The present research

Why is there a perceived disassociation between the current and the future self? The present research aims to (1) identify major components of the perceived relationship between the current and future self, (2) extend and develop empirical measures to address this relationship, and (3) investigate whether and how these components uniquely and jointly relate to important outcomes. To address our aims, we conducted a literature review of the psychological research on the perception of the future self. Notably, Hershfield [4] provided a detailed review of the relevant literature. However, since his review in 2011, this body of literature has gained vast momentum. As such, we expand on his review including articles in peer-reviewed journals that (a) were published before January 2019 and (b) empirically measured people’s view of their future self and/or their view of the relationship between the current and future self. The January 2019 cutoff date was used because this is when the literature review was conducted. However, subsequent articles that were published that are particularly relevant for our research aims are still summarized below when appropriate. Our search using PsycInfo and the keyword “future self” yielded 363 articles of which 71 met these review criteria [5–74].

What makes up perceptions of the future self?

Consistent with Hershfield [4, 75], our literature review shows that most of the research in this area focuses on one or two of the three following components in people’s conception of their future self: (1) the perceived similarity and connectedness between the current and future self, (2) the degree of vividness when the future self is imagined, and (3) the degree of positivity felt toward the future self. It should be noted that researchers have used different labels in referring to these components. Below, we summarize the extant research for each of the three components, discuss the appropriateness of the labels given the items used to assess them, and then suggest novel labels wherever needed to provide clarity in the conceptualization of the construct.

The relatedness component: Similarity and connectedness to the future self

Twenty-six of the 71 articles in our review included a measure of similarity (and/or connectedness). Much of this research on perceptions of the future self draws on the theoretical work of the philosopher Derek Parfit [76, 77]. According to Parfit [77], the self can be conceptualized as a collection of selves representing the individual at different time points, and sharing a certain degree of psychological similarities and connectedness with one another. This degree of connection and/or similarity between the present and future self is a frequently studied component in recent research on the future self.

Twenty of the 26 articles in our review used a version of the 2-item Future Self-Continuity Measure [21], which was adapted from the Inclusion of Other in the Self Scale [78]. The two items ask about the perceived connectedness and similarity between the present and future self. Participants select one of seven pairs of circles that overlap to an increasing degree, in which one circle represents the current self and the other represents the distal future self. A greater degree of overlap between the two circles indicates greater perceived similarity and connectedness between the current self and the future self. Notably, this measure asks participants to consider their future self at a specific point in time (e.g., 10 years in the future) but does not ask participants to consider the future self in a specific domain (e.g., career, social relationships). In other words, participants express their self-view in a relatively decontextualized (or global) way, thus allowing them to define themselves freely as they see fit. Individuals who score highly on this measure are more likely to engage in future-oriented behavior, such as increased long-term health behaviors [64], greater monetary saving, less-myopic temporal discounting tendencies [21], and decreased acceptance of unethical behaviors [29].

Some researchers have used a modification of the Inclusion of Other in the Self Scale [78] that focuses on the connectedness (i.e., not similarity) to the future self, especially in terms of core identity and psychological characteristics [6, 79–82]. Similarly, research from this perspective finds that connectedness to future selves affects how an individual chooses between rewards that occur at different points in time (i.e., intertemporal choices) [6]. If an individual feels less connected to a self that is more distant in the future, he or she is less likely to adopt behaviors or make choices that would benefit that distant future self (e.g., saving for retirement) and are more likely to engage in behaviors that benefit the present self (e.g., spending all disposable income on short-term consumption). The two labels most commonly used in the literature—similarity and connectedness—do not fully capture both aspects of the component. Therefore, we propose the label “relatedness” which encompasses both critical aspects of similarity and connectedness. According to the Self-Determination Theory [83], relatedness is one of the innate psychological needs critical to self-motivation. In that literature, relatedness entails feelings of connection and similarity with close others. These others act as a role model and reinforce self-motivation by representing what the future self could look like. For example, perceived relatedness between students and their role models such as teachers or parents predicts students’ drive for academic achievement [84, 85]. Considering the future self as a special case of the other, we suggest that relatedness represents both perceived connection and similarity between the current and future self. For the remainder of this article, we henceforth use the term relatedness to refer to the component of the future self that represents individual differences in perceived similarity and connectedness.

The positivity component: Liking and a positive valence towards the future self

Our review found that 35 articles included at least a measure of perceived positivity or negativity of the future self. Akin to the existing literature on the relatedness component, the literature on positive views of the future self has used several different labels. For example, some studies measure “self-prediction positivity” [68], “perceived valence of future time” [44], or “positivity of the future self” [4]. Despite differences in the labels, this component captures the degree to which individuals see their future self as desirable and positive. As such, we agree with Hershfield’s conclusion that the label “Positivity” clearly represents this component [86].

In 13 articles, researchers used a “decontextualized/global” measure of the future-self positivity (i.e., participants indicated how positive they imagine their distal future would be or how much they like their future self in general). The remaining 22 articles included measures of the perceived positivity (or negativity) of future self in terms of specific attributes (e.g., calm, confident, positive, fearful, unhappy) [56, 67] or domains (e.g., future social relationships, leisure activities, finances, and work) [38]. In 15 of these 22 articles, the analyses focused on an overall composite of positivity of future self by averaging participants’ positivity ratings across attributes/domains, or formed separate aggregates for positivity/hoped-for future self and negativity/feared future self [45, 74]. Consistently, previous findings show that positivity of future self predicted greater preparation or beneficial behavior for the future [38] and decreased adolescent delinquency [87, 88].

