When it doesn’t get better: Leaning into minoritized identity to brace for decline

Introduction

The joy that some Americans felt in response to Barack Obama’s election as president in 2008 set off a phase of cautious optimism about the state of society. This optimism was marked by assertions that remaining politicized identity challenges were temporary – a claim made most concisely in the title of Dan Savage’s book with his husband Terry Miller, It Gets Better (2012) – and that the path of human history had shifted away from collective bloodshed and towards reason, empathy, and more enlightened forms of human engagement, as argued in Steven Pinker’s The Better Angels of Our Nature (2011). In optimistic times, it can be easy to imagine human history as ultimately “bending toward justice,” in the words of preacher Theodore Parker (1853), words later echoed and made famous by U.S. civil rights leader Martin Luther King, Jr.

But the idea that any individual will be guaranteed to observe overall positive social change in their lifetime is an ahistorical illusion: upward trajectories are not inevitable. The many examples of democratic decline currently occurring worldwide – the topic of this special issue – remind us of this possibility. The consequences of democratic decline, and the political chaos that typically ensues, often target minoritized groups most acutely (see, e.g., Glick, 2010). Here we focus on perceptions of societal decline among one historically minoritized group: sexual minorities. This group (including lesbian, gay, bisexual, transgender, and queer individuals; sometimes abbreviated LGBTQ+) has been very successful in effecting social change in Western liberal democracies, moving from criminalisation to public celebration within a relatively short period of time. These successes, however, seem to be giving way to a more recent period characterised by pushback (e.g., Ayoub & Stoeckl, 2024). Concretely reflecting this new situation, in June 2023 the US-based Human Rights Campaign (HRC, 2023) declared a “state of emergency” following the introduction of anti-LGBTQ+ bills in numerous state legislatures.

As minority group members contemplate the implications of democratic decline and the increasingly hostile societal contexts these create, some perspectives might expect them to “lean out” – that is, to psychologically disengage (e.g., Major & Schmader, 1998), to conceal identity (Goffman, 1963; Yoshino, 2007), or to migrate from prejudicial contexts in search of more supportive environments (e.g., Esposito & Calanchini, 2022). In contrast to these (understandable) responses, we consider when members of the minority might “lean into” collective identity, displaying identity vigilance in difficult times. In using the term “identity vigilance,” we allude to the concept of consolidating awareness around a potentially threatened identity (see the use of the term ‘vigilance’ in Steele et al., 2002) and mobilizing psychological resources for its protection. To investigate this hypothesized phenomenon, we analyzed data collected from multiple waves of LGBTQ+ participants that we gathered over an eight-year period marked by substantial global social change – including discourse about democratic decline in the United States and globally. These data provide a window through which to observe psychological responses to an unfolding situation in light of minority identity status.

Reactions to Social Decline

Within social psychology, social transformations are often understood as the product of collective action, which in turn is understood as being grounded in a perception that change is possible. From the perspective of social identity theory (SIT; Tajfel & Turner, 1979), membership in a socially subordinated or devalued group imposes negative identity on the individual. Exactly how individuals negotiate this situation to restore a positive self-concept is theorised to depend on appraisals of the social structure and the opportunities for change this implies. SIT specifies that when social structures are perceived to be fixed (i.e., stable and legitimate) but group boundaries permeable, individualistic identity management strategies are likely to prevail – for example, through attempts to distance the self from group membership and affiliate with more socially valued groups (i.e., ‘individual mobility’ in SIT terms). If fixed social structures instead involve group boundaries that are perceived to be impermeable, negative identity can only be managed through attempts to reimagine the value of the group despite its societal devaluation (i.e., ‘social creativity’). However, when social structures instead signal the possibility of change (e.g., when hierarchies are perceived to be unstable and/ or illegitimate), members of devalued groups are likely to become motivated to engage collectively in actions that seek to challenge and change the status quo (i.e., ‘social competition’). According to the theory, instability in the social structure opens up the capacity to imagine ‘cognitive alternatives’ to current social arrangements, supplying members of devalued groups with a vision of what to work toward. Empirical work confirms these theoretical predictions: perceptions of likely and desirable social changes have been found to crystallise identification within socially disadvantaged groups (e.g., Bettencourt et al., 2001; Ellemers, 1993; Mummendey et al., 1999), especially among those previously uncommitted (Doosje et al., 2002); increased identification, in turn, prompts support for group-based collective action to ensure that anticipated changes are realised (Kelly, 1993; van Zomeren et al., 2008; Wright et al., 2022).

Although supported by a large body of research – for a summary, see Reicher and Haslam (2012) – much of this evidence is based on cross-sectional surveys or short-term experiments that induce momentary shifts in perceptions of the (in)security of structural relations between groups. The unfolding dynamics of social perceptions, identification and group-based responses are rarely studied over longer time periods (see Bleh et al., 2025, for a similar argument), perhaps not only because the arc of history is long but also because of the logistical challenges inherent in such initiatives (see Johal et al., 2023). One recent study, however, did capture this unfolding dynamic in the context of environmental activist groups (Bleh et al., 2025). Here, improving perceptions of cognitive alternatives to the status quo (i.e., perceptions that desired social changes were a plausible possibility; see also Wright et al., 2020) were related to across-time increases in group identification and to strengthened feelings of collective efficacy, and through these to increased support for actions intended to reinforce commitment to group-based action. This suggests an upward spiral from growing optimism about possible futures, through identification, to further action in the service of social change.

Yet other work within the same theoretical tradition suggests an alternative role for identification in response to social perceptions. The rejection-identification model (see Branscombe et al., 1999) argues that collective identification is a psychological resource that can be mobilized to protect members of disadvantaged groups from the negative effects of their devalued status. Drawing on the logic of SIT, this protective role of minority identification is argued to emerge when group devaluation is construed as a pervasive rather than isolated experience (e.g., through repeated discrimination). Pervasive devaluation indicates stability in the social system – stability that spurs minority group members to locate their sense of self and belonging away from the dominant group. Within this model, achieving a sense of belonging within the minority community offsets the otherwise negative effects of pervasive devaluation on individual well-being.

