Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 May 13;64(3):e12900. doi: 10.1111/bjso.12900

Primary and secondary generalization effects from Black and gay contact: Longitudinal evidence of between‐ and within‐person effects

Hanna Puffer 1,✉, Gordon Hodson 1
PMCID: PMC12076039  PMID: 40364634

Abstract

The contact hypothesis stipulates that contact between social groups can reduce intergroup prejudice, implying that contact changes people (i.e., within‐person effects). However, recent research suggests that more intergroup contact might simply be associated with less intergroup prejudice (i.e., between‐person effects). We explore primary but also secondary contact effects, whereby contact with one outgroup theoretically improves attitudes towards other uninvolved groups. White, heterosexual Americans' contact with Black and gay people was assessed at four timepoints, 3 weeks apart (T1 N = 456; 51.6% women, M age = 46.71, SD = 15.30); multilevel modelling parsed between‐ from within‐person contact effects on intergroup outcomes (attitudes, humanization, collective action intentions). We found consistent evidence of predicted primary contact effects, reflecting both within‐ and between‐subjects relations. For secondary contact, between‐subjects gay‐to‐Black associative generalization was observed: greater contact (quantity and quality) with gay people was observed among those expressing more positive Black intergroup outcomes. Within‐subjects secondary effects were primarily observed in terms of assessing contact quantity, where more contact with Black people predicted more positive gay intergroup outcomes downstream (i.e., Black‐to‐gay process generalization). Contrary to recent concerns, the current study promisingly shows that contact with a primary outgroup can change people in ways that generate positive outcomes towards primary and (some) secondary outgroups.

Keywords: intergroup contact, multilevel modelling, secondary transfer effect

INTRODUCTION

In his seminal work, Allport (1954) proposed the contact hypothesis, positing that contact between members of social groups can reduce intergroup prejudice. Indeed, research shows that the average contact effect is r = −.22 (Pettigrew & Tropp, 2006); the more contact one experiences with members of an outgroup, the less prejudice one expresses towards that group (Dovidio et al., 2017; Hodson & Hewstone, 2013). In the decades since the contact hypothesis was first formulated, the effect has proven to be quite robust, with contact effects observed across different types of contact, including imagined contact (Miles & Crisp, 2014), intergroup friendship (Page‐Gould et al., 2008), and media contact (Lissitsa & Kushnirovich, 2020), as well as among children (Beelmann & Heinemann, 2014; Vezzali et al., 2018), real‐world samples outside the lab (Albuja et al., 2024; Lemmer & Wagner, 2015), non‐WEIRD (Western, Educated, Industrialized, Rich, Democratic) contexts (Van Assche et al., 2023), and even in high‐conflict areas (Paolini et al., 2014). Furthermore, contact effects are observed when considering various types of intergroup outcomes, including (but not limited to) intergroup attitudes (Boin et al., 2024; De Coninck et al., 2021; Fuochi et al., 2020), collective action intentions (willingness to take social or political action to enhance the lives of an outgroup; Di Bernardo et al., 2019; Górska & Tausch, 2023), and humanization of the outgroup (Capozza et al., 2013, 2017).

Between‐ versus within‐subjects effects of contact

Importantly, contact effects are theorized to play out over time. Indeed, previous longitudinal contact research shows evidence of intergroup contact reducing prejudice towards the primary outgroup (Dhont et al., 2011, 2014; Vezzali et al., 2010, 2023). Of note however, intergroup contact research employing longitudinal designs have historically utilized cross‐lagged panel modelling (CLPM) to test the bidirectional effects of contact on intergroup attitudes (see Albarello et al., 2020; Dhont et al., 2014; Meleady et al., 2019; Swart et al., 2011). For example, Tropp et al. (2012) used CLPM to assess the bidirectional relationships between ethnic group (African Americans, Latino Americans, and Asian Americans) friendships with White people, minority participants' perceived discrimination, and ethnic activism at three timepoints. However, CLPM conflates within‐person effects (i.e., temporal variance or change) and between‐person effects (i.e., stable variance) (Hamaker et al., 2015; Lucas, 2023; Mulder & Hamaker, 2021). Researchers are increasingly recommending analytic procedures that isolate rather than confound these effects, and for good reason. Theoretically, the effects of intergroup contact are assumed to represent within‐person effects (i.e., an individual's increase in contact predicts an increase in their favourable attitudes downstream), not between‐person effects (i.e., those experiencing greater contact also reporting more positive attitudes) (Friehs et al., 2024). Yet, recent research by Friehs and colleagues found significant between‐subjects, not within‐subjects, effects of contact on intergroup attitudes (see also Bohrer et al., 2019; Hodson & Meleady, 2024; McKeown et al., in press; Sengupta et al., 2023; Shulman et al., in press).

Such findings draw into question earlier interpretations of CLPM‐based contact research, which were previously assumed to reflect within‐person effects but may instead reflect between‐person differences. As Friehs et al. (2024) note, finding between‐person but not within‐person effects potentially contradicts the premise of the contact hypothesis; the tacit assumption underlying the contact hypothesis is that intergroup contact changes individuals in ways that lead to reduced prejudice. Promisingly, recent primary contact research by Boin et al. (2024) found both between‐subjects and within‐subjects of contact when considering various outcomes towards immigrants in Italy (i.e., attitudes, prejudice, and outgroup variability). Similarly, Górska and Tausch (2023) found within‐subjects effects of Polish peoples' contact with Ukrainians on collective action intentions towards the outgroup, but not on intergroup attitudes. Overall, the field is amassing evidence that within‐person change is often not observed for attitudes or prejudice but is often observed for other, related intergroup variables. In light of these new discoveries, researchers have called for more longitudinal contact research employing analyses that parse apart between‐ and within‐subjects effects to better understand the effects of intergroup contact on intergroup attitudes (see Hodson & Meleady, 2024). Although the random intercept cross‐lagged panel model (RI‐CLPM) has been used to separate between‐ and within‐person effects in the contact field, longitudinal multilevel modelling (MLM) is a related, well‐established analytic procedure that also parses apart these different effects, and has been widely used by contact researchers for this purpose (e.g., Boin et al., 2020, 2024; Wu et al., 2024; Yip et al., 2013).

Secondary contact

If contact effects reflect between‐person differences rather than within‐person changes, this has implications not only for primary contact, but also for secondary contact, the potential for contact effects to extend to groups not directly involved in the contact situation, commonly referred to as secondary transfer effects (STEs; Pettigrew, 2009). Whereas primary contact effects refer to contact effects on outcomes towards the primary outgroup, secondary contact effects refer to the effect of intergroup contact on outcomes towards secondary, uninvolved outgroups. Of note, whenever we discuss ‘secondary contact effects’ or ‘secondary effects’ in the present article, we are referring to STEs. Pettigrew (2009), for instance, found that Germans' positive contact with foreign residents was associated with positive attitudes towards Muslims, homeless people, and gay men/lesbians. More recently, Vezzali et al. (2022) demonstrated that British and Italian peoples' quality of contact with Eastern Europeans/immigrants was associated with positive outcomes towards Roma people, refugees, people with disabilities, gay men/lesbians, and obese people. Furthermore, Tausch et al. (2010) examined contact effects in Northern Ireland, which is largely White (Protestant and Catholic), finding that contact with religious outgroups was associated with less prejudice towards non‐White racial minorities.

With regard to secondary contact, between‐ and within‐subjects contact effects theoretically reflect different types of generalization. Between‐subjects effects can reflect associative generalization, where there is meaningful association between contact and prejudice across target groups but not of the nature that is considered an action (i.e., change) or process being captured during the interval of investigation (but rather was pre‐existing). Such generalization is relevant to generalized prejudice, the “generalization of devaluing sentiments across different group domains” (Bergh et al., 2016, p. 368). Generalized prejudice reflects a broader, underlying tendency whereby people who are prejudiced against one group (e.g., immigrants) tend to also be prejudiced against other groups (e.g., women; Hodson et al., 2017; Hodson & Puffer, in press). Relatedly, associative generalization might reflect a general tendency whereby those who experience more contact with some outgroups (e.g., gay people) are more favourable towards other outgroups (e.g., Black people). As between‐person effects, such links reflect associations without assertions or claims about directional or causal effects. In contrast, within‐subjects effects can reflect what we term process generalization, where generalization effects occur within people over time. For instance, contact with Black people might generalize by generating more favourable gay attitudes over time, as is theoretically implied by the STE.