We identified three articles in which researchers used an alternative to self-report for assessing positivity. In these articles, positive and negative future selves were assessed with a me/not me task in which participants were exposed to positive (e.g., wise, calm, confident) and negative (e.g., dull, fearful, stubborn) adjectives of the future self [55, 65, 67].

The vividness component: Clarity and ease of imagining the future self

The final component, vividness of the future self, refers to the extent to which individuals conjure a clear image of themselves in the future. In our review of the literature, we found 12 articles in which researchers measured vividness of the future. In the majority of the studies, researchers used self-report measures to assess clarity or vividness of images of the future self [8, 51]. In contrast, two studies used a qualitative approach, coding vividness from participants’ brief descriptions of their future [18, 41]. In terms of the domain-specific versus global focus, seven articles assessed vividness of a global, decontextualized future self [8, 69] and five assessed vividness in specific future domains, including retirement) [20] and best possible self [57]. Examples of domain-specific measures of vividness assess the clarity of specific possible future events (e.g., “the next time you meet a friend”) [89].

Much of the research on the vividness component focuses on manipulating the clarity of future images to experimentally increase the degree of relatedness between the present and future self. For instance, interacting with computer-aged renderings of the self leads individuals to feel more connected between their current and future selves [90], thus suggesting that the extent to which individuals can vividly imagine their life in the future is related to their ability to feel psychologically connected to that future self. This increased level of future self-vividness then leads to behaviors indicative of a more long-term temporal focus, including reduced levels of delinquent behavior [91, 92] and temporal discounting [90]. Similarly, research on episodic future thinking has shown that a higher capacity to imagine events that may occur in the future has a variety of beneficial outcomes. These include increased likelihood of waiting for a larger, delayed reward and more effective emotion regulation skills [89, 93].

Researchers studying vividness have used several labels for this component including “vivid mental imagery” [7], “episodic future thinking” [94], “future clarity” [51], and “vividness of the future self” [4]. Although these programs of research use diverse labels, their primary focus is on measuring the clarity of mental images of the future self. As such, the vividness label proposed by Hershfield [4] clearly captures the essence of this component.

Labeling the triadic model

Hershfield [4] proposed the future self-continuity model to encompass the three components of the future self. However, we suggest that the construct label, future self-continuity, presents issues that might lead to a lack of clarity in this literature. Below, we (1) discuss a theoretical issue associated with the use of future self-continuity, (2) articulate why this label does not fully reflect the proposed construct, and (3) provide an alternative label—future self-identification.

Hershfield’s “future self-continuity” label is derived from Parfit’s [76, 77] concept of continuity between temporal selves. According to Parfit, continuity is the overall strength of overlapping links between all sequential pairs of selves. Contrastingly, the three components in Hershfield’s [4] model and the extant literature concern the direct relation between the current and a future self at a given point in time. In other words, the three components only include perceptions of the relatedness, vividness, and positivity of a particular future self. Consequently, an individual may feel highly related, positive, and vivid about a particular future self but may or may not feel a high level of continuity between his or her entire chain of temporal selves.

Additionally, the future self-continuity label focuses exclusively on relatedness between temporal selves, but overlooks the positivity and vividness components. While Parfit [77] discussed vividness and positivity as elements of the future self, they are not necessary to his notion of continuity. Under his conceptualization of continuity, the current and future self may be perceived as highly psychologically continuous without a highly positive or vivid view of the self in the future. An individual with chronic pain may feel highly connected or continuous with a negative future self. Similarly, upon being laid off from a job, a person may fail to see a clear vision of what the future holds despite a strong sense of connection to that future self. This distinction is important because the three components are conceptualized as related but distinct [4]. Consequently, the existing label fails to reflect distinctions in the valence and visualization of the future and does not lend itself to representing all three proposed components under one unified construct.

In light of these issues, we suggest a new label—future self-identification—to represent the overall construct that includes the three components (relatedness, vividness, and positivity). According to the American Psychological Association [95], “identification is the process of associating the self closely with other individuals.” The strength of identification plays an important role in decision-making, as those with higher identification are more likely to consider the shared connection and values with those they identify with and less likely to act based on self-interest [96]. Previous research found that individuals who indicated higher group identification were more likely to stay committed to the group and less likely to leave and join another [97]. Individuals were more likely to contribute to a failing group only when group identification was salient as opposed to personal identification [98].

Although targets of identification are generally others or groups, we suggest the process of identification applies to the perception of the future self. This aligns with the analogy of the “future-self-as-other”. Of course, a future self is still physically tied to a current self, but psychologically there are parallels between how people treat future selves and how they treat other people. Research utilizing fMRI techniques found that making decisions about a distal future self did not elicit the same activation in the medial prefrontal cortex and rostral anterior cingulate as did making decisions about the current self [2]. These cortical midline structures are involved in the processing of self versus other information [99], though there is debate about the precise neural location of this similar processing [3]. The difference in activation between these brain regions for current-self-related and future-self-related decisions also predicted subsequent measures of temporal discounting [2]. Recently, Molouki and Bartels [100] found that there were parallels between how people treat future selves and others, such as similar factors underlying the allocation of resources to both future selves and other individuals. However, this research also showed that people generally allocate more resources to future selves than they allocate to others, suggesting that future selves are not treated exactly like other people.