There are many differences between the two accounts outlined above and these need not be seen as oppositional. Indeed, it is perfectly plausible that different pathways to identification operate in parallel. In the context of unstable social systems, in which desired social changes are perceived as possible, identification within devalued groups anticipates improvements in group status (e.g., Doosje et al., 2002) and becomes the psychological mechanism for achieving these improvements collectively. In the context of pervasive discrimination, in which social structures seem less open to change, identification reflects the salience of group boundaries conveyed through differential treatment and invokes solidarity with others who are similarly disadvantaged.

And yet, there is also some conceptual fuzziness between accounts. For example, the rejection-identification model explicitly draws on SIT’s concept of stability to reason why pervasive discrimination would sharpen identity within minority groups, despite other SIT research arguing and showing that instability, and the capacity for instability to open up desirable cognitive alternatives, catalyses identification and collective action toward social change. Empirically, some studies find that activating perceptions of social change reduces perceptions of discrimination, resulting in lower levels of identification among minoritized groups (e.g., Schmitt et al., 2009; Spoor & Schmitt, 2011), but other studies find that activating the possibility of social change enhances minority identification and well-being (e.g., Zhang et al., 2013). Further complicating this picture, higher levels of identification have also been argued to feed into societal perceptions, not just reflect them – for example, by sensitizing minorities to possible discrimination (e.g., Branscombe et al., 1999), contributing to scepticism about the rate and extent of social change (cf. Eibach & Ehrlinger, 2006). Across various literatures linked to SIT, core constructs of identification and perceptions of (actual and possible) social change align with each other in multiple ways and thus deserve further attention.

The Current Research

Despite strong theory in this domain, the existence of complex and nuanced relationships among constructs leaves open questions about whether – and how – identification operates in response to societal arrangements that are promising versus worrying from the perspective of devalued minority groups. It is unclear whether stability or instability in a currently devaluing social system should be most likely to activate minority identity. In the more specific context of theorising and research about instability, it is further unclear whether minority identity activation is limited to the typically studied versions of this that support optimism and hope rather than instability in the form of erosion or decline. These questions remain open partly because the trajectory of social change is often assumed to be slow, but ultimately positive (i.e., things are always getting better); yet the scope of assessment within psychological research is typically (very) short term. Where research has considered trajectories of social change, these have been more rapid, often signalled by sudden disruptions to the social order triggered by natural disasters, organizational change, or dramatic political events (e.g., see de la Sablonnière & Tougas, 2008; de la Sablonnière et al., 2009; Smith, Livingstone, & Thomas, 2019) or within the context of relatively short-term movement-building activities (e.g., a climate camp: Bleh et al., 2025). To our knowledge, there are very few projects tracking minority responses to developing societal trajectories over a longer period of time.

To address this issue, we present data that document evolving perceptions of the possibilities for progressive social change, and the correlates of this, among (largely US-based) LGBTQ+ individuals over a wider time frame spanning eight years. Our successive waves of sampling were originally conceived as repeated replications of an experiment testing the effects of progressive social change on the meaning and value of minority identity (see Salvatore & Morton, 2017). As elaborated in our pre-registered experimental predictions, this interest was grounded in assumptions that social change was becoming so positive that it might be undermining the ongoing value and relevance of the group – raising theoretically interesting questions about how different individuals within the group position their identity in relation to this “new reality.” As we describe below, this starting assumption was violated by the actual trajectory of social change we observed over iterations of our experiment. This guided our decision to engage with the across-time changes instead of the originally focal experimental effects. This paper reports those across-time changes, the context of which we briefly outline below.

Prior to the period of sampling, the direction of social change reflected a clear ascendancy of liberal values (e.g., under the Obama administration) and policies and practices perceived to be supportive of sexual/gender minorities. At this juncture, social conditions were perceived as being so good that some within the LGBTQ+ community were asking whether this identity should continue to be relevant for self-definition (see, e.g., Savin-Williams, 2009). Even during the first Trump administration, one commentator – perhaps an outlier in the community – declared that “the struggle for gay rights is over” (Kirchick, 2019). But others at that time (e.g., Finn, 2019) saw the trajectory of social change as shifting in a markedly different direction. The newly unfolding social reality – especially in the US where most of our research participants were located – was instead characterised by increasingly prominent administrative opposition to previously gained social and legal protections for sexual/gender minorities (especially trans individuals). Capturing that shift in both rhetoric and policy, the first three samples in our sequence cover the beginning, middle, and end of the first Trump presidency (i.e., 2017, 2019 & 2020, respectively). Our fourth sample was collected at the end of 2021, a year after the office of president had returned to a liberal candidate (Biden); but by the fifth and final sample, in early 2025, the presidency had returned to Trump (see Footnote 1). As noted above, the situation for sexual/gender minorities toward the end of this sampling period was sufficiently concerning that the US-based Human Rights Campaign declared a “state of emergency” (HRC, 2023).

The shifting direction of social change provided a unique opportunity for insight into what happens within minority groups when things don’t just get better. These data also allowed us to explore competing theoretical ideas about whether optimistic or pessimistic perceptions of the possibilities for progressive social change are more activating of minority identity. Work within SIT generally highlights the value of positive perceptions of social change in catalysing identity by, for example, positioning improving cognitive alternatives to the status quo as a mechanism for collective mobilization toward a desired future. Instead, drawing on the notion of collective identity as a coping resource, articulated by the rejection-identification model, we tested the idea that minority individuals might ‘brace’ themselves collectively in the face of social decline and draw strength from shared identity exactly when things are going from bad to worse, rather than when they are getting better. This bracing response implies an oppositional relationship between unfolding pessimism over the possibilities of progressive social change on the one hand, and re-investment in shared identity on the other. It also implies that although pessimistic appraisals of the opportunities for social change should have negative consequences – for emotional reactions to the state of society (Badaan et al., 2020; Greenaway et al., 2016) and feelings about the efficacy of minoritized groups to create change (Bleh et al., 2025; Greenaway et al., 2016) – these consequences should be offset by re-investment in the group itself.