Present investigation

Here we consider outgroup attitudes, humanization, and collective action as criteria of longitudinal contact effects. Both collective action and humanization have become increasingly of interest to contact research (e.g., Capozza et al., 2017; Górska & Tausch, 2023; Hodson & Meleady, 2024; MacInnis & Hodson, 2019), with researchers increasingly encouraging the field to not solely study attitudes (Dixon & Levine, 2012; Tropp & Mallett, 2011). In terms of humanization, previous research, both experimental and cross‐sectional, reveals that contact predicts greater humanization of outgroups (Capozza et al., 2013, 2017). Additionally, Prati and Loughnan (2018) found that imagined contact with Gypsy (i.e., Roma) and Japanese people was associated with increased attributions of human uniqueness and human nature, respectively. In terms of collective action intentions, Vezzali et al. (2022) found that advantaged group members' contact with immigrants was not only associated with greater collective action intentions towards immigrants, but also towards Roma people and gay men and lesbians. Of particular interest to the present study, Górska and Tausch (2023) assessed Polish people's contact with Ukranian people and found within‐person effects of cross‐group friendship predicting greater collective action intentions towards Ukranian people. Furthermore, Puffer and Hodson (2024) found that contact with gay people was indirectly associated with positive intergroup outcomes towards Black people through humanization of Black people. Thus, humanization can conceptually be considered as an outcome of intergroup contact, but potentially also a mediator of contact effects.

Our goal is to test primary and secondary longitudinal contact effects on these intergroup outcomes, while disentangling within versus between‐person variance. Specifically, we assess primary and secondary contact effects with Black and gay people as target outgroups, because they are generally viewed as distinct groups (Petsko & Bodenhausen, 2019), thus providing a relatively conservative test of generalization effects. Related research examining not contact but rather regional diversity has found that residing in more racially diverse counties is associated with less stereotypic implicit attitudes about sexual orientation (Tiv & Spence, 2025). Of note, previous contact research has not often examined how contact effects generalize across these two groups. In an exception, Puffer and Hodson (2024) examined the generalization of Black and gay contact effects in a cross‐sectional sample, finding that White heterosexual people's contact with gay people generalizes to outcomes towards Black people, but that contact with Black people did not generalize to outcomes towards gay people. However, being a cross‐sectional sample, its methodology could not parse apart between‐ and within‐person contact effects of the sort needed to differentiate associative generalization from process generalization.

We hypothesized that contact with a primary outgroup will: (1) predict prejudice (negatively) and collective action intentions (positively) towards the primary outgroup; and (2) predict prejudice (negatively) and collective action intentions (positively) towards the secondary outgroup, while controlling for contact with the secondary group,1 in keeping with the spirit of our original preregistration that considered contact as the predictor and the other three variables (i.e., outgroup attitudes, humanization, and collective action intentions) as criteria.

METHODS

Participants and procedure

Hypotheses and data analysis plans were preregistered via AsPredicted.org. (https://aspredicted.org/zdnt‐rmvk.pdf). Data were collected from an online CloudResearch Connect sample of 500 White, heterosexual US residents, population‐matched on age, gender, and ethnicity (i.e., Hispanic origin). This sample size was chosen to optimize both sample size and number of assessments (i.e., timepoints), both of which are factors that increase power in multilevel analysis (Lane & Hennes, 2018), while taking financial constraints into consideration. Data are available online at https://osf.io/dekj5/). Participants completed surveys at four timepoints separated by 3‐week intervals. They were removed from the analytic sample if they failed two attention checks embedded in the survey, did not respond to majority of the measures (i.e., they began the survey but exited out of it after completing a few items), did not identify as White,2 or identified as an LGBTQ+ member, leaving a final sample of 456 participants at the first timepoint (48.4% men, 51.6% women, M age = 46.71, SD = 15.30).

We retained a large proportion of participants across timepoints: 405 (88.82% of Wave 1 sample) at Wave 2, 371 (81.36% of Wave 1 sample) at Wave 3, 350 participants (76.75% of Wave 1 sample) at Wave 4. Participants did not need to complete all four surveys to be kept in the analytic sample, providing that they completed the first wave. Any missing Level 1 variables were handled using pairwise deletion (Woltman et al., 2012).

This study received approval by the Brock University Research Ethics Board. All participants provided informed consent prior to beginning each survey.

Materials

Quantity of contact

To assess quantity (i.e., frequency) of contact, a three‐item measure assessed how much contact participants experienced with each outgroup over the past week, once for Black people and once for gay people, adapted from Hodson et al. (2009). Example items include: ‘How much contact did you have with [Black/gay] people?’ (1 = none, 7 = a lot), and ‘How many hours per week did you typically spend interacting with [Black/gay] people?’ (1–10+). An overall score was created by standardizing each item and averaging the standardized items for both Black and gay people respectively. Higher scores reflected greater quantity of contact with Black people (T1 α = .92, T2 α = .93, T3 α = .94, T4 α = .95) and gay people (T1 α = .94, T2 α = .96, T3 α = .94, T4 α = .96).

Quality of contact

To assess quality (i.e., valence) of contact, a three‐item measure assessed perceived contact valence regarding each outgroup (Hodson et al., 2009; Voci & Hewstone, 2003) over the previous week. Example items include: ‘Did you find the contact pleasant?’ (1 = not at all, 7 = very), and ‘Did you find the contact cooperative?’ (1 = not at all, 7 = very). Higher scores reflected greater quality of contact with Black people (T1 α = .68, T2 α = .72, T3 α = .72, T4 α = .75) and gay people (T1 α = .74, T2 α = .73, T3 α = .68, T4 α = .72).

Humanization

We conducted a pilot study to determine the most viable humanization measure (see Appendix S1 for a discussion of the pilot study). From that study, we chose a humanization measure adapted from Santoro and Monin (2023), whereby respondents attributed perceived capacity for Black and gay people to have uniquely human experiences (i.e., if Black and gay people typically have experiences that are unique to humans). Example items include: ‘On average, do you think [Black/gay] people typically have culture?’ (1 = not at all, 7 = extremely), and ‘On average, do you think [Black/gay] people are moral?’ (1 = not at all, 7 = extremely). Higher scores reflected greater attribution of human‐like experiences to Black people (T1 α = .93; T2 α = .94, T3 α = .93, T4 α = .94) and gay people (T1 α = .92, T2 α = .93, T3 α = .94, T4 α = .94).

Outgroup attitudes

Participants rated their feelings towards Black people and gay people from 0 (extremely unfavourable attitudes) to 100 (extremely favourable attitudes) (Hodson & Meleady, 2024; Vezzali et al., 2022).

Collective action

Participants completed a measure adapted from Choma et al. (2021) tapping willingness take action to enhance the social and political lives of outgroup, regarding Black people and gay people. An example item reads: ‘How willing would you be to sign a petition to support the [Black/gay] community?’ (1 = not at all willing, to 7 = extremely willing). Higher scores reflected greater collective action intentions towards Black people (T1 α = .93, T2 α = .93, T3 α = .93, T4 α = .94) and gay people (T1 α = .95, T2 α = .95, T3 α = .95, T4 α = .95).

RESULTS

As per our preregistration, scores from outliers greater than 3 standard deviations (SD) from mean scores were Winsorized (i.e., given a score at 3 SD from the mean). Prior to model testing, the data were tested for normality, revealing largely normally distributed variables, although the data violated the assumption of homoscedasticity (implying that there is variability in residuals across levels of the predictor variables). Although there is not a completely uniform distribution of residuals, the violation of homoscedasticity is not extreme. Prior to analysis, data were assessed for missing data and attrition. With respect to missing data, 46.5% (n = 212) of the sample was missing on at least one model variable. Little's (1988) Missing Completely at Random test indicated that data were not missing completely at random, χ 2(2134) = 2280.67, p = .014 (See the Appendix S1 for discussion of how those who were missing variables and timepoints differed from those not missing on any variables or timepoints). Please see Tables S7–S10 and Figures S9–S12 for full results of the tests of normality, Figures S13–S22 for stability of the variables across time, as well to Table S11 for correlations between key variables; see Table 1 for descriptive statistics of key variables.

TABLE 1.

Descriptive statistics for key variables.