We propose that people who feel more related to the future self, have more positive feelings towards it, and can imagine it more clearly, will more strongly identify with the future self than those who do not. These elements closely align with the components that impact identification within the group identification literature. Specifically, Henry, Arrow, and Carini [101] argued that group identification will be stronger in groups with greater intra-individual similarity [102, 103], more positive feelings towards group members, and sharing more vivid goals [104, 105]. Given the parallel between the process of identification and the perceived relation between the current and future self, we suggest a departure from previous research by assigning a new label—future self-identification—as the overarching construct encompassing the three most common components of future self in the literature. Together, the three components of future self-identification tap into the extent to which an individual relates their current self to their future self, how clearly they envision their future self, and how much they like their future self.

Relationships among the three components

At the time of our review, no research had empirically examined the relationships among the three components of future self-identification and few researchers even examined the relation between two of the components in one study. However, results of the existing research suggest positive relationships among the three components. For example, one study found positive associations between measures of liking the future self and measures of feeling similar to the future self, suggesting that the positivity and relatedness components may be positively related [21]. This finding is consistent with research on Temporal Self-Appraisal Theory showing that individuals often report feeling closer to favorable than unfavorable future selves [73, 106, 107]. Moreover, a recent study found that experimentally manipulating the valence of expected change in personal characteristics (e.g., personality, preferences, morality, experiences) affected perceived continuity between current and future selves [108]. Specifically, positive changes in the future produced higher future self-continuity, whereas negative changes were particularly disruptive to perceived self-continuity with future selves. These results further suggest a relationship between the valence of perceived future selves and degrees of relatedness.

Considering the relationship between relatedness and vividness, experimental studies show that exposure to computer-aged images (i.e., a manipulation that presumably induced vividness of the future self) enhanced perceived relatedness to the future self [40, 90]. Additionally, individuals who could vividly imagine their future self were more likely to relate to the future self, suggesting that the vividness and relatedness components may be positively related [8]. This recent study also found that individuals higher in future self-vividness experienced steeper increases in relatedness to the future self over time. However, another experimental study did not find a significant relationship between vividness and relatedness [49]. This research reported that, after writing about their possible future selves, participants’ ratings of vividness of their future selves did not predict ratings of the similarity of their future self. Although vividness of future self ratings was a negative predictor of number of undesirable words generated for the future self, they did not predict the number of desirable words for the future self [49]. Thus, the findings from studies of the relationship between vividness and positivity are inconclusive.

Another notable observation of our review is that the majority of studies on the validity of future self-identification focused on predicting temporal discounting [2, 6, 29], and with few exceptions, other important outcomes have received less attention. For instance, only three studies examined achievement outcomes. These studies found that relatedness to future self positively predicted college grade point averages and negatively predicted procrastination [5, 7], and positivity of future self predicted higher academic motivation [55].

Further, several studies examined one of the components of future self-identification and its relations with intrapsychic outcomes. These studies found that (a) relatedness predicted more positive and less negative affect [7], more psychological well-being [14], and less depression [67]; (b) positivity predicted less anxiety and depression [19, 31, 67], more life satisfaction and state self-esteem [33, 39], and higher quality of life [64]; and (c) vividness predicted higher life satisfaction, greater positive affect, and less negative affect and depression [49].

A gap in previous research involves testing the independent predictive validity of the three components. One exception examined two of the components, finding that vividness, but not relatedness, predicted more positive and less negative affect [51]. Given the positive relationships that have been observed among the three components, it is important to test the independent predictive validity of each component.

Since the time of our literature review, Sokol and Serper [109] tested a measure of all three components within a single framework. A number of interesting findings emerged from this research. First, the results of their confirmatory factor analysis show that the three components are positively correlated ranging from .46 to .59 and they are independent factors, supporting a triadic model of this construct. Second, each component was significantly correlated with one or more indicators of less temporal discounting (e.g., consideration of future and immediate consequences) and intrapsychic adjustment (e.g., life satisfaction and hopelessness). Third, comparisons of the strength of the correlations between each component suggest that positivity was the strongest predictor of consideration of the future consequences. However, the three components did not differ in terms of their correlations with a temporal discounting task.

We acknowledge the important contribution of this work. However, two crucial empirical issues remain. First, although this study examined all three components, it did not report the extent to which the three components independently predict outcome variables of interest. This is important because each component may show different patterns of unique relationships with outcomes depending on how much overlap it shares with the other components. This is a crucial issue as researchers interested in future self-identification may wish to understand the unique contribution of each component for different outcomes (e.g., for the design of experimental manipulation/intervention). Second, the convergent, divergent, and independent predictive validity of the frequently used “global and decontextualized” measures of the future self remain unknown. In Sokol and Serper’s [109] assessment of the three components, participants considered the domains of their future self chosen by the researchers (e.g., values, personality, beliefs, appearance, and family relations). However, much of the literature on this topic assessed the components with global, decontextualized measures, emphasizing participants’ freedom to consider the aspects of the future self that are most important to them. Accordingly, we consider these issues in the present research.

Overview of the present research

Our review of the literature supports the idea that there are three major components of future self-identification that have downstream consequences. In the present research, we will address three major research goals: (1) advance the conceptualization of the relationship between the current and future self, (2) test a measurement model of the factor structure of future self-identification by extending and developing items tapping into the three components of future self-identification, and (3) examine whether and how the components of future self-identification uniquely relate to intrapsychic and achievement outcomes. To study all three components in one study, we aim to minimize methodological differences in the measures by framing all items in terms of the decontextualized, global future self. Our literature review shows that few studies on this topic have examined the relationships between multiple components of future self-identification and outcomes related to intrapsychic adjustment and objective markers of achievement. Therefore, the present research focuses on the following four outcome domains to help clarify the unique relationships with the three components of future self-identification: (1) psychological well-being (self-esteem and state hope), (2) imagination of the future (visual imagery of future events and perceived temporal distance), (3) self-control, and (4) academic performance.