Our eventual dataset was not longitudinal at the level of the individual (i.e., it does not include repeated measurement from the same participants). Instead, it was longitudinal at the level of the sample: it involves cross-sectional surveys collected from distinct samples of LGBTQ+ respondents at five time points over an eight-year period. The time period of sampling spans significant shifts in the social-political landscape, especially from a minority perspective, thereby offering insight into how this situation was being metabolized psychologically within the target population. We used structural equation modelling to test a path model in which (1) the perceived possibilities for progressive social change and (2) identification with the minority group were positioned as mediators of emotional and psychological responses to the unfolding state of society and evolving status of the minority group itself, and of vigilant desires to protect minority identity.

Method

We pre-screened potential respondents on Prolific to restrict participation to those at least 18 years of age who did not identify as heterosexual. Respondents were compensated between 0.90 and 1.25 GBP. All but the first sample were restricted to US residents to constrain the focal political context. The full datafile includes full or partial entries from 3828 individuals. In the first sample, we retained for analysis only those respondents who resided in Western cultural contexts (i.e., North America, Western Europe, Australia), reasoning that relevant rights and protections were broadly similar (9 indicated residing in non-Western contexts; 117 did not provide this information and were also excluded). From the full sample, we also excluded a small percentage of participants in each sample who – despite the pre-screen – described themselves as heterosexual in a free-response item (n = 48) or who did not provide a meaningful response to this question (n = 33). The final sample after exclusions included 3627 individuals. Respondents’ age, (US) residence, ((cis) male) gender identity, and (White) ethnicity were controlled in all analyses. Table 1 provides more detail about the samples, available demographics, and the context for data collections.

Table 1

Summary Information About, and Context For, Five Waves of Data Collection

The design and procedure of all data collections were identical. A brief introductory text explained the survey’s focus on non-heterosexual people’s perceptions of “social changes that are happening in many western nations today.” After giving informed consent, participants were asked how they would describe their own sexual identity (open-ended).

Next, they indicated their agreement with six items assessing identification with the LGBTQ+ community. With the one exception noted below, these and all other items were responded to on a 7-point Likert-type scale ranging from 1 = strongly disagree to 7 = strongly agree. This scale comprised two sub-scales from Leach et al.’s (2008) multidimensional measure of social identification: solidarity (example item: I feel committed to the LGBTQ+ community) and centrality (example item: I often think about the fact that I am a member of the LGBTQ+ community). Details of all individual items and all scale reliabilities at each wave are given in SOM #2; Cronbach’s alpha for the identification composite scale ranged from .92 to .94 across waves.

At this point, participants within each sample were randomly assigned to one of three experimental conditions and experienced stimuli related to their condition. Because identification was measured prior to random assignment, note that it cannot be affected by condition. For information about the experimental manipulation and tests relating to its potential impact on the measures that followed, see SOM #3.

Following this, all waves included a core set of common measures, with different additional measures (not mentioned further in this manuscript) in each iteration of the study. These common measures included the following.

First, we measured perceptions of the possibility of progressive change, or participants’ beliefs that progressive social change was occurring and would plausibly continue into the future, with 4 items that we developed (example item: Full equality for sexual minorities within broader society is achievable within the not too distant future). Cronbach’s alpha for the composite scale ranged from .65 to .79 across waves.

We measured participants’ emotions using six positive items (e.g., Hope) and six negative items (e.g., Concern). The order of the twelve items was randomised. The prompt for these items referenced how participants felt about the current social changes they perceived around sexual identities. Unlike the other items, the response format for the emotion items was 1 = not at all to 5 = a lot. Across the five waves, Cronbach’s alpha ranged from .81 to .90 (positive emotions composite) and .83 to .87 (negative emotions composite).

We measured group-based self-definition and group-based satisfaction with items again drawn from Leach et al.’s (2008) measure of in-group identification. The former scale comprised the four items that together measure individual self-stereotyping (i.e., self-group similarity) and ingroup homogeneity (example item: I am similar to the average LGBTQ+ person). Cronbach’s alpha for this composite ranged from .79 to .89 across waves. The latter scale comprised the four satisfaction items (example item: It is pleasant to be a member of the LGBTQ+ community; see Footnote 2). Cronbach’s alpha for this composite ranged from .91 to .92 across waves.

We measured collective efficacy with five items (example: the LGBTQ+ community can usually handle whatever comes our way) adapted from Reicher and Haslam (2006). Cronbach’s alpha for this composite ranged from .92 to .93 across waves.

Lastly, we measured desires to preserve (2 items) versus relinquish (3 items) collective identity as part of the process of social change. Although these five items were intended to measure a single identity maintenance construct, factor analysis indicated two weakly (negatively) correlated sub-dimensions (r = -.13). As indicated in SOM #1, the interitem reliability of the ‘relinquish’ subscale did not reach conventional standards, so we report only the ‘preserve’ subscale here (example item: In addition to pursuing social change, it is important that the LGBTQ+ community protects the unique spaces that it has inhabited in the past). Across the entire sample, the two items in that subscale were correlated at r = .59, p < .001, and across waves, Cronbach’s alpha ranged from .67 to .78.

The common set of measures also included two items assessing the desired speed of social change, which we do not analyze here as it is unrelated to identity or identity vigilance (contact authors for details). The pooled data and an example of the full survey text can be found at https://osf.io/tzu6j.

Results

Preliminary Checks and Analytic Strategy

Data were analysed using IBM SPSS version 30.0.0.0 and jamovi version 2.6.45.0 (The jamovi project, 2025). Prior to analysis, we merged the datasets to create an analysis file containing the common measures. As already noted, all datasets included an experimental manipulation directing participants to contemplate social change as it was manifest in the ingroup, the outgroup, or in an unspecified way in a control condition. Here we instead focus on the non-experimental aspects of these data – namely, the patterns of across-time change on key variables and the structure of correlations among them. Nonetheless, we checked for effects of condition prior to testing time-related patterns.