Variable N Scale range M SD Skewness Kurtosis α
Quantity Black Contact T1 456 0–1 0.00 0.93 0.75 −0.55 .92
Quantity Black Contact T2 405 – 0.00 0.94 0.77 −0.47 .93
Quantity Black Contact T3 371 – 0.00 0.94 0.74 −0.59 .94
Quantity Black Contact T4 350 – 0.00 0.96 0.75 −0.63 .95
Quantity Gay Contact T1 456 0–1 0.00 0.95 1.18 0.31 .94
Quantity Gay Contact T2 405 – 0.00 0.96 1.31 0.64 .96
Quantity Gay Contact T3 370 – 0.01 0.95 1.25 0.61 .94
Quantity Gay Contact T4 349 – −0.01 0.96 1.05 −0.09 .96
Quality Black Contact T1 442 1–7 5.85 1.14 −0.91 0.12 .68
Quality Black Contact T2 389 – 5.78 1.21 −0.81 −0.19 .72
Quality Black Contact T3 354 – 5.80 1.17 −0.78 −0.16 .72
Quality Black Contact T4 334 – 5.85 1.18 −0.86 −0.24 .75
Quality Gay Contact T1 314 1–7 5.53 1.43 −0.91 0.21 .74
Quality Gay Contact T2 370 – 5.59 1.34 −0.76 −0.33 .73
Quality Gay Contact T3 334 – 5.51 1.35 −0.56 −0.60 .68
Quality Gay Contact T4 323 – 5.59 1.39 −0.78 −0.10 .72
Black Humanization T1 455 1–7 6.25 0.72 −1.04 0.92 .93
Black Humanization T2 405 – 6.28 0.71 −1.22 1.32 .94
Black Humanization T3 369 – 6.29 0.69 −0.93 0.40 .93
Black Humanization T4 349 – 6.28 0.72 −1.02 0.77 .94
Gay Humanization T1 455 1–7 5.97 0.92 −1.04 0.84 .92
Gay Humanization T2 405 – 5.95 0.95 −1.12 1.05 .93
Gay Humanization T3 369 – 5.96 1.00 −1.28 1.50 .94
Gay Humanization T4 349 – 6.01 0.95 −1.10 1.08 .94
Black Attitudes T1 455 0–100 78.97 20.15 −0.78 −0.36 –
Black Attitudes T2 404 – 80.20 20.52 −0.91 −0.06 –
Black Attitudes T3 369 – 78.35 20.69 −0.80 −0.19 –
Black Attitudes T4 349 – 78.20 21.33 −0.85 −0.15 –
Gay Attitudes T1 422 0–100 73.82 25.42 −0.90 0.08 –
Gay Attitudes T2 403 – 73.69 26.36 −0.94 0.08 –
Gay Attitudes T3 369 – 72.44 26.42 −0.92 0.11 –
Gay Attitudes T4 349 – 72.86 25.92 −0.91 −0.16 –
Black Collective Action T1 454 1–7 4.47 1.75 −0.28 −0.93 .93
Black Collective Action T2 403 – 4.36 1.79 −0.22 −1.10 .93
Black Collective Action T3 367 – 4.12 1.81 −0.12 −1.20 .93
Black Collective Action T4 350 – 4.03 1.86 −0.04 −1.24 .94
Gay Collective Action T1 453 1–7 3.98 1.98 −0.04 −1.27 .95
Gay Collective Action T2 404 – 3.88 1.99 −0.02 −1.32 .95
Gay Collective Action T3 368 – 3.74 1.94 0.07 −1.30 .95
Gay Collective Action T4 350 – 3.64 1.98 0.16 −1.28 .95
Age 456 18–100 46.71 15.30 – – –
Open in a new tab

Note: Quantity of contact variables are standardized. Age of participants was only collected at T1. Attitudes were scored so that higher values represent more favourable attitudes.

Analytic strategy

We analysed our data with multilevel modelling using HLM Version 8 (Raudenbush & Congdon, 2021), testing six overall models. Model 1 involved quantity of contact with each outgroup predicting attitudes towards each outgroup. Model 2 involved quality of contact with each outgroup predicting attitudes towards each outgroup. Model 3 involved quantity of contact with each outgroup predicting humanization towards each outgroup. Model 4 involved quality of contact with each outgroup predicting humanization towards each outgroup. Model 5 involved quantity of contact with each outgroup predicting collective action intentions towards each outgroup. Model 6 involved quality of contact with each outgroup predicting collective action intentions towards each outgroup.

The intraclass correlations for Black attitudes (ICC = .78), gay attitudes (ICC = .82), Black collective action intentions (ICC = .82), gay collective action intentions (ICC = .87), Black humanization (ICC = .66), and gay humanization (ICC = .76) indicated clustering within individuals that should be accounted for in our analyses. We conducted two‐level models with a random intercept for each participant, allowing the average amounts of each outcome to vary between individuals. We included Time in each model to account for any third variables that may covary with time as a fixed effect. Additionally, Time was entered as a random slope to allow for different trajectories over time (Bolger & Laurenceau, 2013).

When entering the variables into the models, a between‐subjects version of each predictor was created by averaging participants' responses across time and subtracting the grand mean from their response (i.e., grand‐mean centring; Hayes, 2006). A within‐subjects version of each predictor was created by subtracting the participants' overall mean from their score at each timepoint (i.e., group mean centring; Hayes, 2006). Both versions of our two predictors (i.e., the between‐subjects and within‐subjects predictors) were entered into models to simultaneously test between‐ and within‐person effects.

Model results

For each model, we first report results from between‐subjects level tests (reflecting stable, between‐person associations), and then results from within‐subjects level tests (reflecting changes within people). We report unstandardized betas with robust standard errors for all results.

Model 1: Quantity of contact predicting attitudes

Primary contact

At the between‐subjects level, greater Black contact was associated with more positive Black attitudes (b = 4.13, p < .001; see Table 2 and Figure 1). Greater gay contact was associated with more positive gay attitudes (b = 8.62, p < .001). At the within‐subjects level, more frequent Black contact did not significantly predict Black attitudes (b = 0.31, p = .629), but more frequent gay contact significantly predicted more positive gay attitudes (b = 1.75, p = .022; see Figure 2).

TABLE 2.

Model 1 results of contact quantity predicting outgroup attitudes.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quantity of Contact Black Attitudes 0.31 .629 4.13 <.001
Gay Quantity of Contact −0.33 .595 1.38 .191
Black Quantity of Contact Gay Attitudes 0.16 .811 −2.34 .093
Gay Quantity of Contact 1.75 .022 8.62 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 1.

FIGURE 1

Open in a new tab

Model 1 between‐subjects results of contact quantity predicting outgroup attitudes. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 2.

FIGURE 2

Open in a new tab

Model 1 within‐subjects results of contact quantity predicting outgroup attitudes. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

Secondary contact

At the between‐subjects level, greater gay contact was not associated with Black attitudes (b = 1.38, p = .191), and greater Black contact was not associated with gay attitudes (b = −2.34, p = .093; Figure 1). At the within‐subjects level, greater gay contact did not predict Black attitudes (b = −0.33, p = .595), and greater Black contact did not predict gay attitudes (b = 0.16, p = .811; Figure 2).

Model 2: Quality of contact predicting attitudes

Primary contact

At the between‐subjects level, greater quality Black contact was related to more positive Black attitudes (b = 10.09, p < .001; Table 3 and Figure 3), and greater quality gay contact was related to more positive gay attitudes (b = 13.04, p < .001). At the within‐subjects level, greater quality Black contact significantly predicted more positive Black attitudes (b = 2.50, p < .001), and greater quality gay contact significantly predicted more positive gay attitudes (b = 1.86, p < .001; see Figure 4).

TABLE 3.

Model 2 results of contact quality predicting outgroup attitudes.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quality of Contact Black Attitudes 2.50 <.001 10.09 <.001
Gay Quality of Contact 0.20 .611 0.62 .444
Black Quality of Contact Gay Attitudes 0.92 .099 0.73 .540
Gay Quality of Contact 1.86 <.001 13.04 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 3.

FIGURE 3

Open in a new tab

Model 2 between‐subjects results of contact quality predicting outgroup attitudes. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 4.

FIGURE 4

Open in a new tab

Model 2 within‐subjects results of contact quality predicting outgroup attitudes. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

Secondary contact

At the between‐subjects level, greater quality gay contact was not related to Black attitudes (b = 0.62, p = .444), and greater quality Black contact was not related to gay attitudes (b = 0.73, p = .540; see Figure 3). At the within‐subjects level, greater quality gay contact did not predict Black attitudes (b = 0.20, p = .611), and greater quality Black contact did not predict gay attitudes (b = 0.92, p = .099; see Figure 4).

Model 3: Quantity of contact predicting humanization

Primary contact

At the between‐subjects level, more frequent Black contact was associated with Black humanization (b = 0.07, p = .033; Table 4 and Figure 5), and more frequent gay contact was associated with gay humanization (b = 0.26, p < .001). At the within‐subjects level, more frequent Black contact directly predicted Black humanization (b = 0.08, p = .014), but more frequent gay contact did not predict gay humanization (b = 0.04, p = .138; see Figure 6).

TABLE 4.

Model 3 results of contact quantity predicting outgroup humanization.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quantity of Contact Black Humanization 0.08 .014 0.07 .033
Gay Quantity of Contact 0.00 .998 0.09 .004
Black Quantity of Contact Gay Humanization 0.10 .007 −0.05 .292
Gay Quantity of Contact 0.04 .138 0.26 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts.