We organize the report below into four parts. It was first necessary to extend and develop items that tap into all three components of future self-identification. In Part I, we examine whether the future self-identification items have a triadic factor structure. In Part II, we further test the reliability of the future self-identification factors by assessing test-retest reliability and the replicability of the factor structure in a new sample. In Parts III and IV, we then ascertain the unique relationships between the three components of future self-identification and our outcomes of interest. Because prior research often only included measurement of one component of future self-identification, Parts III and IV will help clarify the independent relationships between the three components and important intrapsychic and achievement outcomes.

A crucial aspect of this research is the life stage of our participants. We chose participants going through a major age-normative transition—starting college—because such transitions are junctures in people’s lives. Although older students (e.g., third-year college students) may have more information in predicting their senior year or beyond, we expect that differences in how entering college students conceive of their future may shape important outcomes such as academic performance. As many students drop out of college after the first year [110], research targeting entering freshmen such as the present one may give insight into early interventions to increase retention. Finally, our future self-identification scale used a future self in 10 years as a reference point. Five or six years after graduation (approximately 10 years from starting college) is when many critical life decisions are made (e.g., career choice, marriage, and family decisions). Degrees of identification with a future self at this time frame may be impactful on academic outcomes during college.

Part I: Factor structure of future self-identification items

In Part I, we used confirmatory factor analysis to test whether the extended and newly developed items of the three components of future self-identification represent discriminant factors. Specifically, we tested whether the best fit to the future self-identification items was provided by a three-factor model. As described above, relatedness, vividness, and positivity were expected to be separate components of perceived psychological identification with the future self. Given that each of the three components focuses on different aspects of the future self (e.g., degree of imagination, valence), we predicted that the three components would be discriminant (but positively correlated) factors within an overall model.

Method

Analysis.

We used confirmatory factor analysis to examine the factor structure of the future self- identification items. Five models were constructed to test competing conceptualizations of the future self-identification construct. In the one-factor model, all six future self-identification items loaded onto a single latent variable. A three-factor model was fit where the two relatedness items were indicators of the first factor, the two vividness items were indicators of the second factor, and the two positivity items were indicators of the third factor (see Fig 1 for a visual representation of the three-factor model). For completeness, we also tested all of the possible two-factor models: (a) relatedness items were loaded on one factor and positivity/vividness items were loaded on a second factor, (b) vividness items were loaded on one factor and relatedness/positivity items were loaded on a second factor, and (c) positivity items were loaded on one factor and relatedness/vividness items were loaded on a second factor.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The three-factor model of future self-identification.

See the Method section in Part I for a description of the future self-identification items. The Method section in Part IV includes description of the revised future self-identification items.

https://doi.org/10.1371/journal.pone.0242504.g001

Across all five models, we implemented the following specifications: (a) factor means and variances were set to zero and one, respectively; (b) the loadings for the unique factors were set to one; (c) the means of the unique factors were set to zero; (d) the factor loadings, the intercepts of the indicators, and the variances of the unique factors were freely estimated. For the multi-factor models (i.e., the two-factor and three-factor models), covariances between the latent variables were also freely estimated.

Results

Table 1 shows the descriptive statistics for the six future self-identification items. The fit statistics of the five confirmatory factor analyses are presented in Table 2. Fit statistics included the chi-square test, the comparative fit index (CFI), the root mean square error of approximation (RMSEA) with 90% confidence intervals, and the standardized root mean square residual (SRMR). Due to the chi-square test being influenced by sample size, the CFI, RMSEA, and SRMR were also used as the indices to examine model fit in the study. Per general conventions, CFI values above .90, RMSEA below .08, and SRMR below .08 were used as indications of reasonable model fit [111, 112].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics of the future self-identification items.

https://doi.org/10.1371/journal.pone.0242504.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Fit statistics of the five confirmatory factor analyses in Part I.

https://doi.org/10.1371/journal.pone.0242504.t002

The one-factor model and the three two-factor models did not reach acceptable levels of fit on any of the indices (e.g., RMSEAs > .13), suggesting these models did not adequately account for the covariance structure of the data. The fit of the three-factor model was good and reached acceptable levels on all fit indices. One common method to assess whether one model provides a better fit to the data than another model is if the change in CFI (ΔCFI) is more than .01 [113]. Using this criterion, the three-factor model provided a better fit to the data than the one-factor model and the three two-factor models.

Table 3 contains the standardized factor loadings. In the three-factor model, all items loaded significantly (ps < .001) on their designated factors. The correlations among the three factors were all positive and significant: relatedness and vividness (r = .38, p < .001), relatedness and positivity (r = .38, p < .001), vividness and positivity (r = .61, p < .001). These results support a three-factor structure of the future self-identification items.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Standardized factor loadings of the three-factor model of future self-identification.

https://doi.org/10.1371/journal.pone.0242504.t003

Method

Results

Test-retest reliability of the three-factor model.

We first report the test-retest reliability of the three-component model of future self-identification. The test-retest reliability analyses reported below utilized data from two time periods separated by approximately five weeks (Time 1 and Time 2). The three-factor model from Part I (see Fig 1) was fit to the Time 1 and Time 2 data simultaneously. This produced six latent factors (the relatedness, vividness, and positivity factors at both Time 1 and Time 2). We implemented the same specifications as was done in Part I, with the following additional specifications: (a) all six latent factors were allowed to covary; (b) unique factors were allowed to covary across time periods (e.g., the unique factor for the relatedness item at Time 1 with the unique factor for the relatedness item at Time 2).