Analysis testing the originally predicted interaction between condition and identification across the pooled data (via multi-level modelling with a random intercept for wave; using the GAMLj3 module) revealed sporadic and sometimes contradictory effects (see SOM #3 for detail). Similarly, analyses testing interactions between study wave and the experimental manipulation (via a series of general linear models; using the GAMLj3 module) revealed sporadic and sometimes contradictory effects (also SOM #3). Furthermore, a robustness check using only data from the control condition showed the same broad findings we report in Table 2 below (see Table SOM 3-4). Because the effects of condition do not appear to be systematic, we felt comfortable collapsing across conditions for the purpose of the current analyses – however, given the existence of some effects of experimental condition on specific measures at specific times, we control for these when reporting across time effects below.

To understand how the unfolding political context was metabolized psychologically, we first examined mean levels of all constructs as a function of sample wave (which we conceptualise as representing the passage of time) via analysis of variance across all measures (conducted via a MANCOVA in SPSS). Then, we examined correlational relationships among individual variables (controlling for time; partial correlations via jamovi’s default jmv package) before connecting across-time shifts to their consequences for group identity via a path model (run via the pathj module). Code for the main analyses is provided in the supplementary file (see SOM #4).

Across-Time Social Decline

Because individual datasets varied in terms of the proportion identifying as a gender or ethnic majority (i.e., (cis) male, white), age and exclusively US residence (explained above), we control for these demographic factors (along with experimental condition effects) when analysing between-wave differences. Across-time trends are depicted in Figure 1; variable means, standard deviations, and significance tests for comparisons are presented in Table 2.

Across the 8-year period of sampling, we observed declining perceptions of the possibilities for progressive social change. That downward trajectory is accompanied by a consolidation of identity, evidenced by increases across all identity-related measures – identification, group-based self-definition, group-based satisfaction, and collective efficacy – and was also evident in increases in expressed desires to preserve the historical symbols and spaces of identity. Over time emotional reactions to the state of social change were more complex: positive emotions fluctuated somewhat but there was no discernible pattern; negative emotions remained low and stable until the final timepoint (at the beginning of the second Trump presidency) when there was a sharp increase. The portrait that emerges from these across-time comparisons is one that implies not a divestment from collective identification in response to negative social change, but a vigilant (re-)investment.

Figure 1

Trends in Key Variables Across the Eight-Year Period of Data Collection

Table 2

Variable Descriptives as a Function of Sample

Individual Differences and their Correlates

Having identified the across-time trajectory, we next explored the correlates of perceiving increased or reduced possibilities for progressive change at the individual level (i.e., independent of across-time changes). Bivariate correlations controlling for sample/time, age, (cis) male gender, white ethnicity, and US residence are presented in Table 3. (See SOM #5 for correlations broken out by wave.)

Respondents who perceived more potential for progressive social change reported more positive emotion and less negative emotion, as well as greater feelings of collective efficacy, than respondents who perceived less potential. Critically, reflecting the across-time pattern, perceiving progressive change as more possible was negatively correlated with identification. In other words, those who thought things were getting better reported more psychological distance from the group (i.e., lower identification) than those who were sceptical that improvement was happening. Furthermore, they were relatively less desirous of preserving collective identity. As would be expected, individual identification was correlated with all other identity-related outcomes: those who identified more strongly with the minority group also defined themselves more strongly in these terms, reported more satisfaction with their group membership, and experienced stronger collective efficacy; they also reported stronger desires to maintain identity. It is interesting to note that while identification was associated with more positive emotions about social change, it was also associated with more negative emotions, suggesting a certain degree of ambivalence about social change among highly identified minority group members.

Table 3

Correlations between Focal Variables

Indirect Consequences of Across-Time Pessimism about the Potential for Progressive Social Change

The analyses reported above identify an interesting hydraulic relationship between the perceived possibility of progressive change and identity maintenance: minorities report less vigilance about the conservation of identity to the extent that full equality is (or appears) ever closer; and conversely, to the extent that social change is (or appears to be) moving in a negative direction this is associated with higher degrees of vigilance and consolidation around identity. Although reporting reduced possibilities for progressive change was associated with negative emotions and with lower group-based self-definitions and lower feelings of collective efficacy, reporting higher identification was associated positively with these outcomes, as well as with vigilance (which was also correlated with perceptions of declining possibilities).

To more formally test the roles of changing social perceptions and levels of identification with the minority group in explaining across-time patterns on the other variables, we specified a path model. This model included emotions, group-based self-definition, group satisfaction, collective efficacy and desires to preserve identity as parallel outcomes. Following the logic we have been developing above, perceived possibilities for progressive change and identification were positioned as sequential mediators – that is, we assumed that across-time changes in the actual political context were first reflected in perceptions of the possibilities for progressive change and that these perceptions subsequently shaped identification with the minority group.

In this model, Time was the exogenous variable and was treated as categorical – that is, each time point was compared to the previous one (using the ‘repeated’ contrast for the exogenous factor in jamovi) to model change between timepoints. The default coding of this variable produces positive coefficients when the prior time point is higher than the subsequent one (i.e., decline over time), andnegative coefficients when the prior timepoint is lower than the subsequent one (i.e., increases over time). To avoid confusion between the coding produced by jamovi and the meaning of these effects, in Figures 2 and 3, and Tables 4 and 5 below, we have adjusted the sign of all coefficients (and associated confidence intervals, b & z) involving time so that positive coefficients indicate increases over time, and negative coefficients declines. As for the previous analyses, we included demographics and experimental conditions as covariates for all paths in the model. To evaluate model fit we examined the Comparative Fit Index (CFI), the Tucker-Lewis index (TLI), root mean squared error of approximation (RMSEA) and standardized root mean square residual (SRMR), applying the thresholds recommended by Hu and Bentler (1998): CFI > 0.90, TLI > 0.90, SRMR < 0.10, RMSEA < 0.08.

