Defining and Categorizing Transgender People | Stargel and Bell Ratings between 0 degrees and 50 degrees mean that you would not consider this person to be transgender. Ratings between 50 degrees and 100 degrees mean that you would consider this person to be transgender. You would rate the person at the 50 degree mark if you do not know if you would consider the person transgender or not transgender. consistent with past research (α = .98 in Billard, 2018). This approach to measuring categorization is simi lar to past research that asked participants to categorize biracial targets on scales that ranged from “completely White” to “completely Black,” with scale midpoints representing a categorization as biracial or multiracial (Ho et al., 2013, 2015). The current work was not con cerned with accuracy of categorization, but rather the extent to which people consider targets to be more or less transgender, and the task was developed with this in mind. Categorization scores were created by a mean composite (M = 50.5, SD = 17.3), and items across the measure demonstrated high reliability (α = .91). Transgender Prejudice A mean composite of the 12-item Attitudes Toward Transgender Men and Women Scale (Billard, 2018) evaluated the extent of participants’ prejudice (e.g., “transgender men will never really be men” and “trans gender women are unnatural.”). Ratings were made on a 1 (strongly disagree) to 7 (strongly agree) scale; higher scores reflect more prejudice (M = 4.1, SD = 1.8). The measure demonstrated high reliability (α = .99) that is TABLE 2 Pearson Correlations Between Variables (N = 497) Variable M(SD) 1 1. Transnormative definition agreement 5.5 (1.0) -- 2 3 2. Umbrella definition agreement 5.8 (1.3) .14** -- 3. Transgender prejudice 4.1 (1.8) –.02 –.25*** -- 4. Gender essentialism 4.2 (1.3) –.01 –.14 .63*** -- 5. Target categorization 51.2 (17.4) .10* .45*** –.09 –.03 -- R2 p ** 4 5 Note. p< .05 p < .01. p < .001. * ** *** TABLE 3 R and Significance of Each Model 2 Outcome Variable Predictor Moderator Umbrella definition agreement (UDA) Transgender prejudice Transgender prejudice Transgender prejudice Gender essentialism Gender essentialism Gender essentialism Transnormative definition agreement (TDA) Target categorization Mediators -- .09 < .001 -- .00 TDA & UDA .98 .20 < .001 Gender Essentialism Brescoll and colleague’s (2013) gender essentialism scale consists of four statements (e.g., “there is something ‘essen tial’ that causes men and women to behave differently” and “men get in more physical fights than women because men have more of an inherent tendency for aggression.”). All ratings were made on a scale ranging from 1 (strongly disagree) to 7 (strongly agree), and higher scores reflect higher beliefs of gender essentialism (M = 4.2, SD = 1.3). The measure demonstrated high reliability (α = .79) compared to past work (α = .63 in Brescoll et al., 2013). Results Correlations Pearson correlations for all measured variables are available in Table 2. There was a significant predicted positive correlation between transgender prejudice and gender essentialism r(497) = .63, p < .001. As predicted, umbrella definition agreement was positively correlated with average categorization score, r(497) = .45, p < .001. Unexpectedly, agreement with transnormative definitions was also positively correlated with average categorization score, r(497) = .10, p = .02. Agreement with umbrella definitions was also negatively correlated with transgender prejudice, r(497) = –.25, p < .001, and gender essentialism, r(497) = –.14, p < .01. Finally, there was a positive correlation between transnorma tive definition agreement and umbrella definition agreement, r(497) = .14, p < .01. There were no other significant relationships between variables. The subse quent analyses aim to elucidate the directional nature of the correlations. Parallel Moderated Mediation Analysis We tested whether transnormative and umbrella definition agreement simultaneously mediated the relationship between transgender prejudice and target categorization, and whether gender essential ism moderates the relationship between transgender prejudice and each definition agreement. The parallel moderated-mediation analysis was conducted through the PROCESS macro in SPSS (Model 7; Hayes, 2013; Preacher & Hayes, 2004). As seen in Figure 1, results indicated that transgender prejudice is indirectly related to transgender categorization through its relationship with umbrella definition agreement (as moderated by gender essentialism) but not through transnormative definition agreement. See Table 3 for R2 values that highlight the proportion of variance in categorization scores explained by each model. 280 COPYRIGHT 2022 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 27, NO. 4/ISSN 2325-7342)