Factorial invariance testing. Prior to measuring the test-retest reliabilities of the future self-identification factors, we first established factorial invariance across the two time periods. This ensured that the latent variables (e.g., the relatedness, vividness, and positivity factors) measured the same constructs at the two time periods. To test for factorial invariance, we followed the steps outlined by Widaman, Ferrer, and Conger [115]. Specifically, we ran three models with increasingly strict constraints placed on the estimated parameters: (1) configural invariance, where the same factor structure is fit at the two time periods but all free parameters are estimated separately across time; (2) weak factorial invariance, where factor loadings are constrained to be equal across time; and (3) strong factorial invariance, where factor loadings and item intercepts are constrained to be equal across time.

If adding constraints to the model leads to a significant change in model fit, factorial invariance would not be achieved. The CFI model fits were good for all three models: the configural invariant model (.991), the weak factorial invariant model (.989), and the strong factorial invariant model (.977). Moreover, ΔCFI between the configural and weak factorial invariant models (.002) and between the weak and strong factorial invariant models (.012) did not or slightly exceeded .01. As a result, we concluded that strong factorial invariance was established for the three-factor model across the two time periods.

Test-retest results. The overall fit of the strong factorial invariant model reached acceptable levels on the fit indices, χ²(45) = 98.74, p < .001; CFI = .977; RMSEA = .057 [.042, .072]; SRMR = .038. The test-retest reliabilities of the three factors were as follows (see Table 4): relatedness (r = .65, p < .001), vividness (r = .80, p < .001), and positivity (r = .85, p < .001). These significant correlations demonstrate the stability of individual differences in the future self-identification components over the five-week period between Time 1 and Time 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Correlations between the three factors of future self-identification at time 1 and time 2 in Part II.

https://doi.org/10.1371/journal.pone.0242504.t004

Psychological well-being

The relatedness component is positively related to satisfaction with life [116]. Perceiving greater relatedness between current and future selves appears to confer benefits to present psychological and subjective well-being. Additionally, relatedness of the future self is positively related to positive affect and negatively related to negative affect [7]. Relative to individuals who perceive less relatedness between selves over time, those who perceive greater relatedness appear to spend more time in positive affective states (e.g., happy, joyful, and fun times) and less time in negative states (e.g., depressed, unhappy, and frustrated).

By measuring all three components of future self-identification, we were able to test their unique relationships with psychological well-being adjusting for the other components. In Part III, self-esteem was chosen as the psychological well-being variable of interest. Due to the conceptual similarity between the future self-positivity items and self-esteem, we expected that the positivity component would be the strongest predictor of self-esteem. Because the positivity component relates specifically to attitudes toward one’s future self, we modified the Rosenberg Self-Esteem Scale [117] with individuals’ future selves as the reference. This allowed us to assess the unique effects of the three future self-identification components on both current and future self-esteem. We predicted a stronger relationship between the positivity component and future self-esteem as compared to current self-esteem. These predictions highlight the idea that individuals can have different evaluations of self-worth and personal value (i.e., self-esteem) for selves at different time periods (e.g., past, current, and future selves). Moreover, it remained an empirical question if relatedness and vividness demonstrate independent relationships with current and future self-esteem after adjusting for positivity.

Visual imagery of future events

Self-reported ability to create vivid mental images is positively associated with future-oriented behavior, such as reduced levels of procrastination [7]. Individuals may also feel more psychologically connected to a future self if they can more easily visualize details associated with events surrounding that future self. In one study [8], college students at the beginning of an academic semester were measured on how vividly they could imagine aspects of their future self at the end of the semester. Ratings of vivid imagery were positively related to perceived relatedness between the current self and the future self at the end of the academic semester.

To measure visual imagery of future events in the current study, we chose events surrounding two future time periods: (1) vividness of college graduation, and (2) vividness of a typical work week five years after college. These future time periods were chosen because events surrounding college graduation and five years post-graduation are relevant for the long-term goals students possess throughout college. Because visual imagery of future events relies upon the extent to which one holds a vivid conception of the future, we predicted that the vividness component of future self-identification would have the strongest relationship with the visual imagery outcomes. As the current study included all three future self-identification components within a single model, we were able to test if relatedness and positivity demonstrate unique relationships with visual imagery of future events after adjusting for future self-vividness.

Self-control

Resisting the temptation of immediate gratification in order to maximize long-term rewards typically requires a degree of self-control [118–122]. High future self-identification may promote self-control. If people feel strongly identified with their future selves, they may be better able to eschew behaviors that lead to short-term gratification but negatively impact long-term well-being. In past research, the relatedness component of future self-identification correlated positively with a measure of self-control [5, 29]. Individuals high in future self-identification may be better able to shift their temporal focus to the future instead of overly weighing present priorities. In fact, greater perceived relatedness to the future self is positively related to consideration of future consequences and negatively related to consideration of immediate consequences [5].

Prior research only studied the relationship between the relatedness component of future self-identification and self-control. Consequently, it remains unclear which component(s) of future self-identification predict self-control. By including all three components within a single study, the present investigation helps clarify the nature of the relationship between future self-identification and self-control.