As a precursor to testing our theoretical model, we first fitted a model with only direct effects of time on outcomes with no mediators included. This model was near saturated, did not produce a χ² test for the model, and fit indices are not sensible to interpret: CFI = 1.00, TLI = 1.00, SRMR = .000, RMSEA = .000. For this model, illustrated graphically in Figure 2, we instead focus on the significance of pathways and their coefficients. The patterns of direct effects observed in this model mostly reproduce the patterns reported in Table 2.

Figure 2

Direct Effects of Time on Outcome Variables

The theoretical model, including sequential mediators, was first fitted without covariates. This showed acceptable fit to the data; χ²(24) = 472, p <.001; CFI = 0.961, TLI = 0.902, SRMR = .032, RMSEA = .072. With inclusion of covariates, the model showed improved fit on most indicators, but impaired fit on the TLI: χ² (24) = 350, p <.001; CFI = 0.972, TLI = 0.874, SRMR = .019, RMSEA = .061. Since the TLI punishes model complexity, and multiple covariates were included in all path estimations, we conclude that the specified model shows acceptable fit to the data, even with covariates. We report the results of the model including all covariates. Figure 3 illustrates the theoretical model graphically, including standardised beta weights for each direct path. Table 4 reports tests of the two-step sequential indirect paths from time to perceived possibility of progressive changes to identification to different outcomes. Table 5 reports the component single-step indirect paths connecting time and outcomes via perceived possibility of progressive change and identification separately. For detail of all coefficients in this model see SOM#6.

As Figure 3 and Table 4 show, the negative consequences of across-time downward shifts in the perceived possibility of progressive change were offset by countervailing increases in identification with the minority community. As such, although perceiving decline could have been associated with less positive affect (and more negative affect), and less positive feelings about the group (i.e., group-based self-definition and satisfaction) and its capacities (i.e., collective efficacy), the association between perceived decline and increased identification rendered across-time shifts on all outcomes positive, not negative. The different consequences of perceiving reduced possibilities for progressive change over time on each of our outcomes alone versus indirectly via identification is revealed in the contrasting coefficients of the single-step versus two-step indirect pathways (reported in Tables 5 and 4, respectively). Ultimately, then, perceiving reduced possibilities for progressive change over time were associated with greater consolidation of identity and more vigilant desires to preserve this. The only exception to this pattern was the third temporal transition, between 2020 and 2021 – that is, the period spanning Biden’s presidency. Here the coefficients of all indirect paths via perceived possibilities for progressive change reversed, suggesting that time this period was registered as a reprieve from the downward trajectory that then continued with the return of Trump’s second presidency at the final timepoint.

Figure 3

Path Model of Theorised Sequential Mediation Among Variables

Table 4

Tests of Sequential Indirect Paths Between Timepoint and Different Outcomes.

Table 5

Tests of Component Indirect Paths Between Time Point and Different Outcomes.

Discussion

Across a large and diverse set of samples collected during a period of significant political uncertainty and instability, we find evidence of across-time shifts, patterns of association, and mediated relationships that are all consistent with a vigilant investment in minoritized (sexual) identity in opposition to perceived social decline. As the possibility for progressive social change was perceived as more and more distant, minority identification shifted upward – both across-time and at the level of individual correlations. The hydraulic relationship between perceiving social change as less possible and greater investment in collective identity seemed to protect minority individuals against the otherwise deleterious consequences of perceiving across-time social decline. For example, although perceptions of progressive social change were associated with stronger self-definition in collective terms and stronger feelings of collective efficacy, in step with identification, these variables also developed in an upward direction over time, in spite of the downward trajectory of social change perceptions.

The progressive strengthening of minority identification and collective efficacy observed in our data could lead to an expectation that collective action (e.g., protest) might increase among US-based LGBTQ+ individuals (e.g., van Zomeren et al., 2008). Instead, after an uptick in 2022 and 2023, there has been a more recent reduction in both pro- and anti-LGBTQ+ ‘mobilization’ (ACLED, 2024). Clearly, there are further nuances to the patterns we document here, and future theory and research could productively consider potential moderators of the relationships among social change perceptions, identification and action-related outcomes.

In our data, it is especially striking that faith in collective efficacy remained strong even when perceptions of the possibilities for desired change materialising were low. In our view, the contrast between these two observations underscores our analytic point about the value of vigilant re-investment in minority identity and aligns with the idea of identification as a coping resource (e.g., in Branscombe et al., 1999). Yet, the observed patterns contrast with other research – for example work showing that growing optimism about future prospects within society reinforces both identification and feelings of collective efficacy (Bleh et al., 2025). The different contexts seem relevant to understanding this divergence: the positive implications of growing optimism observed by Bleh and colleagues emerged in the context of environmental movements; our observation of protective identity vigilance in the face of growing pessimism refers to a historically stigmatized group. In line with this, we would suggest that any further unpacking of exactly when, how, and why optimistic versus pessimistic perceptions of society mobilize identity and action should involve close attention to the specifics of the groups being investigated (see Borghi et al., 2025, for a similar argument in relation to anxiety about the future). We would also suggest that this unpacking could incorporate a more dynamic understanding of how narratives of both social progress and social decline can be used in rhetoric to mobilise groups towards action (e.g., see Coenen et al., 2024).

In addition, although the rejection-identification model has been supported in many contexts, there are also observed variabilities across studied groups and times when devaluation is met with identity distancing and the prioritisation of individual coping over collective action (e.g., Liaquat, Balcetis, & Jost, 2025). As such, there is also further theoretical and empirical work to be done to identify more precisely when minorities draw strength from identity in the face of hardship and when identity might instead become a liability when things get really bad. In seeking to advance this literature, some have focused on the individual differences that might predispose people to different pathways of response to devaluing circumstances (e.g., Liaquat et al., 2025). In line with our across-time focus, we would suggest value in considering individual differences in a wider frame that includes the historical trajectories within which groups, and the individuals that comprise these, are embedded.

