Survey description
Goals
What we want to replicate from past studies
For these measures, in the “none” condition, we expected that participants would rate the woman higher on voice quality, interest, voice solicitation, and would select her more often.
New for this study
- We are manipulating who is the purported expert. We would expect
that participants would select the man (vs. the woman) more frequently
in the “Man expert” condition. We would also expect for the man to be
seen as having higher voice quality, interest, and solicitation.
- We are also testing how participants evaluate each target: whether they see women (versus men) as more committed, tokenized, or likely to receive backlash.
Procedure
- Participants were asked to imagine that they are a supervising
manager at their firm. They lead a team of employees with a similar
function as them.
- Then, we told that there are increasing demands to improve DEI, and that their firm wants a DEI advisor. We provided more info about this role.
- We listed two candidates for the role: Phillip Dunn and Judith Clark. These are our within-subjects manipulation: the target of evaluation.
- Each candidate had qualifications. We manipulated the qualifications (this was our between subjects manipulation: the target’s expertise)
Number of between-subjects conditions: 3 (target with expertise: Man expert, no expertise, woman expert)
Distribution of qualifications across between-subjects conditions
| No expert condition. | The expert is the man | The expert is the woman | |
|---|---|---|---|
| Judith Clark’s expertise | [randomly assigned functional committee] | [randomly assigned functional committee] | Equity and Belonging in ${e://Field/function} committee |
| Phillip Dunn’s expertise | [randomly assigned functional committee] | Equity and Belonging in ${e://Field/function} committee | [randomly assigned functional committee] |
Number of within-subjects conditions: 2 (target identity: man or woman)
Target: Phillip Dunn or Judith Clark
Measures (after participants read the vignette above)
Replication items
For these measures, in the “none” condition, we expected that participants would rate the woman higher on voice quality, interest, voice solicitation, and would select her more often.
Voice solicitation
- I would ask for help/advice from them on this initiative.
- I would encourage them to speak out on this initiative.
Voice quality
- I think they can offer useful ideas for this initiative.
- I think their ideas will likely have a lot of value for improving this initiative.
Interest
- How interested do you think each candidate would be in offering their perspectives or insights on this role?
Who would you select for this role?
- Phillip Dunn (coded as 0 if selected), Judith Clark (Coded as 1 if selected)
New items: Participant’s perceptions of each candidate
Participants see this for EACH candidate. The text for Judith Clark is below.
Prompt read: “Imagine that you asked Judith Clark to provide her perspective or insights about this role.”
Tokenization
- feel like her identity as a woman would be made salient.
- feel like she would have to represent other women.
- feel singled out.
Commitment
- be willing to put in a great deal of effort beyond that normally
expected in order to help this initiative be successful.
- want to talk up this organization to others as a great organization
to work for.
- wonder if her values and the organization’s are similar..
- be extremely glad that she chose this organization to work for.
Backlash
If I choose Judith Clark to provide her perspective or insights on this role, she would probably think that…
- it’s risky to challenge existing processes because it may be seen as
questioning the status quo.
- speaking up to suggest a better way is likely to offend
people.
- it is not good to question the way things are done because people are likely to take it personally.
Analyses
Selection likelihood
Note: 1 = selected woman, 0 = selected man
Comparing likelihood of selecting men (versus women) when there is no expert
When there is not a DEI expert, participants were significantly more likely to select the woman.