Academic performance

Lastly, we examined the relationship between future self-identification and academic performance. Academic achievement often requires the eschewing of immediate gratification (e.g., studying instead of going out with friends) in the service of long-term benefits (e.g., higher grade point average). Similar to the relationship between future self-identification and self-control described above, future self-identification may promote academic achievement by regulating behaviors to be more aligned with long-term goals. Evidence of this includes the negative relationship between perceived relatedness with future selves and academic procrastination [7]. Individuals with a greater perceived overlap between current and future selves appear to be less likely to delay completion of academic tasks.

As it relates to academic performance, a significant relationship was observed between the relatedness component of future self-identification and college grade point average [5]. Individuals with greater perceived relatedness between current and future selves at the beginning of an academic semester have higher grade point averages at the end of the semester. However, it remains an open question whether vividness and positivity would also be significant predictors of subsequent academic performance. The inclusion of all three components of future self-identification within a single study allows us to examine this research question. Moreover, we were also able to test the indirect effects of the three components on academic performance through self-control. This indirect effect was observed for relatedness by Adelman et al. [5] but further testing whether indirect effects are observed for the other two components may help clarify the possible mechanisms that underlie the relationship between future self-identification and academic outcomes.

Method

Results

Given the differences in scale formats for the future self-identification items, we first standardized the items and averaged the two items separately for each of the three components. We then used participants’ scores on these three components to predict the outcome variables of interest. We report both the zero-order correlations and multiple regression coefficients. The latter analyses were to test the unique relationships between the three future self-identification components and the dependent variables. Finally, a path analytic model is included at the end of the Results section to test the indirect effects of the future self-identification components on academic performance through self-control. For all regression models, multicollinearity was not an issue with all variance inflation factors being less than two.

Self-esteem.

We examined the zero-order correlations of the three future self-identification components with the dependent variables (Tables 5–8). For the self-esteem variables, the correlations were positive and significant for all three components (ranging from .25 to .64), with the strongest correlations being observed for positivity. For both current and future self-esteem, the correlation with positivity was significantly higher than with relatedness and vividness (zs > 3.81, ps < .001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Multiple regression analyses of future self-identification components predicting psychological well-being.

https://doi.org/10.1371/journal.pone.0242504.t005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Multiple regression analyses of future self-identification components predicting visual imagery of future events and perceived temporal distance.

https://doi.org/10.1371/journal.pone.0242504.t006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Multiple regression analysis of future self-identification components predicting self-control.

https://doi.org/10.1371/journal.pone.0242504.t007

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Multiple regression analysis of future self-identification components predicting Grade Point Average (GPA) in Part III and GPA expectations in Part IV.

https://doi.org/10.1371/journal.pone.0242504.t008

We next carried out multiple regression analyses with the three future self-identification components predicting current and future self-esteem separately. As predicted, the future self-identification components accounted for more variance in future self-esteem, F(3, 759) = 177.52, p < .001, R² = .41, than current self-esteem, F(3, 752) = 71.73, p < .001, R² = .23. Specifically, positivity was the strongest independent predictor of self-esteem, and its relationship was stronger with future self-esteem (β = .60) than with current self-esteem (β = .36). This pattern of results supports the concurrent relationship between future self-identification and self-esteem, with a stronger relationship between the positivity component and future self-esteem. Relatedness and vividness did remain significant predictors of current self-esteem after adjusting for the effect of positivity. However, these two predictors failed to reach significance for future self-esteem (see Table 5). Taken together, these results suggest that the three components of future self-identification can independently predict current levels of self-esteem, but that future self-esteem is more exclusively related to the positivity component.

One possible explanation for the above pattern of results is that the future self-positivity items simply measure a global self-evaluative construct. If this is the case, the strong relationship between the positivity component and self-esteem could just be due to a general tendency towards positivity. To test this possibility, a second model was performed with current self-esteem added as a predictor of future self-esteem (Table 5). The overall regression model was significant, F(4, 751) = 300.38, p < .001, R² = .62. Importantly, positivity remained a significant unique predictor of future self-esteem (β = .40) adjusting for current self-esteem. These results suggest that the positivity component does not simply reflect a general positive self-disposition, but is strongly tied to future self-attitudes.

Path analytic results.

To examine whether future self-identification relates to GPA indirectly through self-control, a path model was tested. The three future self-identification components were exogenous variables in the model; self-control and GPA were endogenous variables. Direct paths from the future self-identification components to GPA were included in the model, as well as indirect paths through self-control (see Fig 2 for a visualization of the path model). In this and all following path models, 95% bias-corrected confidence intervals stemmed from 5,000 bootstrapped samples [128]. The significance of the indirect effects was determined by ascertaining whether the 95% confidence intervals included zero or not.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Path analytic models in Parts III and IV.

Self-control, hope, or perceived temporal distance acted as intermediary variables (psychological resources) between the three future self-identification components and academic outcomes (cumulative GPA or semester GPA expectations). Covariances between the future self-identification components were included in all models.

https://doi.org/10.1371/journal.pone.0242504.g002

The direct and indirect effects of the path model are included in Table 9. Both relatedness and self-control had significant direct effects on GPA. Nevertheless, all three future self-identification components had significant indirect effects on GPA through self-control. That is, for all three indirect effects, the 95% bias-corrected confidence intervals did not include zero (see Table 9). These results suggest that all three future self-identification components have indirect relationships with college GPA, with self-control identified as a possible intermediary variable.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. Path analytic results relating the future self-identification components to academic outcomes (cumulative GPA in Part III and semester GPA expectations in Part IV) through intermediary psychological resources (self-control, hope, or perceived temporal distance).

https://doi.org/10.1371/journal.pone.0242504.t009

Method

Results

The results section for Part IV is organized as follows. We first report the confirmatory factor analysis of the modified future self-identification items. We then report the unique relationships of the three future self-identification components and the variance they explain together in each domain of interest: psychological well-being (self-esteem, hope), imagination of the future (visual imagery of future events, perceived temporal distance), self-control, and semester GPA expectations. We conclude the Results section with path analytic models that probe the indirect effects of the future self-identification components on semester GPA expectations through the variables of self-control, hope, and perceived temporal distance.