To further unfold the processes that occur across time and in response to social changes that are not what one may have hoped – or even previously fought – for, integration with additional theories may be fruitful. For example, to better understand the dynamics between identity vigilance versus disengagement and distancing, it might be helpful to draw on perspectives from the stress and coping literature. Classic theory in this domain, for example Selye’s (1950) physiological stress model, argues that responses to a stressor develop and change across time. After an initial ‘alarm’ response, in which the body rapidly activates, the body settles into a second ‘resistance’ phase that can be much more prolonged. In this second phase, the body seeks to adapt while continuing to fight the stressor. The energy it takes to sustain that resistance is substantial and prolonged periods of this can lead to exhaustion and burnout.

Perceiving social decline can be conceived as a stressor (or a threat to well-being; see Wu et al., 2026) and psychological resources are critical for effective resistance in the face of this. In the logic of the rejection-identification model, leaning into identity is a way to mobilize the (individual and collective) psychological resources necessary to fight back. Yet, prolonged resistance may have drawbacks and resources can diminish. It is unclear whether and for how long identity vigilance can be (self-)sustaining, especially if perceived societal decline continues over a much longer period. Thus, over time, the same individuals may go from leaning in to leaning out of their identity as a function of whether evolving social circumstances are recent versus ongoing, acute versus chronic. Practically speaking, for individuals who feel exhausted by prolonged democratic instability and ongoing threats, shared energy across a community (or across communities with common interests; see Rogbeer & Efrén, 2026) could be an under-recognized resource to be harnessed. Especially when individual minorities are not aware that others are “leaning in”, community centers, community organizers, nonprofits, etc. could serve as hubs for connecting individuals and channelling shared energies in ways that build hope in otherwise dark times.

We situated the timeline of our datasets within salient political events in the US. However, this was a period of time in which similar political shifts, including “anti-woke” discourses targeting gender and sexual minorities, were also apparent in other parts of the world. Our project is unable to disentangle perceived decline in the United States from more global perceptions of change in “society” (which could, plausibly, be construed as including other nations). Here, we acknowledge that the status of LGBTQ+ rights looks very different across the globe, and that the vigilant and agentic responses we observe are manifest in a country that remains democratic and liberal – even if these values might be perceived by some in our sample to be declining. In political contexts that are not just negative in tone, but actively threatening to the safety of minority individuals, leaning away from identity – for example, through concealment, passing, or other actions that obscure the visibility of identity – can make very good sense. Along these lines, whether shared identity is embraced, how it is expressed collectively, and how it is experienced by the individual, has to be understood against the backdrop of the material rights and protections afforded by the local context.

It is important to recognize constraints on generalizability of our analysis (see Bauer, 2023). In particular, our samples of US-based LGBTQ+ individuals are based on non-panel convenience sources, and as such are not representative. Furthermore, one might argue that trans and gender-diverse individuals should be more central to our analysis, especially inasmuch as they were directly targeted politically during the 8-year time period we studied. Unfortunately, only the final three samples included a question allowing us to differentiate cis from trans and non-binary participants. Readers may also be curious about differences between various sexual-identity and gender-identity subgroups. Our open-ended question inviting participants to articulate their own identity resulted in a range of rich and interesting descriptions, but did not provide enough contextual information to create defensibly distinct categories on which to base sub-group analyses. Another variable we did not have information about was social class; Prolific does not automatically gather data on socioeconomic status, nor did we collect such information from participants. Lastly, the first sample differs from the later samples in including some non-US-based participants. While this choice renders that sample less comparable to the others, we felt it was important to maximise available information from the 2017 timepoint.

In addition to these sampling-related limitations, there are also limitations relating to causal inference. Our path model, involving mediation, makes untested assumptions about causal direction. The sequential mediators could plausibly be swapped, with identification preceding societal perceptions, and any of the variables designated as outcome could have been situated earlier in the causal chain. The model we tested is theoretically plausible based on SIT, which positions identification as prior to collective appraisals and feelings, and as something that might be impacted by the perceived state of society. From an empirical standpoint, however, this sequence remains highly uncertain, and causality should not be inferred from the correlational data.

All this, in combination with the lack of longitudinal measurement of the same individuals, limits our ability to speak more closely to how societal changes are perceived, metabolized, and responded to. Although we see great value in the data we have collected for documenting unfolding perceptions of societal change and the psychological responses that were engaged by these shifting perceptions, we are also aware that our data are still limited by being temporally organized snapshots. Longitudinal data at the level of the individual participant would speak most powerfully to within-individual across-time changes and their consequences for identity processes. However, datasets spanning timeframes as extended as what we have here are rare, and where they do exist (e.g., in representative population surveys) they rarely contain the necessary granularity of theory-relevant measures.

To address the kinds of questions raised by growing awareness of social decline, a much wider view is needed than the typically severely time-limited cross-sectional or relatively short-term longitudinal studies that currently populate the literature in this domain. We hope our large-scale investigation of the societal perceptions of LGBTQ+ people across successive US presidential administrations provides some necessary movement in this direction. In addition to inspiring further longer-scale work in this area – and specifically more rigorously longitudinal time-series data collection, where possible – we also hope that our data serves as a reminder to researchers that we should not assume that things always just get better.

Conflicts of Interest

The authors declare no competing interests.

Data Availability Statement

Data and materials are available via https://osf.io/tzu6j/.

Supplementary Materials

The supplementary materials can be found here.

Author Contributions

T.M. and J.S. contributed fully and equally to all aspects of this research, including conceptualisation, design, data collection (excepting the 2025 wave), analysis and write-up.