Chi-squared comparison
##
## Chi-squared test for given probabilities
##
## data: table(exp_clean1$select)
## X-squared = 14, df = 1, p-value = 0.0002Comparing selections of women across all three conditions
## select
## condition_text2 0 1
## 1. Man expert 58 22
## 2. No expert 20 54
## 3. Woman expert 8 66Reference: Man
##
## Call:
## glm(formula = select ~ condition_text2, family = "binomial",
## data = exp_clean1)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.969 0.250 -3.87 0.00011 ***
## condition_text22. No expert 1.963 0.362 5.42 0.000000060211 ***
## condition_text23. Woman expert 3.080 0.450 6.84 0.000000000008 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 302.18 on 227 degrees of freedom
## Residual deviance: 231.17 on 225 degrees of freedom
## AIC: 237.2
##
## Number of Fisher Scoring iterations: 4Reference: None
##
## Call:
## glm(formula = select ~ condition_text, family = "binomial", data = exp_clean1)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.993 0.262 3.79 0.00015 ***
## condition_text2_ManExpert -1.963 0.362 -5.42 0.00000006 ***
## condition_text3_WomanExpert 1.117 0.457 2.45 0.01448 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 302.18 on 227 degrees of freedom
## Residual deviance: 231.17 on 225 degrees of freedom
## AIC: 237.2
##
## Number of Fisher Scoring iterations: 4Do participants select the expert more when the expert is a woman (versus man)
expertisecomp <- exp_clean1 %>%
mutate(expertselection = case_when(
condition == "man" & select == 0 ~ 1,
condition == "man" & select == 1 ~ 0,
condition == "woman" & select == 0 ~ 0,
condition == "woman" & select == 1 ~ 1
)) %>%
filter(condition != "none")
summary(glm(expertselection~condition, expertisecomp, family = "binomial"))##
## Call:
## glm(formula = expertselection ~ condition, family = "binomial",
## data = expertisecomp)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.969 0.250 3.87 0.00011 ***
## conditionwoman 1.141 0.450 2.53 0.01131 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 151.88 on 153 degrees of freedom
## Residual deviance: 144.80 on 152 degrees of freedom
## AIC: 148.8
##
## Number of Fisher Scoring iterations: 4left_join(
exp(coef(glm(expertselection~condition, expertisecomp, family = "binomial"))) %>% as.data.frame()%>% rownames_to_column("variable") %>% rename(oddsratio = "."),
exp(confint(glm(expertselection~condition, expertisecomp, family = "binomial"))) %>% as.data.frame() %>% rownames_to_column("variable"))## Waiting for profiling to be done...## Joining with `by = join_by(variable)`VQ
