417144111

29.

Trope

Liberman

(2003). Temporal construal. Psychological Review, 110(3), 403–421. https://doi.org/10.1037/0033-295X.110.3.403

30.

Trope

Liberman

(2010). Construal-level theory of psychological distance. Psychological Review, 117(2), 440–463. https://doi.org/10.1037/a0018963

31.

Valenti

Libby

L. K.

Eibach

R. P.

(2011). Looking back with regret: Visual perspective in memory images differentially affects regret for actions and inactions. Journal of Experimental Social Psychology, 47(4), 730–737. https://doi.org/10.1016/j.jesp.2011.02.008

32.

Vasquez

N. A.

Buehler

(2007). Seeing future success: Does imagery perspective influence achievement motivation? Personality and Social Psychology Bulletin, 33(10), 1392–1405. https://doi.org/10.1177/0146167207304541

33.

Williams

J. M. G.

Barnhofer

Crane

Herman

Raes

Watkins

Dalgleish

(2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 133(1), 122–148. https://doi.org/10.1037/0033-2909.133.1.122

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

0.33 MB