References

Armed Conflict Location Event Data Project (ACLED; July 23, 2024). Pride and protest: A downward trend in LGBTQ+ mobilization this June. Accessed Dec. 16, 2025 at https://acleddata.com/report/pride-and-protest-downward-trend-lgbtq-mobilization-june

Ayoub, P., & Stoeckl, K. (2024). The global resistance to LGBTIQ rights. Journal of Democracy, 35(1), 59-73. https://dx.doi.org/10.1353/jod.2024.a915349

Badaan, V., Jost, J.T., Fernando, J., & Kashima, Y. (2020). Imagining better societies: A social psychological framework for the study of utopian thinking and collective action. Social & Personality Psychology Compass, 14:e12525. https://doi.org/10.1111/spc3.12525

Bauer, P. J. (2023). Generalizations: The grail and the gremlins. Journal of Applied Research in Memory and Cognition, 12(2), 159–175. https://doi.org/10.1037/mac0000106

Bettencourt, B. A., Charlton, K., Dorr, N., & Hume, D. L. (2001). Status differences and in-group bias: A meta-analytic examination of the effects of status stability, status legitimacy, and group permeability. Psychological Bulletin, 127(4), 520–542. https://doi.org/10.1037/0033-2909.127.4.520

Bleh, J., Masson, T., Köhler, S., & Fritsche, I. (2025). From imagination to activism: Cognitive alternatives motivate commitment to activism through identification with social movements and collective efficacy. British Journal of Social Psychology, 64(1), e12811. https://doi.org/10.1111/bjso.12811

Borghi, O., Niraki, M., Seremeta, E., Smets, K., & Tsakiris, M. (2025). Facing a dark future: Young people’s future anxiety and political attitudes in the UK and Greece. advances.in/psychology, 2, e555124. https://doi.org/10.56296/aip00042

Branscombe, N. R., Schmitt, M. T., & Harvey, R. D. (1999). Perceiving pervasive discrimination among African-Americans: Implications for group identification and well-being. Journal of Personality and Social Psychology, 77, 135-149. https://doi.org/10.1037/0022-3514.77.1.135

Coenen, A. C., Jüttemeier, M., Obaidi, M., Power, S. A., & Kunst, J. R. (2024). The role of temporal analogies in collective movements. European Journal of Social Psychology, 54(3), 670-687. https://doi.org/10.1002/ejsp.3037

de la Sablonnière, R., & Tougas, F. (2008). Relative deprivation and social identity in times of dramatic social change: the case of nurses. Journal of Applied Social Psychology, 38(9), 2293-2314. https://doi.org/10.1111/j.1559-1816.2008.00392.x

de la Sablonnière, R., Tougas, F., & Lortie-Lussier, M. (2009). Dramatic social change in Russia and Mongolia: Connecting relative deprivation to social identity. Journal of Cross-Cultural Psychology, 40(3), 327-348. https://doi.org/10.1177/0022022108330986

Doosje, B., Spears, R., & Ellemers, N. (2002). Social identity as both cause and effect: The development of group identification in response to anticipated and actual changes in the intergroup status hierarchy. British Journal of Social Psychology, 41(1), 57-76. https://doi.org/10.1348/014466602165054

Eibach, R. P., & Ehrlinger, J. (2010). Reference points in men’s and women’s judgments of progress toward gender equality. Sex Roles, 63, 882-893. https://doi.org/10.1007/s11199-010-9846-7

Ellemers, N. (1993). The influence of socio-structural variables on identity management strategies. European Review of Social Psychology, 4(1), 27-57. https://doi.org/10.1080/14792779343000013

Esposito, E., & Calanchini, J. (2022). Examining selective migration as attitudinal fit versus gay migration. Journal of Experimental Social Psychology, 101, 104307. https://doi.org/10.1016/j.jesp.2022.104307

Finn, J. (July 2, 2019). The struggle for ‘gay rights’ is NOT over. Accessed Dec. 22, 2025, at https://medium.com/james-finn/the-atlantic-cant-imagine-real-lgbtq-equality-36a0ea0a2d35

Gast, D. L., & Baekey, D. H. (2014). Withdrawal and reversal designs. In Single case research methodology (pp. 211-250). Routledge. https://doi.org/10.4324/9780203521892

Glick, P. (2010). Scapegoating. In The Corsini Encyclopedia of Psychology (Eds. I.B. Weiner and W.E. Craighead). Wiley. https://doi.org/10.1002/9780470479216.corpsy0817

Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Simon & Schuster.

Greenaway, K.H., Cichocka, A., van Veelen, R., Likki, T. & Branscombe, N.R. (2016). Feeling hopeful inspires support for social change. Political Psychology, 37: 89-107. https://doi.org/10.1111/pops.12225

Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd edition). Guilford.

HRC (June 6, 2023). For the first time ever, Human Rights Campaign officially declares ‘state of emergency’ for LGBTQ+ Americans. Accessed Dec. 16, 2025, at https://www.hrc.org/press-releases/for-the-first-time-ever-human-rights-campaign-officially-declares-state-of-emergency-for-lgbtq-americans-issues-national-warning-and-guidebook-to-ensure-safety-for-lgbtq-residents-and-travelers

Hu, L. T., & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods, 3(4), 424-453. https://doi.org/10.1037/1082-989X.3.4.424

Johal, S. K., Batra, R., & Ferrer, E. (2023). Collecting longitudinal data: Present issues and future challenges. In H. Cooper, M. N. Coutanche, L. M. McMullen, A. T. Panter, D. Rindskopf, & K. J. Sher (Eds.), APA handbook of research methods in psychology: Research designs: Quantitative, qualitative, neuropsychological, and biological (2nd ed., pp. 385–408). American Psychological Association. https://doi.org/10.1037/0000319-018

Kelly, C. (1993). Group identification, intergroup perceptions and collective action. European Review of Social Psychology, 4, 59–83. https://doi.org/10.1080/14792779343000022

Kirchick, J. (June 28, 2019). The battle for gay rights is over. Accessed Dec. 19, 2025, at https://www.theatlantic.com/ideas/archive/2019/06/battle-gay-rights-over/592645/