FIGURE 5.

FIGURE 5

Open in a new tab

Model 3 between‐subjects results of contact quantity predicting outgroup humanization. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 6.

FIGURE 6

Open in a new tab

Model 3 within‐subjects results of contact quantity predicting outgroup humanization. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

Secondary contact

At the between‐subjects level, greater gay contact was related to Black humanization, also reflecting associative generalization (b = 0.09, p = .004; see Figure 5), but greater Black contact was not related to gay humanization (b = −0.05, p = .292). At the within‐subjects level, greater gay contact did not predict Black humanization (b = 0.00, p = .998), but greater Black contact directly predicted gay humanization (b = 0.10, p = .007; see Figure 6), again reflecting process generalization.

Model 4: Quality of contact predicting humanization

Primary contact

At the between‐subjects level, greater quality Black contact was associated with Black humanization (b = 0.27, p < .001; Table 5 and Figure 7), and greater quality gay contact was associated with gay humanization (b = 0.40, p < .001). At the within‐subjects level, greater quality Black contact directly predicted Black humanization (b = 0.09, p < .001), and greater quality gay contact predicted gay humanization (b = 0.10, p < .001; see Figure 8).

TABLE 5.

Model 4 results of contact quality predicting outgroup humanization.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quality of Contact Black Humanization 0.09 <.001 0.27 <.001
Gay Quality of Contact 0.02 .236 0.08 .002
Black Quality of Contact Gay Humanization 0.02 .308 0.07 .121
Gay Quality of Contact 0.10 <.001 0.40 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts.

FIGURE 7.

FIGURE 7

Open in a new tab

Model 4 between‐subjects results of contact quality predicting outgroup humanization. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

FIGURE 8.

FIGURE 8

Open in a new tab

Model 4 within‐subjects results of contact quality predicting outgroup humanization. Dotted lines represent non‐statistically significant pathways. Black/gay attitudes refers to attitudes towards Black/gay people.

Secondary contact

At the between‐subjects level, greater quality gay contact was related to Black humanization (b = 0.08, p = .002; see Figure 7), also reflecting associative generalization, but greater quality Black contact was not related to gay humanization (b = 0.07, p = .121). At the within‐subjects level, greater quality gay contact did not predict Black humanization (b = 0.02, p = .236), and greater quality Black contact did not directly predict gay humanization (b = 0.02, p = .308; see Figure 8).

Model 5: Quantity of contact predicting collective action intentions

Primary contact

At the between‐subjects level, greater Black contact was associated with Black collective action intentions (b = 0.38, p < .001; Table 6 and Figure 9), and greater gay contact was associated with more gay collective action intentions (b = 1.02, p < .001). At the within‐subjects level, greater Black contact significantly predicted more Black collective action intentions (b = 0.12, p = .039), but greater gay contact did not significantly predict gay collective action intentions (b = 0.01 p = .784; see Figure 10).

TABLE 6.

Model 5 results of contact quantity predicting collective action intentions.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quantity of Contact Black Collective Action 0.12 .039 0.38 <.001
Gay Quantity of Contact −0.05 .301 0.59 <.001
Black Quantity of Contact Gay Collective Action 0.18 .002 −0.03 .756
Gay Quantity of Contact 0.01 .784 1.02 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts.

FIGURE 9.

FIGURE 9

Open in a new tab

Model 5 between‐subjects results of contact quantity predicting collective action intentions. Dotted lines represent non‐statistically significant pathways.

FIGURE 10.

FIGURE 10

Open in a new tab

Model 5 within‐subjects results of contact quantity predicting collective action intentions. Dotted lines represent non‐statistically significant pathways.

Secondary contact

At the between‐subjects level, more frequent gay contact was related to Black collective action intentions (b = 0.59, p < .001), reflecting associative generalization, but more frequent Black contact was not related to gay collective action intentions (b = −0.03, p = .756; see Figure 9). At the within‐subjects level, more frequent gay contact did not predict Black collective action intentions (b = −0.05, p = .301), but more frequent Black contact predicted gay collective action intentions (b = 0.18, p = .002; see Figure 10), reflecting process generalization.

Model 6: Quality of contact predicting collective action intentions

Primary contact

At the between‐subjects level, greater quality Black contact was associated with Black collective action intentions (b = 0.34, p < .001; Table 7, Figure 11). Greater quality gay contact was associated with more gay collective action intentions (b = 0.87, p < .001). At the within‐subjects level, greater quality Black contact significantly predicted more Black collective action intentions (b = 0.10, p = .005), but greater quality gay contact did not significantly predict gay collective action intentions (b = 0.05 p = .153; see Figure 12).

TABLE 7.

Model 6 results of contact quality predicting collective action intentions.

Predictor Outcome Level 1 (within‐subjects) Level 2 (between‐subjects)
b p b p
Black Quality of Contact Black Collective Action 0.10 .005 0.34 <.001
Gay Quality of Contact 0.00 .925 0.33 <.001
Black Quality of Contact Gay Collective Action 0.04 .221 −0.05 .619
Gay Quantity of Contact 0.05 .153 0.87 <.001
Open in a new tab

Note: All coefficients use robust standard errors. Time (i.e., timepoint) was included each model to account for any third variables that may covary with time as a fixed effect. All predictors are modelled with random intercepts.

FIGURE 11.

FIGURE 11

Open in a new tab

Model 6 between‐subjects results of contact quality predicting collective action intentions. Dotted lines represent non‐statistically significant pathways.

FIGURE 12.

FIGURE 12

Open in a new tab

Model 6 within‐subjects results of contact quality predicting collective action intentions. Dotted lines represent non‐statistically significant pathways.

Secondary contact

At the between‐subjects level, greater quality gay contact was related to Black collective action intentions (b = 0.33, p < .001), reflecting associative generalization, but greater quality Black contact was not related to gay collective action intentions (b = −0.05, p = .619; see Figure 11). At the within‐subjects level, greater quality gay contact did not predict Black collective action intentions (b = 0.00, p = .925), and greater quality Black contact did not predict gay collective action intentions (b = 0.04, p = .221; see Figure 12).

GENERAL DISCUSSION

Intergroup contact is recognized as one of the most influential ways to reduce prejudice (Hodson & Hewstone, 2013; Pettigrew & Tropp, 2006), involving both primary and secondary effects. However, recent longitudinal contact findings draw into question primary contact effects, with unclear implications for secondary contact. We weigh in on this issue with four waves of data across 3‐week intervals, assessing experiences with two unrelated groups and three types of criteria, and using statistical methods that parse apart between‐ and within‐subjects effects.

A conceptual summary of the results from our study can be found in Tables 8 and 9. We found clear evidence of the basic contact hypothesis (i.e., primary contact effects) at the between‐subjects and within‐subjects levels. More specifically, both quantity and quality of contact with the primary outgroup predicted positive outcomes towards the primary outgroup. Specifically, quantity and quality of contact predicted humanization, attitudes, and collective action intentions towards the primary outgroup. These results were significant for virtually every model tested, at both the between‐ and within‐subjects level (with the following exceptions at the within‐subjects level: gay contact (both types) predicting gay collective action intentions; gay contact quantity predicting gay humanization; Black contact quantity predicting Black attitudes).

TABLE 8.

Summary of contact quantity effects.

Level 1 (within‐subjects) Level 2 (between‐subjects)
Black → Black Gay → Gay Black → Gay Gay → Black Black → Black Gay → Gay Black → Gay Gay → Black
Primary contact
Contact → Attitudes ✘ ✔ – – ✔ ✔ – –
Contact → Collective Action ✔ ✘ – – ✔ ✔ – –
Contact → Humanization ✔ ✘ – – ✔ ✔ – –
Secondary contact
Contact → Attitudes – – ✘ ✘ – – ✘ ✘
Contact → Collective Action – – ✔ ✘ – – ✘ ✔
Contact → Humanization – – ✔ ✘ – – ✘ ✔
Open in a new tab

Note: Tick marks signify statistically significant results (p < .05); crosses indicate results that do not reach statistical significance.

TABLE 9.

Summary of contact quality effects.

Level 1 (within‐subjects) Level 2 (between‐subjects)
Black → black Gay → gay Black → gay Gay → black Black → black Gay → gay Black → gay Gay → black
Primary contact
Contact → Attitudes ✔ ✔ – – ✔ ✔ – –
Contact → Collective Action ✔ ✘ – – ✔ ✔ – –
Contact → Humanization ✔ ✔ – – ✔ ✔ – –
Secondary contact
Contact → Attitudes – – ✘ ✘ – – ✘ ✘
Contact → Collective Action – – ✘ ✘ – – ✘ ✔
Contact → Humanization – – ✘ ✘ – – ✘ ✔
Open in a new tab

Note: Tick marks signify statistically significant results (p < .05); crosses indicate results that do not reach statistical significance.