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
VS
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
Interest
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
FSR
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
Token
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
Commitment
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
Backlash
Means
Anova
Compares evaluations of man vs. woman WITHIN subjects
Compares gender evaluations across BTWN subjects manipulations
Comparing men in the “men expert” condition to women in the other two conditions
Mediation
Mediator: VQ
DV: VS
##
## ****************** PART 1. Regression Model Summary ******************
##
## PROCESS Model Code : 7 (Hayes, 2018; www.guilford.com/p/hayes3)
## PROCESS Model Type : Moderated Mediation
## - Outcome (Y) : vs
## - Predictor (X) : person_num
## - Mediators (M) : vq
## - Moderators (W) : condition
## - Covariates (C) : -
## - HLM Clusters : pid
##
## All numeric predictors have been grand-mean centered.
## (For details, please see the help page of PROCESS.)
##
## Formula of Mediator:
## - vq ~ person_num*condition + (1 | pid)
## Formula of Outcome:
## - vs ~ person_num + condition + vq + (1 | pid)
##
## CAUTION:
## Fixed effect (coef.) of a predictor involved in an interaction
## denotes its "simple effect/slope" at the other predictor = 0.
## Only when all predictors in an interaction are mean-centered
## can the fixed effect denote the "main effect"!
## ##
## Model Summary
##
## ───────────────────────────────────────────────────────────────────
## (1) vs (2) vq (3) vs
## ───────────────────────────────────────────────────────────────────
## (Intercept) 5.183 *** 5.400 *** 5.163 ***
## (0.076) (0.113) (0.069)
## person_num 0.577 *** -0.937 *** 0.025
## (0.153) (0.216) (0.073)
## conditionnone -0.268 0.071
## (0.164) (0.099)
## conditionwoman -0.353 * -0.007
## (0.164) (0.100)
## person_num:conditionnone 1.512 ***
## (0.311)
## person_num:conditionwoman 3.235 ***
## (0.311)
## vq 0.915 ***
## (0.025)
## ───────────────────────────────────────────────────────────────────
## AIC 1752.316 1620.835 1138.556
## BIC 1768.806 1653.815 1167.413
## Num. obs. 456 456 456
## Num. groups: pid 228 228 228
## Var: pid (Intercept) 0.000 0.098 0.084
## Var: Residual 2.665 1.860 0.590
## ───────────────────────────────────────────────────────────────────
## Note. * p < .05, ** p < .01, *** p < .001.
##
## ************ PART 2. Mediation/Moderation Effect Estimate ************
##
## Package Use : ‘mediation’ (v4.5.0), ‘interactions’ (v1.2.0)
## Effect Type : Moderated Mediation (Model 7)
## Sample Size : 456
## Random Seed : set.seed()
## Simulations : 10000 (Bootstrap)
##
## Direct Effect: "person_num" (X) ==> "vs" (Y)## ────────────────────────────────────────────────────────────
## Effect S.E. t p [95% CI]
## ────────────────────────────────────────────────────────────
## Direct (c') 0.025 (0.073) 0.344 .731 [-0.119, 0.169]
## ────────────────────────────────────────────────────────────
##
## Interaction Effect on "vq" (M)
## ───────────────────────────────────────────────
## F df1 df2 p
## ───────────────────────────────────────────────
## person_num * condition 54.08 2 225 <.001 ***
## ───────────────────────────────────────────────
##
## Simple Slopes: "person_num" (X) ==> "vq" (M)
## (Conditional Effects [a] of X on M)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. t p [95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.938 (0.216) -4.348 <.001 *** [-1.360, -0.515]
## none 0.574 (0.224) 2.562 .011 * [ 0.135, 1.014]
## woman 2.297 (0.224) 10.246 <.001 *** [ 1.858, 2.737]
## ─────────────────────────────────────────────────────────────
##
## Running 10000 * 3 simulations...
## Indirect Path: "person_num" (X) ==> "vq" (M) ==> "vs" (Y)
## (Conditional Indirect Effects [ab] of X through M on Y)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. z p [MCMC 95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.858 (0.199) -4.321 <.001 *** [-1.254, -0.475]