Leach, C. W., van Zomeren, M., Zebel, S., Vliek, M. L. W., Pennekamp, S. F., Doosje, B., Ouwerkerk, J. W., & Spears, R. (2008). Group-level self-definition and self-investment: A hierarchical (multicomponent) model of in-group identification. Journal of Personality and Social Psychology, 95(1), 144–165. https://doi.org/10.1037/0022-3514.95.1.144

Liaquat, U., Balcetis, E., & Jost, J.T. (2025). Multiple identities model of collective inaction: How belonging to psychologically incompatible groups reinforces the status quo. Social and Personality Psychology Compass, 19, e70065. https://doi.org/10.1111/spc3.70065

Major, B., & Schmader, T. (1998). Coping with stigma through psychological disengagement. In C. Stangor & J. Swim (Eds.), Prejudice: The target’s perspective (pp. 219-241). Academic Press.

Mummendey, A., Klink, A., Mielke, R., Wenzel, M., & Blanz, M. (1999). Socio‐structural characteristics of intergroup relations and identity management strategies: Results from a field study in East Germany. European Journal of Social Psychology, 29(2‐3), 259-285. https://doi.org/10.1002/(SICI)1099-0992(199903/05)29:2/3<259::AID-EJSP927>3.0.CO;2-F

Parker, T. (1853). Of justice and the conscience. Downloaded June 2025 from https://justinalmeida.com/wp-content/uploads/2016/07/theodore-parker-justice-and-the-conscience.pdf

Pinker, S. (2012). The better angels of our nature. Penguin.

Reicher, S., & Haslam, S. A. (2006). Rethinking the psychology of tyranny: The BBC prison study. British Journal of Social Psychology, 45(1), 1-40. https://doi.org/10.1348/014466605X48998

Reicher, S. D., & Haslam, S. A. (2012). Change we can believe in: The role of social identity, cognitive alternatives, and leadership in group mobilization and social transformation. In B. Wagoner, E. Jensen, & J. A. Oldmeadow (Eds.), Culture and social change: Transforming society through the power of ideas (pp. 53–73). IAP Information Age Publishing.

Rogbeer, K. G., & Pérez, E. (2026). Solidarity as a bridge: Shared discrimination is indirectly associated with voting intentions among People of Color. advances.in/psychology, 1, e328013. https://doi.org/10.56296/aip00045

Salvatore, J., & Morton, T.A. (2017). Identity challenges in the context of social change. Presented at the annual meeting of the Society for the Psychological Study of Social Issues, Albuquerque, NM, and a preconference to the triennial EASP meeting sponsored by the International Society for Self and Identity, Granada, Spain.

Savage, D., & Miller, T. (Eds.). (2012). It gets better: Coming out, overcoming bullying, and creating a life worth living. Penguin.

Savin-Williams, R. C. (2009). The new gay teenager. Harvard University Press.

Schmitt, M. T., Spoor, J. R., Danaher, K., & Branscombe, N. R. (2009). Rose-colored glasses: How tokenism and comparisons with the past reduce the visibility of gender inequality. In M. Barreto, M. K. Ryan, & M. T. Schmitt (Eds.), The glass ceiling in the 21st century: Understanding barriers to gender equality (pp. 49–71). American Psychological Association. https://doi.org/10.1037/11863-003

Selye, H. (1950). Stress and the general adaptation syndrome. British Medical Journal, 1(4667), 1383-1392. https://doi.org/10.1136/bmj.1.4667.1383

Spoor, J.R., & Schmitt, M.T. (2011). “Things are getting better” isn’t always better: Considering women’s progress affects perceptions of and reactions to contemporary gender inequality. Basic and Applied Social Psychology, 33(1), 24-36. https://doi.org/10.1080/01973533.2010.539948

Smith, L. G., Livingstone, A. G., & Thomas, E. F. (2019). Advancing the social psychology of rapid societal change. British Journal of Social Psychology, 58(1), 33-44. https://doi.org/10.1111/bjso.12292

Steele, C. M., Spencer, S. J., & Aronson, J. (2002). Contending with group image: The psychology of stereotype and social identity threat. In Advances in experimental social psychology (Vol. 34, pp. 379-440). Academic Press.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of social conflict. In Austin, W.G., & Worchel, S. (Eds.), The social psychology of intergroup relations (pp.33-47). Brooks/Cole.

The jamovi project (2025). jamovi (Version 2.6.45.0) [Computer Software]. Retrieved from https://www.jamovi.org

Van Zomeren, M., Postmes, T., & Spears, R. (2008). Toward an integrative social identity model of collective action: a quantitative research synthesis of three socio-psychological perspectives. Psychological bulletin, 134(4), 504. https://doi.org/10.1037/0033-2909.134.4.504

Wright, J. D., Schmitt, M. T., & Mackay, C. M. L. (2022). Access to environmental cognitive alternatives predicts pro-environmental activist behavior. Environment and Behavior, 54(3), 712–742. https://doi.org/10.1177/00139165211065008

Wright, J. D., Schmitt, M. T., Mackay, C. M., & Neufeld, S. D. (2020). Imagining a sustainable world: Measuring cognitive alternatives to the environmental status quo. Journal of Environmental Psychology, 72, 101523. https://doi.org/10.1016/j.jenvp.2020.101523

Wu, D. J., Law, K. F., Syropoulos, S., & Perry, S. P. (2026). The politics of well-being during democratic backsliding: How partisan affiliation and support for government actions relate to happiness and life satisfaction. advances.in/psychology, 1, e569295. https://doi.org/10.56296/aip00051

Yoshino, K. (2007). Covering: The hidden assault on our civil rights. Random House.

Zhang, A., Jetten, J., Iyer, A., & Cui, L. (2013). “It will not always be this way”: Cognitive alternatives improve self-esteem in contexts of segregation. Social Psychological and Personality Science, 4(2), 159-166. https://doi.org/10.1177/1948550612452890