Particularly novel to the field is our tests regarding the STE. We found evidence of Black‐to‐gay process generalization (i.e., within‐subjects effects), where quantity of contact with Black people predicted: (a) increased gay humanization and; (b) increased gay collective action intentions. That is, more contact with Black people predicted changes within people that lead to more favourable outcomes towards gay people. Interestingly, the opposite pattern of generalization effects occurred at the between‐subjects level, where we found evidence of gay‐to‐Black associative generalization, in that those: (a) with greater contact (quantity or quality) with gay people humanized Black people more, and; (b) with greater contact (quantity or quality) with gay people also reported greater collective action intentions helping Black people. Recall that between‐subjects effects reflect stable associations only, not evidence of an active process captured during the duration of the study.

Of note, it appears that quantity (vs. quality) of contact is a more consistent predictor of secondary contact effects. Although contact valence (vs. frequency) is a stronger predictor of attitudes towards the outgroup in primary contact research (De Coninck et al., 2021; Islam & Hewstone, 1993), in our investigation secondary contact effects were primarily observed in the present study with regard to contact quantity. At the within‐subjects level, frequency of contact with Black people predicted increased collective action intentions towards gay people and also humanized gay people. In contrast, we observed no significant within‐person secondary contact effects for contact quality. Of note, however, is that we did not simultaneously test quantity and quality as predictors of intergroup outcomes to explore their unique predictive effects in light of the model complexity and in light of our preregistered analysis plan. We see value in such an approach, however, and encourage future researchers who are interested in the unique prediction of each type of contact to consider such statistical controls (e.g., the effects of contact quantity controlling for contact quality).

Overall, we found evidence of Black‐to‐gay process generalization (within‐person change) and evidence of gay‐to‐Black associative generalization (between‐person differences). The observed Black‐to‐gay process generalization over time, without the presence of comparable gay‐to‐Black generalization, could reflect the properties of the specific groups we examined. With regard to member‐to‐group generalization, that is, how attitudes towards a member of Group X generalize to Group X as a whole, we know that the process is more easily facilitated when the group member is deemed relatively prototypical of the outgroup (see Brown & Hewstone, 2005). Put another way, member‐to‐group generalization occurs most readily when thinking of the group member in terms of their group (vs. personal) identity. Relatedly, generalization across groups (i.e., secondary transfer effects) might be more easily facilitated when the contact group is represented by the perceiver in more essentialized terms, that is, when their ‘groupness’ is accentuated. Consistent with this possibility, White participants rate Black (vs. gay) targets as higher in natural essentialism (Hodson & Skorska, 2015), meaning that they view Black people as particularly distinct from their outgroups. Greater essentialism might facilitate the spillover effects of STE from Black (vs. gay) contact because the contact is psychologically experienced as relatively more ‘group’ based. Additionally, given the physical salience of interracial relations, as opposed to relations across sexual orientation groups, interactants might be relatively more aware of, and conscious of, the fact that they are engaging in an intergroup contact situation, facilitating generalization effects from Black contact (higher salience) than from gay contact (lower salience). However, because this is a relatively novel finding, it would be beneficial for future research to test the robustness of this pattern and explore why this might be the case.

The present study also demonstrates the importance of studying contact longitudinally. Theorists have argued that studying contact as it unfolds over time can be critical to best understanding its effects (MacInnis & Page‐Gould, 2015), and in the present investigation, doing so has clarified the nature of contact generalization effects. Some researchers in the field have expressed concern that contact effects might reflect between‐subjects differences as opposed to within‐subjects changes (Friehs et al., 2024; Hodson & Meleady, 2024; Sengupta et al., 2023). Our results, however, find clear evidence of within‐person changes as a function of more frequent or greater quality contact, across multiple outcome measures. Part of the difference between our findings and those of past research might involve the specific contact groups in question. Here, where we have found within‐person effects, we examined contact with Black and gay people, groups that have not been examined in past longitudinal research disentangling between‐ from within‐person effects. We also sampled American participants, whereas much of the previous work in this field has used samples from Britain, Europe, or New Zealand. Of note, recent meta‐analytic analyses reveal that contact among children in school exerts significantly stronger effects among American than European samples (Bardach et al., 2024), which might explain why our analyses revealed evidence of within‐person contact effects.

Regardless of these methodological differences, we see good reason for the field to avoid becoming overly pessimistic about the efficacy of intergroup contact, particularly given its clear effects in experimental research as well (e.g., Albuja et al., 2024; Grady et al., 2023). It is also important to keep in mind that longitudinal contact studies capture ‘windows’ of individuals' experiences with other groups where moment‐to‐moment fluctuations in contact predict moment‐to‐moment fluctuations in prejudice (Shulman et al., in press). Importantly, these research methods do not capture changes outside of the window of assessment, including developmental processes if the participants are adults. For example, someone might go on a holiday, meet people from Group X, and start to feel more positive towards Group X (or even Group Y). But when they are assessed 5 years later, their day‐to‐day fluctuations in contact with Group X during that particular study interval fails to predict attitudes towards Group X (or Group Y). In this example, contact clearly led to favourable outcomes towards the outgroup, but this was not captured in the window of assessment in terms of moment‐to‐moment fluctuations during the investigation. Thus, even where researchers find null within‐person effects of contact, it does not necessarily mean that contact (as a construct) does not impact prejudice, but rather that contact and prejudice do not covary in the restricted windows of life examined in those longitudinal studies.

Limitations and future directions

Like all studies, the present study features some limitations. We recruited participants from an online sample, and some researchers have raised concerns about such samples (for review see Johnson, 2015). However, a variety of studies across disciplines have demonstrated that findings from online samples generalize well to broader populations (Coppock, 2018; Merz et al., 2020; Peyton et al., 2021). Moreover, although concerns have been raised about the data quality on online recruitment platforms (Chmielewski & Kucker, 2020), CloudResearch Connect, the platform used to recruit participants in this study, produces higher quality data compared with other platforms (e.g., MTurk, Qualtrics; Douglas et al., 2023). That being said, employment of more generalizable samples (e.g., nationally representative samples) would be a valuable next step for future research.

Another important point of discussion pertains to the specific time‐lags chosen in the present study. Currently, there is no consensus on which time intervals are optimal for examining contact longitudinally. We chose 3‐week intervals between each timepoint to assess contact over a relatively short period of time. Our variables remained quite stable across time (see Figures S13–S22 for graphical depictions of variable stability), potentially indicating that this time‐lag was not long enough to capture significant mean‐level changes in our variables as well. However, other studies finding within‐person changes of contact on intergroup outcomes examine contact over time intervals of 2–4 weeks apart (Boin et al., 2024; Górska & Tausch, 2023). Therefore, although our time intervals were relatively short, our study provides support for a growing body of literature finding that shorter time intervals might be crucial when assessing within‐person effects of intergroup contact, as least for certain social groups (but see Shulman et al., in press).

Additionally, it should be noted that we used the same measures to assess outcomes towards both outgroups in the present study. This method was favoured, in part, because it enabled us to keep the questionnaire short and mitigate participant attrition. But it also allows direct comparisons between the groups (for example, in comparing means and standard deviations), and is consistent with methods used by major polling agencies such as Pew and Gallup. But the method also has potential implications for shared method variance. An alternative could be to use different measures tapping prejudicial attitudes unique to the groups in question (e.g., the Modern Racism Scale or the Attitudes Toward Lesbians and Gay Men Scale), although there would still remain shared method variance given the self‐report nature of these scales. Nonetheless, we see it as beneficial for future research to explore the effects of intergroup contact on outcomes that tap attitudes unique to the social groups examined.

Future research might also consider replicating the present study using a humanization measure that is more commonly used in the contact literature. In the present study, we assessed humanization via a measure adapted by Santoro and Monin (2023), which has been used to assess attributions of humanity to artificial intelligence. However, this construct broadly assesses attributions of uniquely human experiences and capabilities, which is a validated approach to measuring humanization (for review, see Kteily & Landry, 2022). Additionally, research shows that responses on various humanization measures generally correlate with one another (Kteily et al., 2015, Table 3). Thus, although the measure we used is newer, it is built on widely used concepts and theoretically aligns with the ways in which researchers measure humanization. Additionally, we chose this measure based on empirical support from a pilot study we conducted prior to the present study (see Appendix S1), which revealed that the Santoro and Monin (2023) measure correlated with the other humanization measures that we assessed, correlated with prejudice and ideology on a similar magnitude to that of the other humanization measures, and exhibited the best psychometric properties out of the three humanization scales that we assessed. Nevertheless, future contact research could consider utilizing humanization measures more commonly used in the contact literature.