## none 0.523 (0.206) 2.543 .011 * [ 0.123, 0.929]
## woman 2.099 (0.212) 9.918 <.001 *** [ 1.692, 2.520]
## ─────────────────────────────────────────────────────────────
## Monte Carlo (Quasi-Bayesian) Confidence Interval
## (Effect, SE, and CI are estimated based on 10000 Monte Carlo samples.)
##
## Note. The results based on bootstrapping or other random processes
## are unlikely identical to other statistical software (e.g., SPSS).
## To make results reproducible, you need to set a seed (any number).
## Please see the help page for details: help(PROCESS)
## Ignore this note if you have already set a seed. :)DV: Select
Mediator: Interest
DV: VS
##
## ****************** PART 1. Regression Model Summary ******************
##
## PROCESS Model Code : 7 (Hayes, 2018; www.guilford.com/p/hayes3)
## PROCESS Model Type : Moderated Mediation
## - Outcome (Y) : vs
## - Predictor (X) : person_num
## - Mediators (M) : interest
## - Moderators (W) : condition
## - Covariates (C) : -
## - HLM Clusters : pid
##
## All numeric predictors have been grand-mean centered.
## (For details, please see the help page of PROCESS.)
##
## Formula of Mediator:
## - interest ~ person_num*condition + (1 | pid)
## Formula of Outcome:
## - vs ~ person_num + condition + interest + (1 | pid)
##
## CAUTION:
## Fixed effect (coef.) of a predictor involved in an interaction
## denotes its "simple effect/slope" at the other predictor = 0.
## Only when all predictors in an interaction are mean-centered
## can the fixed effect denote the "main effect"!
## ##
## Model Summary
##
## ───────────────────────────────────────────────────────────────────
## (1) vs (2) interest (3) vs
## ───────────────────────────────────────────────────────────────────
## (Intercept) 5.183 *** 5.475 *** 5.187 ***
## (0.076) (0.108) (0.091)
## person_num 0.577 *** -0.850 *** 0.029
## (0.153) (0.216) (0.110)
## conditionnone -0.394 * 0.129
## (0.156) (0.131)
## conditionwoman -0.245 -0.141
## (0.156) (0.131)
## person_num:conditionnone 1.688 ***
## (0.312)
## person_num:conditionwoman 3.120 ***
## (0.312)
## interest 0.771 ***
## (0.036)
## ───────────────────────────────────────────────────────────────────
## AIC 1752.316 1599.648 1442.966
## BIC 1768.806 1632.628 1471.823
## Num. obs. 456 456 456
## Num. groups: pid 228 228 228
## Var: pid (Intercept) 0.000 0.000 0.000
## Var: Residual 2.665 1.866 1.312
## ───────────────────────────────────────────────────────────────────
## Note. * p < .05, ** p < .01, *** p < .001.
##
## ************ PART 2. Mediation/Moderation Effect Estimate ************
##
## Package Use : ‘mediation’ (v4.5.0), ‘interactions’ (v1.2.0)
## Effect Type : Moderated Mediation (Model 7)
## Sample Size : 456
## Random Seed : set.seed()
## Simulations : 10000 (Bootstrap)
##
## Direct Effect: "person_num" (X) ==> "vs" (Y)## ────────────────────────────────────────────────────────────
## Effect S.E. t p [95% CI]
## ────────────────────────────────────────────────────────────
## Direct (c') 0.029 (0.110) 0.266 .791 [-0.186, 0.245]
## ────────────────────────────────────────────────────────────## Interaction Effect on "interest" (M)
## ───────────────────────────────────────────────
## F df1 df2 p
## ───────────────────────────────────────────────
## person_num * condition 50.39 2 450 <.001 ***
## ───────────────────────────────────────────────
##
## Simple Slopes: "person_num" (X) ==> "interest" (M)
## (Conditional Effects [a] of X on M)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. t p [95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.850 (0.216) -3.936 <.001 *** [-1.273, -0.427]
## none 0.838 (0.225) 3.731 <.001 *** [ 0.398, 1.278]
## woman 2.270 (0.225) 10.110 <.001 *** [ 1.830, 2.710]
## ─────────────────────────────────────────────────────────────
##
## Running 10000 * 3 simulations...
## Indirect Path: "person_num" (X) ==> "interest" (M) ==> "vs" (Y)
## (Conditional Indirect Effects [ab] of X through M on Y)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. z p [MCMC 95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.655 (0.171) -3.836 <.001 *** [-0.997, -0.322]