Summary of findings

The present findings show that the three components of future self-identification—relatedness, vividness, and positivity toward one’s future self—represent correlated but independent factors. Most prior measures of future self-identification focused on only one or two of these components [but see 21, 92]. As a result, it remained unclear which aspects of future self-identification were associated with psychological and behavioral outcomes. The results of the current research suggest that the answer to this question is domain dependent, which demonstrates the utility of measuring all three components in future studies.

For self-control, all three components of future self-identification had significant relationships, with the strength of the correlation and regression coefficients being comparable among the three components (though the unique effect of positivity was not significant in Part IV). These results suggest an additive model of the effects of the components of future self-identification for the domain of self-control, in which higher levels of each future self-identification component are uniquely associated with increased self-control. As the path analytic models further demonstrated, the relationship between the future self-identification components and self-control has particular relevance for the academic domain, as significant indirect effects were observed between the components and grade expectations and performance.

In other domains, one of the three components had a noticeably stronger relationship than the other two. These distinctive relationships demonstrate the external convergent validity of the components and the utility of adopting a three-component model of future self-identification. For self-esteem, the positivity component had the strongest relationships compared to relatedness and vividness, with a particularly strong relationship observed between positivity and self-esteem towards the future self. These results support and extend prior research that demonstrated an association between future self-identification and psychological and subjective well-being [7, 116]. Yet, by not taking the positivity component into account, prior research may have overestimated the degree of relationship between future self-relatedness and well-being. Finally, the positivity component does not simply measure a global positive self-evaluation. In models predicting future self-esteem and state hope, the positivity component remained a significant unique predictor even after adjusting for current self-esteem, suggesting that the positivity component taps into the valence attached to future selves.

As expected, visual imagery of future events was most strongly related to the vividness component, though positivity noticeably related to vivid imagery of certain future events as well (e.g., college graduation). Individuals with higher positivity towards future selves may be more likely to spend time imagining the vivid details surrounding happy and celebratory events in the future. The extent that positivity towards future selves relates to vivid imagery of future events that span a range of perceived affective states remains an open question. In Part IV, we included an additional imagination of the future variable, which measured perceived temporal distance of college graduation. For this outcome, both vividness and relatedness negatively predicted perceived temporal distance. Holding a vivid conception of a future self and perceiving relatedness between current and future selves are associated with perceiving college graduation as closer in time. Perceiving future events as being closer in time might be particularly relevant for academic outcomes, due to subjective time perception relating to intertemporal preferences [138]. Individuals who perceive time durations to be subjectively longer tend to discount delayed rewards at steeper rates [139]. The current results suggest that perceiving graduation as closer in time might confer academic benefits, as the vividness component was shown to relate to semester GPA expectations through perceived temporal distance.

Finally, academic outcomes (cumulative GPA and semester GPA expectations) were only directly related to the relatedness component. This latter finding extends previous research by showing that relatedness accounts for variance in academic performance net of the vividness and positivity components. Vividness and positivity did demonstrate indirect effects on academic outcomes, however. These results suggest that all three future self-identification components may be relevant to academic performance, but vividness and positivity possibly exert their influences through other psychological variables. The present study identified three potential intermediary psychological variables (self-control, hope, and perceived temporal distance). But these preliminary results warrant replicability, as well as identification of other intervening variables between future self-identification and academic performance.

With self-control having a significant relationship with academic outcomes, it may be surprising that the vividness and positivity components had significant relationships with self-control but not significant direct relationships with GPA. While all three components tap into individuals’ perceptions of their future self, the relatedness component is the only one explicitly assessing the link between current and future selves; the vividness and positivity items only measure individuals’ perceptions of their future self with no mention of the current self. Being a student is likely to be a salient identity of emerging adults who are experiencing their first year at college. Consequently, academic performance may therefore be more directly tied to the relatedness component because this component taps into their current self-identity as a student.

According to the typology of mediation effects proposed by Zhao, Lynch, and Chen [141], the significant indirect effects of the vividness and positivity components on academic performance would be termed indirect-only mediation effects. These effects are still meaningful in the absence of significant direct or total effects. These findings suggest that merely having strong future self-identification may not always lead to positive outcomes such as better academic performance. Some individuals with a vivid and positive future perception may direct their effort to goal-directed behaviors (i.e., self-control) and in turn perform better in school (i.e., higher GPA). However, others with a vivid and positive future perception may daydream about a positive future without taking action or exercising self-control and do not do well. These opposing indirect effects could then cancel each other out with regards to the total effects of vividness and positivity on academic performance. These findings highlight the importance for future research to address the conditions under which future self-identification enhances performance.

Future longitudinal research will be necessary to extend these indirect effect analyses to more conclusively tease apart the temporal order and causal nature of the relationships. The long-term associations between future self-identification and later behavioral/psychological outcomes is a relatively unexplored topic. In a recent article, Reiff, Hershfield, and Quoidbach [142] found that the relatedness component was positively correlated with a measure of well-being (life satisfaction) ten years later. The current results demonstrated that the future self-identification components had different unique relationships with psychological well-being (e.g., self-esteem, hope). As a result, it is important for future research to test longitudinal relationships between all three future self-components and well-being, as well as investigate the mechanisms that underpin these longitudinal associations.