Lastly, missing values analysis revealed that the variables in our sample were missing not at random (MNAR). In general, those who were missing variables had more positive attitudes towards each outgroup, but had less contact with each outgroup, and those who were missing timepoints had more positive attitudes towards each outgroup and humanized each outgroup more (see Appendix S1 for a more detailed discussion of missing data). This potentially indicates that our sample is not entirely representative of the broader population in terms of observed contact effects. Nonetheless, among the participants with data used in the present analyses, contact appeared to exert positive primary and secondary contact effects over time. However, it would be beneficial for future research to explore contacts effects among those with missing versus complete data, and even to use nationally representative samples, to provide a more representative understanding of these effects.

CONCLUSION

Recently, emerging longitudinal contact research has presented evidence that the intergroup contact effects on outcomes, which are theoretically assumed or implied to be within‐subjects in nature, might instead empirically reflect between‐subjects effects or ‘differences’ between individuals (Friehs et al., 2024; Hodson & Meleady, 2024; Sengupta et al., 2023; Shulman et al., in press). For good reason, such findings have raised concerns about intergroup contact as an intervention tool. Promisingly, our results show that both primary and secondary contact with Black and gay outgroups has the potential to change people in ways that lead to positive intergroup outcomes, as posited by contact theory. Promisingly, intergroup contact has the potential to exert broader effects than mere improvement of attitudes towards the contacted groups, but rather can influence how we think about ‘others’ more generally.

AUTHOR CONTRIBUTIONS

Hanna Puffer: Conceptualization; writing – original draft; formal analysis. Gordon Hodson: Conceptualization; funding acquisition; writing – review and editing; supervision.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to disclose.

Supporting information

Appendix S1

BJSO-64-0-s001.pdf (2.3MB, pdf)

ACKNOWLEDGEMENTS

This project was funded by a Council for Research in Social Sciences (CRISS) grant (112‐222‐920) from Brock University awarded to the second author.

Puffer, H. , & Hodson, G. (2025). Primary and secondary generalization effects from Black and gay contact: Longitudinal evidence of between‐ and within‐person effects. British Journal of Social Psychology, 64, e12900. 10.1111/bjso.12900

Footnotes

1

Note that we initially preregistered humanization as a potential mediator of our contact effects on attitudes and collective action. However, upon initial testing we failed to find significant b‐paths (i.e., humanization predicting attitudes/collective action) in relation to secondary contact effects, meaning tests of indirect effects would not be viable. Due to the complexity of our models, and that mediation effects were not observed, in the following analyses we simply report contact effects on humanization (as we do for attitudes and collective action). In the interest of transparency, the full mediation model can be found in the Tables S1–S4, Figures S1–S8.

2

Participants were retained in the analytic sample if they selected “White” as one of their racialized identities (i.e., they had the option to select multiple racialized identities): 100% identified as White, 0.4% as Black, 0.4% as Asian, 0.9% as Indigenous, 0.2% as Middle Eastern, 8.3% as Hispanic/Latino/South American.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in OSF: [https://osf.io/dekj5/].