## none 0.648 (0.174) 3.718 <.001 *** [ 0.310, 0.995]
## woman 1.751 (0.192) 9.122 <.001 *** [ 1.390, 2.133]
## ─────────────────────────────────────────────────────────────
## Monte Carlo (Quasi-Bayesian) Confidence Interval
## (Effect, SE, and CI are estimated based on 10000 Monte Carlo samples.)
##
## Note. The results based on bootstrapping or other random processes
## are unlikely identical to other statistical software (e.g., SPSS).
## To make results reproducible, you need to set a seed (any number).
## Please see the help page for details: help(PROCESS)
## Ignore this note if you have already set a seed. :)DV: Select
Mediators: VQ+Interest
DV: VS
##
## ****************** PART 1. Regression Model Summary ******************
##
## PROCESS Model Code : 7 (Hayes, 2018; www.guilford.com/p/hayes3)
## PROCESS Model Type : Parallel Multiple Moderated Mediation (2 meds)
## - Outcome (Y) : vs
## - Predictor (X) : person_num
## - Mediators (M) : interest, vq
## - Moderators (W) : condition
## - Covariates (C) : -
## - HLM Clusters : pid
##
## All numeric predictors have been grand-mean centered.
## (For details, please see the help page of PROCESS.)
##
## Formula of Mediator:
## - interest ~ person_num*condition + (1 | pid)
## - vq ~ person_num*condition + (1 | pid)
## Formula of Outcome:
## - vs ~ person_num + condition + interest + vq + (1 | pid)
##
## CAUTION:
## Fixed effect (coef.) of a predictor involved in an interaction
## denotes its "simple effect/slope" at the other predictor = 0.
## Only when all predictors in an interaction are mean-centered
## can the fixed effect denote the "main effect"!
## ##
## Model Summary
##
## ─────────────────────────────────────────────────────────────────────────────────
## (1) vs (2) interest (3) vq (4) vs
## ─────────────────────────────────────────────────────────────────────────────────
## (Intercept) 5.183 *** 5.475 *** 5.400 *** 5.154 ***
## (0.076) (0.108) (0.113) (0.068)
## person_num 0.577 *** -0.850 *** -0.937 *** -0.013
## (0.153) (0.216) (0.216) (0.073)
## conditionnone -0.394 * -0.268 0.099
## (0.156) (0.164) (0.098)
## conditionwoman -0.245 -0.353 * -0.010
## (0.156) (0.164) (0.098)
## person_num:conditionnone 1.688 *** 1.512 ***
## (0.312) (0.311)
## person_num:conditionwoman 3.120 *** 3.235 ***
## (0.312) (0.311)
## interest 0.145 ***
## (0.038)
## vq 0.806 ***
## (0.037)
## ─────────────────────────────────────────────────────────────────────────────────
## AIC 1752.316 1599.648 1620.835 1130.799
## BIC 1768.806 1632.628 1653.815 1163.779
## Num. obs. 456 456 456 456
## Num. groups: pid 228 228 228 228
## Var: pid (Intercept) 0.000 0.000 0.098 0.075
## Var: Residual 2.665 1.866 1.860 0.578
## ─────────────────────────────────────────────────────────────────────────────────
## Note. * p < .05, ** p < .01, *** p < .001.
##
## ************ PART 2. Mediation/Moderation Effect Estimate ************
##
## Package Use : ‘mediation’ (v4.5.0), ‘interactions’ (v1.2.0)
## Effect Type : Parallel Multiple Moderated Mediation (2 meds) (Model 7)
## Sample Size : 456
## Random Seed : set.seed()
## Simulations : 10000 (Bootstrap)
##
## Direct Effect: "person_num" (X) ==> "vs" (Y)## ─────────────────────────────────────────────────────────────
## Effect S.E. t p [95% CI]
## ─────────────────────────────────────────────────────────────
## Direct (c') -0.013 (0.073) -0.174 .862 [-0.156, 0.131]
## ─────────────────────────────────────────────────────────────## Interaction Effect on "interest" (M)
## ───────────────────────────────────────────────
## F df1 df2 p
## ───────────────────────────────────────────────
## person_num * condition 50.39 2 450 <.001 ***
## ───────────────────────────────────────────────
##
## Simple Slopes: "person_num" (X) ==> "interest" (M)
## (Conditional Effects [a] of X on M)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. t p [95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.850 (0.216) -3.936 <.001 *** [-1.273, -0.427]
## none 0.838 (0.225) 3.731 <.001 *** [ 0.398, 1.278]
## woman 2.270 (0.225) 10.110 <.001 *** [ 1.830, 2.710]
## ─────────────────────────────────────────────────────────────
##
## Running 10000 * 3 simulations...
## Indirect Path: "person_num" (X) ==> "interest" (M) ==> "vs" (Y)
## (Conditional Indirect Effects [ab] of X through M on Y)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. z p [MCMC 95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.124 (0.045) -2.718 .007 ** [-0.222, -0.046]
## none 0.121 (0.046) 2.625 .009 ** [ 0.044, 0.223]
## woman 0.330 (0.092) 3.584 <.001 *** [ 0.156, 0.517]
## ─────────────────────────────────────────────────────────────
## Monte Carlo (Quasi-Bayesian) Confidence Interval
## (Effect, SE, and CI are estimated based on 10000 Monte Carlo samples.)
##
## Interaction Effect on "vq" (M)
## ───────────────────────────────────────────────
## F df1 df2 p
## ───────────────────────────────────────────────
## person_num * condition 54.08 2 225 <.001 ***
## ───────────────────────────────────────────────
##
## Simple Slopes: "person_num" (X) ==> "vq" (M)
## (Conditional Effects [a] of X on M)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. t p [95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.938 (0.216) -4.348 <.001 *** [-1.360, -0.515]
## none 0.574 (0.224) 2.562 .011 * [ 0.135, 1.014]
## woman 2.297 (0.224) 10.246 <.001 *** [ 1.858, 2.737]
## ─────────────────────────────────────────────────────────────
##
## Running 10000 * 3 simulations...
## Indirect Path: "person_num" (X) ==> "vq" (M) ==> "vs" (Y)
## (Conditional Indirect Effects [ab] of X through M on Y)
## ─────────────────────────────────────────────────────────────
## "condition" Effect S.E. z p [MCMC 95% CI]
## ─────────────────────────────────────────────────────────────
## man -0.757 (0.177) -4.268 <.001 *** [-1.113, -0.411]
## none 0.459 (0.182) 2.522 .012 * [ 0.108, 0.821]
## woman 1.855 (0.203) 9.146 <.001 *** [ 1.472, 2.270]
## ─────────────────────────────────────────────────────────────
## Monte Carlo (Quasi-Bayesian) Confidence Interval
## (Effect, SE, and CI are estimated based on 10000 Monte Carlo samples.)
##
## Note. The results based on bootstrapping or other random processes
## are unlikely identical to other statistical software (e.g., SPSS).
## To make results reproducible, you need to set a seed (any number).
## Please see the help page for details: help(PROCESS)
## Ignore this note if you have already set a seed. :)