Implications, limitations, and future directions

To test the conceptualization of future self-identification as a triadic construct, our first aim was to develop a more comprehensive measure of future self-identification. We built upon and extended current measures of the construct by including items designed to tap into all three components. Through confirmatory factor analysis, we demonstrated that the three components could be measured as discriminant factors within an overall model. We further reported on the test-retest reliability over a period of five weeks and replicated the model in two additional samples. These results helped establish the reliability and internal validity of the future self-identification measure and its utility for researchers seeking to assess future self-identification in a more nuanced way.

A better understanding of the multi-dimensionality of future self-identification is relevant to intervention research. Recent interventions were designed to increase levels of future self-identification [8, 80, 90]. The goal of these interventions was to foster long-term beneficial behavior such as monetary savings [90] and delinquency [91, 92] by increasing feelings of connection between present and future selves. The current research suggests that the component most relevant to an intervention varies with the type of outcome of interest. Consequently, interventions should focus on the component(s) of future self-identification that have the strongest relationship in the outcome being targeted. For example, if the positivity component has the strongest relationship with a target outcome (e.g., psychological well-being), interventions should focus on positivity rather than relatedness or vividness. Ensuring that interventions are tailored to the most relevant component(s) will be necessary during the planning and design process of subsequent future self-identification interventions. It also might be the case that other components of future self-identification will be identified. Though the current review and analyses supported three components of future self-identification [4], future research will need to explore the possibility of other components.

Another direction for future research is to determine whether the components of future self-identification exert interaction effects on outcomes. We did not have any a priori hypotheses about interaction effects, but we believe extending our findings along these lines is an important next step. For instance, certain outcomes may require a high degree of future self-identification on multiple components. For example, the benefit conferred by a vivid future self on attainment of goals is likely to vary with the valence of the future self. Whereas a vividly imagined positive future may elicit behaviors instrumental to goal attainment, a vividly imagined negative future may evoke self-defeating behaviors due to negative psychological states (e.g., dread) that undermine goal attainment.

Some limitations of the current study should be noted. First, each of the three future self-identification factors was measured with only two indicators. There are statistical benefits to including more than two items per factor when carrying out a factor analysis, such as enhancing reliability [143]. For this reason, we carried out multiple methods of measuring reliability and validity (both internal and concurrent) of the factor model, such as testing the factor structure over multiple time points and across different samples of participants. The findings from this investigation demonstrate that the three components of future self-identification can be measured with a brief scale shown to be both reliable and valid. Nevertheless, developing future self-identification subscales with more items per factor is an important avenue for future research. Having brief and longer measures of each factor enables researchers to choose between the statistical benefits of assessing each factor with more items and the need to measure variables parsimoniously when respondent burden is an issue (e.g., multi-disciplinary longitudinal studies). Moreover, to further validate the future self-identification components, future research should include measures that utilize a variety of method formats. The current future self-identification and psychological resource measures relied exclusively on self-report, so our understanding of the consequences of future self-identification will be aided by a wider diversity of measurement methodologies.

A second limitation is that the current study focused solely on undergraduates in their first year of college. The choice of this sample was pertinent because adolescents entering college are involved in a major life transition. During this stage, feelings toward one’s future self may be particularly important in fostering long-term beneficial behaviors (e.g., academic persistence). Yet, the extent that the three-factor model of future self-identification observed in the current samples generalizes to other age or cultural groups is unknown. Even within the college student population, entering freshmen may not have as much information about their future self with a college degree as compared to more advanced students. It will be fruitful for future longitudinal research to test whether there are changes in the factor structure of future self-identification during college, and the psychological and academic consequences of these changes. The concurrent and predictive validity of the three components might also differ based on age group, the likelihood of life transitions in the near future, and cultural context. For instance, future self-identification may be especially relevant when individuals go through role transitions. The current study focused on students starting college, but other relevant important life transitions could include starting a family, getting divorced, receiving a diagnosis of a chronic disease, re-careering, and retirement. Future research that targets these different populations will help elucidate the critical role that future self-identification plays across different life stages and transitions.

It will be important for future research to investigate whether the effects of future self-identification are moderated by factors such as demographic characteristics. One such example is whether the relationship between future self-identification and academic performance differs across racial/ethnic groups, due to the academic achievement gaps often reported in the literature [125–127]. Cultural backgrounds and experiences may impact the way students conceptualize their future selves, as well as how they perform academically. For instance, college generation status (first-generation vs. continuing-generation students) was shown to moderate the relationship between temporal psychological factors and psychological resources [5]. Demographic groups that differ in college generation status, as well as on other cultural and historical factors, may identify with their post-college future self differently. Therefore, the pathways by which future self-identification impacts academic behavior and performance may vary across demographic groups. We believe this is a promising avenue for future investigation.

Consistent with previous research, we found that the components of future self-identification were related to several psychological resources (e.g., self-control, self-esteem) which influence a wide range of consequential outcomes [124]. Nevertheless, a third limitation of the present study is that we assessed only one behavioral outcome, with future self-identification accounting for a small proportion of variance in cumulative GPA. Prior research has demonstrated an effect of future self-identification on a variety of behavioral outcomes, including savings behavior [21], exercise rates [64], and delinquency [92]. Thus, future research is warranted that investigates the unique relationship between the three future self-identification components and behavioral outcomes such as work productivity and adherence to medical regimens.