REFERENCES

  1. Albarello, F. , Crocetti, E. , & Rubini, M. (2020). Prejudice and inclusiveness in adolescence: The role of social dominance orientation and multiple categorization. Child Development, 91(4), 1183–1202. 10.1111/cdev.13295 [DOI] [PubMed] [Google Scholar]
  2. Albuja, A. F. , Gaither, S. E. , Sanchez, D. T. , & Nixon, J. (2024). Testing intergroup contact theory through a natural experiment of randomized college roommate assignments in the United States. Journal of Personality and Social Psychology, 127(2), 277–290. 10.1037/pspa0000393 [DOI] [PubMed] [Google Scholar]
  3. Allport, G. W. (1954). The nature of prejudice. Addison‐Wesley. [Google Scholar]
  4. Bardach, L. , Röhl, S. , Oczlon, S. , Schumacher, A. , Lüftenegger, M. , Lavelle‐Hill, R. , Schwarzenthal, M. , & Zitzmann, S. (2024). Cultural diversity climate in school: A meta‐analytic review of its relationships with intergroup, academic, and socioemotional outcomes. Psychological Bulletin, 150(12), 1397–1439. 10.1037/bul0000454 [DOI] [PubMed] [Google Scholar]
  5. Beelmann, A. , & Heinemann, K. S. (2014). Preventing prejudice and improving intergroup attitudes: A meta‐analysis of child and adolescent training programs. Journal of Applied Developmental Psychology, 35(1), 10–24. 10.1016/j.appdev.2013.11.002 [DOI] [Google Scholar]
  6. Bergh, R. , Akrami, N. , Sidanius, J. , & Sibley, C. G. (2016). Is group membership necessary for understanding generalized prejudice? A re‐evaluation of why prejudices are interrelated. Journal of Personality and Social Psychology, 111(3), 367–395. 10.1037/pspi0000064 [DOI] [PubMed] [Google Scholar]
  7. Bohrer, B. , Friehs, M. , Schmidt, P. , & Weick, S. (2019). Contacts between natives and migrants in Germany: Perceptions of the native population since 1980 and an examination of the contact hypotheses. Social Inclusion, 7(4), 320–331. 10.17645/si.v7i4.2429 [DOI] [Google Scholar]
  8. Boin, J. , Fuochi, G. , & Voci, A. (2020). Deprovincialization as a key correlate of ideology, prejudice, and intergroup contact. Personality and Individual Differences, 157, 109799. 10.1016/j.paid.2019.109799 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Boin, J. , Fuochi, G. , Voci, A. , & Hewstone, M. (2024). The intra‐individual power of contact: Investigating when, how and why intergroup contact and intergroup outcomes fluctuate together. European Journal of Social Psychology, 54, 298–313. 10.1002/ejsp.3016 [DOI] [Google Scholar]
  10. Bolger, N. , & Laurenceau, J. (2013). Intensive longitudinal methods: An introduction to diary and experience sampling research. Guilford Press. [Google Scholar]
  11. Brown, R. , & Hewstone, M. (2005). An integrative theory of intergroup contact. In Zanna M. (Ed.), Advances in experimental social psychology (Vol. 37, pp. 255–343). Academic Press. [Google Scholar]
  12. Capozza, D. , Di Bernardo, G. A. , & Falvo, R. (2017). Intergroup contact and outgroup humanization: Is the causal relationship uni‐ or bidirectional? PLoS One, 12(1), e0170554. 10.1371/journal.pone.0170554 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Capozza, D. , Trifiletti, E. , Vezzali, L. , & Favara, I. (2013). Can intergroup contact improve humanity attributions? International Journal of Psychology, 48(4), 527–541. 10.1080/00207594.2012.688132 [DOI] [PubMed] [Google Scholar]
  14. Chmielewski, M. S. , & Kucker, S. C. (2020). An MTurk crisis? Shifts in data quality and the impact on study results. Social Psychological and Personality Science, 11(4), 464–473. 10.1177/1948550619875149 [DOI] [Google Scholar]
  15. Choma, B. L. , Hodson, G. , Sumantry, D. , Hanoch, Y. , & Gummerum, M. (2021). Ideological and psychological predictors of COVID‐19‐related collective action, opinions, and health compliance across three nations. Journal of Social and Political Psychology, 9(1), 123–143. 10.5964/jspp.5585 [DOI] [Google Scholar]
  16. Coppock, A. (2018). Generalizing from survey experiments conducted on mechanical Turk: A replication approach. Political Science Research and Methods, 7(3), 613–628. 10.1017/psrm.2018.10 [DOI] [Google Scholar]
  17. De Coninck, D. , Igartua, J. J. , & D'Haenens, L. (2021). The contact hypothesis during the European refugee crisis: Relating quality and quantity of (in)direct intergroup contact to attitudes toward refugees. Group Processes & Intergroup Relations, 24(6), 881–901. 10.1177/1368430220929394 [DOI] [Google Scholar]
  18. Dhont, K. , Van Hiel, A. , & Hewstone, M. (2014). Changing the ideological roots of prejudice: Longitudinal effects of ethnic intergroup contact on social dominance orientation. Group Processes & Intergroup Relations, 17(1), 27–44. 10.1177/1368430213497064 [DOI] [Google Scholar]
  19. Dhont, K. , Van Hiel, A. , De Bolle, M. , & Roets, A. (2011). Longitudinal intergroup contact effects on prejudice using self‐ and observer‐reports. British Journal of Social Psychology, 51(2), 221–238. 10.1111/j.2044-8309.2011.02039.x [DOI] [PubMed] [Google Scholar]
  20. Di Bernardo, G. A. , Vezzali, L. , Stathi, S. , McKeown, S. , Cocco, V. M. , Saguy, T. , & Dixon, J. (2019). Fostering social change among advantaged and disadvantaged group members: Integrating intergroup contact and social identity perspectives on collective action. Group Processes & Intergroup Relations, 24(1), 26–47. 10.1177/1368430219889134 [DOI] [Google Scholar]
  21. Dixon, J. , & Levine, M. (Eds.). (2012). Beyond prejudice: Extending the social psychology of conflict, inequality and social change. Cambridge University Press. [Google Scholar]
  22. Douglas, B. D. , Ewell, P. J. , & Bräuer, M. (2023). Data quality in online human‐subjects research: Comparisons between MTurk, prolific, CloudResearch, Qualtrics, and SONA. PLoS One, 18(3), e0279720. 10.1371/journal.pone.0279720 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Dovidio, J. F. , Love, A. , Schellhaas, F. M. H. , & Hewstone, M. (2017). Reducing intergroup bias through intergroup contact: Twenty years of progress and future directions. Group Processes & Intergroup Relations, 20(5), 606–620. 10.1177/1368430217712052 [DOI] [Google Scholar]
  24. Friehs, M. , Bracegirdle, C. , Reimer, N. K. , Wölfer, R. , Schmidt, P. , Wagner, U. , & Hewstone, M. (2024). The between‐person and within‐person effects of intergroup contact on outgroup attitudes: A multi‐context examination. Social Psychological and Personality Science, 194855062311530, 125–141. 10.1177/19485506231153017 [DOI] [Google Scholar]
  25. Fuochi, G. , Voci, A. , Veneziani, C. A. , Boin, J. , Fell, B. , & Hewstone, M. (2020). Is negative mass media news always associated with outgroup prejudice? The buffering role of direct contact. Group Processes & Intergroup Relations, 23(2), 195–213. 10.1177/1368430219837347 [DOI] [Google Scholar]
  26. Górska, P. , & Tausch, N. (2023). Dynamic, yet stable: Separating within‐ and between‐person components of collective action in support of a disadvantaged outgroup and its antecedents. Social Psychological and Personality Science, 14(7), 875–887. 10.1177/19485506221133882 [DOI] [Google Scholar]
  27. Grady, C. , Wolfe, R. , Dawop, D. , & Inks, L. (2023). How contact can promote societal change amid conflict: An intergroup contact field experiment in Nigeria. Proceedings of the National Academy of Sciences of the United States of America, 120(43), e2304882120. 10.1073/pnas.2304882120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hamaker, E. L. , Kuiper, R. M. , & Grasman, R. P. P. P. (2015). A critique of the cross‐lagged panel model. Psychological Methods, 20(1), 102–116. 10.1037/a0038889 [DOI] [PubMed] [Google Scholar]
  29. Hayes, A. F. (2006). A primer on multilevel modeling. Human Communication Research, 32(4), 385–410. 10.1111/j.1468-2958.2006.00281.x [DOI] [Google Scholar]
  30. Hodson, G. , & Hewstone, M. (Eds.). (2013). Advances in intergroup contact. Psychology Press. [Google Scholar]
  31. Hodson, G. , & Meleady, R. (2024). Replicating and extending Sengupta et al. (2023): Contact predicts no within‐person longitudinal outgroup‐bias change. American Psychologist, 79(3), 451–462. 10.1037/amp0001210 [DOI] [PubMed] [Google Scholar]
  32. Hodson, G. , & Puffer, H. (in press). The evolving nature of generalized prejudice toward marginalized groups in the United States 2004–2020. Social Psychological and Personality Science. [Google Scholar]
  33. Hodson, G. , & Skorska, M. N. (2015). Tapping generalized essentialism to predict outgroup prejudices. British Journal of Social Psychology, 54(2), 371–382. 10.1111/bjso.12083 [DOI] [PubMed] [Google Scholar]
  34. Hodson, G. , Harry, H. , & Mitchell, A. (2009). Independent benefits of contact and friendship on attitudes toward homosexuals among authoritarians and highly identified heterosexuals. European Journal of Social Psychology, 39(4), 509–525. 10.1002/ejsp.558 [DOI] [Google Scholar]
  35. Hodson, G. , MacInnis, C. C. , & Busseri, M. A. (2017). Bowing and kicking: Rediscovering the fundamental link between generalized authoritarianism and generalized prejudice. Personality and Individual Differences, 104, 243–251. 10.1016/j.paid.2016.08.018 [DOI] [Google Scholar]
  36. Islam, M. R. , & Hewstone, M. (1993). Dimensions of contact as predictors of intergroup anxiety, perceived out‐group variability, and out‐group attitude: An integrative model. Personality and Social Psychology Bulletin, 19(6), 700–710. 10.1177/0146167293196005 [DOI] [Google Scholar]
  37. Johnson, J. S. (2015). Improving online panel data usage in sales research. Journal of Personal Selling & Sales Management, 36(1), 74–85. 10.1080/08853134.2015.1111611 [DOI] [Google Scholar]
  38. Kteily, N. , & Landry, A. P. (2022). Dehumanization: Trends, insights, and challenges. Trends in Cognitive Sciences, 26(3), 222–240. 10.1016/j.tics.2021.12.003 [DOI] [PubMed] [Google Scholar]
  39. Kteily, N. , Bruneau, E. , Waytz, A. , & Cotterill, S. (2015). The ascent of man: Theoretical and empirical evidence for blatant dehumanization. Journal of Personality and Social Psychology, 109(5), 901–931. 10.1037/pspp0000048 [DOI] [PubMed] [Google Scholar]
  40. Lane, S. P. , & Hennes, E. P. (2018). Power struggles: Estimating sample size for multilevel relationships research. Journal of Social and Personal Relationships, 35(1), 7–31. 10.1177/0265407517710342 [DOI] [Google Scholar]
  41. Lemmer, G. , & Wagner, U. (2015). Can we really reduce ethnic prejudice outside the lab? A meta‐analysis of direct and indirect contact interventions. European Journal of Social Psychology, 45(2), 152–168. 10.1002/ejsp.2079 [DOI] [Google Scholar]
  42. Lissitsa, S. , & Kushnirovich, N. (2020). Is negative the new positive? Secondary transfer effect of exposure to LGBT portrayals in TV entertainment programs. Journal of Applied Social Psychology, 50(2), 115–130. 10.1111/jasp.12644 [DOI] [Google Scholar]
  43. Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83(404), 1198–1202. 10.1080/01621459.1988.10478722 [DOI] [Google Scholar]
  44. Lucas, R. E. (2023). Why the cross‐lagged panel model is almost never the right choice. Advances in Methods and Practices in Psychological Science, 6(1), 1–22. 10.1177/2515245923115 [DOI] [Google Scholar]
  45. MacInnis, C. C. , & Hodson, G. (2019). Extending the benefits of intergroup contact beyond attitudes: When does intergroup contact predict greater collective action support? Journal of Theoretical Social Psychology, 3(1), 11–22. 10.1002/jts5.23 [DOI] [Google Scholar]
  46. MacInnis, C. C. , & Page‐Gould, E. (2015). How can intergroup interaction be bad if intergroup contact is good? Exploring and reconciling an apparent paradox in the science of intergroup relations. Perspectives on Psychological Science, 10(3), 307–327. 10.1177/1745691614568482 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. McKeown, S. , Schaefer, C. D. , Ali, S. , Dupont, P. , Manley, D. , Rao, S. , Taylor, L. K. , & Meleady, R. (in press). The longitudinal relationship between youth intergroup contact and social cohesion outcomes in two divided societies. European Journal of Social Psychology. 10.1002/ejsp.3121 [DOI] [Google Scholar]
  48. Meleady, R. , Crisp, R. J. , Hodson, G. , & Earle, M. (2019). On the generalization of intergroup contact: A taxonomy of transfer effects. Current Directions in Psychological Science, 28(5), 430–435. 10.1177/0963721419848682 [DOI] [Google Scholar]
  49. Merz, Z. C. , Lace, J. W. , & Eisenstein, A. M. (2020). Examining broad intellectual abilities obtained within an MTurk internet sample. Current Psychology, 41(4), 2241–2249. 10.1007/s12144-020-00741-0 [DOI] [Google Scholar]
  50. Miles, E. , & Crisp, R. J. (2014). A meta‐analytic test of the imagined contact hypothesis. Group Processes & Intergroup Relations, 17(1), 3–26. 10.1177/1368430213510573 [DOI] [Google Scholar]
  51. Mulder, J. , & Hamaker, E. L. (2021). Three extensions of the random intercept cross‐lagged panel model. Structural Equation Modeling, 28(4), 638–648. 10.1080/10705511.2020.1784738 [DOI] [Google Scholar]
  52. Page‐Gould, E. , Mendoza‐Denton, R. , & Tropp, L. R. (2008). With a little help from my cross‐group friend: Reducing anxiety in intergroup contexts through cross‐group friendship. Journal of Personality and Social Psychology, 95(5), 1080–1094. 10.1037/0022-3514.95.5.1080 [DOI] [PubMed] [Google Scholar]
  53. Paolini, S. , Harwood, J. , Rubin, M. , Husnu, S. , Joyce, N. , & Hewstone, M. (2014). Positive and extensive intergroup contact in the past buffers against the disproportionate impact of negative contact in the present. European Journal of Social Psychology, 44(6), 548–562. 10.1002/ejsp.2029 [DOI] [Google Scholar]
  54. Petsko, C. D. , & Bodenhausen, G. V. (2019). Racial stereotyping of gay men: Can a minority sexual orientation erase race? Journal of Experimental Social Psychology, 83, 37–54. 10.1016/j.jesp.2019.03.002 [DOI] [Google Scholar]
  55. Pettigrew, T. F. (2009). Secondary transfer effect of contact: Do intergroup contact effects spread to noncontacted outgroups? Social Psychology, 40(2), 55–65. 10.1027/1864-9335.40.2.55 [DOI] [Google Scholar]
  56. Pettigrew, T. F. , & Tropp, L. R. (2006). A meta‐analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90(5), 751–783. [DOI] [PubMed] [Google Scholar]
  57. Peyton, K. , Huber, G. A. , & Coppock, A. (2021). The generalizability of online experiments conducted during the COVID‐19 pandemic. Journal of Experimental Political Science, 9(3), 379–394. 10.1017/xps.2021.17 [DOI] [Google Scholar]
  58. Prati, F. , & Loughnan, S. (2018). Imagined intergroup contact promotes support for human rights through increased humanization. European Journal of Social Psychology, 48(1), 051–061. 10.1002/ejsp.2282 [DOI] [Google Scholar]
  59. Puffer, H. , & Hodson, G. (2024). Exploring secondary transfer generalisation effects from Black and gay contact: The role of humanisation. Journal of Community & Applied Social Psychology, 34(6), 1–14. 10.1002/casp.70009 [DOI] [Google Scholar]
  60. Raudenbush, S. W. , & Congdon, R. T. (2021). HLM 8: Hierarchical linear and nonlinear modeling. Scientific Software International, Inc. [Google Scholar]
  61. Santoro, E. , & Monin, B. (2023). The AI effect: People rate distinctively human attributes as more essential to being human after learning about artificial intelligence advances. Journal of Experimental Social Psychology, 107, 104464. 10.1016/j.jesp.2023.104464 [DOI] [Google Scholar]
  62. Sengupta, N. K. , Reimer, N. K. , Sibley, C. G. , & Barlow, F. K. (2023). Does intergroup contact foster solidarity with the disadvantaged? A longitudinal analysis across 7 years. American Psychologist, 78(6), 750–760. 10.1037/amp0001079 [DOI] [PubMed] [Google Scholar]
  63. Shulman, D. , Meleady, R. , Hodson, G. , & Crisp, R. J. (in press). Fluctuations in prejudice do not track fluctuations in ordinary contact in three 5‐wave “Shortitudinal” studies examining daily, weekly, or monthly intervals. Social Psychological and Personality Science. 10.1177/19485506241284079 [DOI] [Google Scholar]
  64. Swart, H. , Hewstone, M. , Christ, O. , & Voci, A. (2011). Affective mediators of intergroup contact: A three‐wave longitudinal study in South Africa. Journal of Personality and Social Psychology, 101(6), 1221–1238. 10.1037/a0024450 [DOI] [PubMed] [Google Scholar]
  65. Tausch, N. , Hewstone, M. , Kenworthy, J. B. , Psaltis, C. , Schmid, K. , Popan, J. R. , Cairns, E. , & Hughes, J. (2010). Secondary transfer effects of intergroup contact: Alternative accounts and underlying processes. Journal of Personality and Social Psychology, 99(2), 282–302. 10.1037/a0018553 [DOI] [PubMed] [Google Scholar]
  66. Tiv, M. , & Spence, C. (2025). Does ethnoracial context diversity predict implicit associations of sexual orientation? Evidence from linked Census Bureau and implicit association test data. Group Processes & Intergroup Relations, 28(3), 571–599. 10.1177/13684302241290919 [DOI] [Google Scholar]
  67. Tropp, L. R. , & Mallett, R. K. (Eds.). (2011). Moving beyond prejudice reduction: Pathways to positive intergroup relations. American Psychological Association. 10.1037/12319-000 [DOI] [Google Scholar]
  68. Tropp, L. R. , Hawi, D. R. , Van Laar, C. , & Levin, S. (2012). Cross‐ethnic friendships, perceived discrimination, and their effects on ethnic activism over time: A longitudinal investigation of three ethnic minority groups. British Journal of Social Psychology, 51(2), 257–272. 10.1111/j.2044-8309.2011.02050.x [DOI] [PubMed] [Google Scholar]
  69. Van Assche, J. , Swart, H. , Schmid, K. , Dhont, K. , Al Ramiah, A. , Christ, O. , Kauff, M. , Rothmann, S. , Savelkoul, M. , Tausch, N. , Wölfer, R. , Zahreddine, S. , Saleem, M. , & Hewstone, M. (2023). Intergroup contact is reliably associated with reduced prejudice, even in the face of group threat and discrimination. American Psychologist, 78(6), 761–774. 10.1037/amp0001144 [DOI] [PubMed] [Google Scholar]
  70. Vezzali, L. , Di Bernardo, G. A. , Stathi, S. , Cadamuro, A. , Lášticová, B. , & Andraščiková, S. (2018). Secondary transfer effect among children: The role of social dominance orientation and outgroup attitudes. British Journal of Social Psychology, 57(3), 547–566. 10.1111/bjso.12248 [DOI] [PubMed] [Google Scholar]
  71. Vezzali, L. , Giovannini, D. , & Capozza, D. (2010). Longitudinal effects of contact on intergroup relations: The role of majority and minority group membership and intergroup emotions. Journal of Community & Applied Social Psychology, 20(6), 462–479. 10.1002/casp.1058 [DOI] [Google Scholar]
  72. Vezzali, L. , Lolliot, S. , Trifiletti, E. , Cocco, V. M. , Rae, J. R. , Capozza, D. , & Hewstone, M. (2023). Effects of intergroup contact on explicit and implicit outgroup attitudes: A longitudinal field study with majority and minority group members. British Journal of Social Psychology, 62(1), 215–240. 10.1111/bjso.12558 [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Vezzali, L. , Pagliaro, S. , Di Bernardo, G. A. , McKeown, S. , & Cocco, V. M. (2022). Solidarity across group lines: Secondary transfer effect of intergroup contact, perceived moral distance, and collective action. European Journal of Social Psychology, 53(3), 450–470. 10.1002/ejsp.2914 [DOI] [Google Scholar]
  74. Voci, A. , & Hewstone, M. (2003). Intergroup contact and prejudice toward immigrants in Italy: The mediational role of anxiety and the moderational role of group salience. Group Processes & Intergroup Relations, 6(1), 37–54. 10.1177/1368430203006001011 [DOI] [Google Scholar]
  75. Woltman, H. , Feldstain, A. , MacKay, J. C. , & Rocchi, M. (2012). An introduction to hierarchical linear modeling. Tutorial in Quantitative Methods for Psychology, 8(1), 52–69. 10.20982/tqmp.08.1.p052 [DOI] [Google Scholar]
  76. Wu, Y. , Xu, J. , Shen, Y. , Wang, Y. , & Zheng, Y. (2024). Daily agreeableness and acculturation processes in ethnic/racial minority freshmen: The role of inter‐ethnic contact and perceived discrimination. Journal of Personality, 92(5), 1299–1314. 10.1111/jopy.12889 [DOI] [PubMed] [Google Scholar]
  77. Yip, T. , Douglass, S. , & Shelton, J. N. (2013). Daily intragroup contact in diverse settings: Implications for Asian adolescents' ethnic identity. Child Development, 84(4), 1425–1441. 10.1111/cdev.12038 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix S1

BJSO-64-0-s001.pdf (2.3MB, pdf)

Data Availability Statement

The data that support the findings of this study are openly available in OSF: [https://osf.io/dekj5/].

Articles from The British Journal of Social Psychology are provided here courtesy of Wiley

RESOURCES