Data Analysis
Julie
8/15/2019
Install libraries, upload data
library(tidyverse)## ── Attaching packages ───────────────## ✔ ggplot2 3.1.1 ✔ purrr 0.3.0
## ✔ tibble 2.1.3 ✔ dplyr 0.8.1
## ✔ tidyr 0.8.2 ✔ stringr 1.4.0
## ✔ readr 1.3.1 ✔ forcats 0.3.0## ── Conflicts ────────────────────────
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()library(umx) #For Cronbach Alpha calculation## Loading required package: OpenMx## To take full advantage of multiple cores, use:
## mxOption(NULL, 'Number of Threads', parallel::detectCores()) #now
## Sys.setenv(OMP_NUM_THREADS=parallel::detectCores()) #before library(OpenMx)## For an overview type '?umx'##
## Attaching package: 'umx'## The following object is masked from 'package:stats':
##
## loadingslibrary(lmerTest) #For analyses using lmer## Loading required package: lme4## Loading required package: Matrix##
## Attaching package: 'Matrix'## The following objects are masked from 'package:OpenMx':
##
## %&%, expm## The following object is masked from 'package:tidyr':
##
## expand##
## Attaching package: 'lmerTest'## The following object is masked from 'package:lme4':
##
## lmer## The following object is masked from 'package:stats':
##
## steplibrary(emmeans) #For posthoc analyses
library(scales) #For rescaling y axis for AVI demographics plots##
## Attaching package: 'scales'## The following object is masked from 'package:purrr':
##
## discard## The following object is masked from 'package:readr':
##
## col_factor#Load data so far (update as you collect more data)
Part1 <- read.csv("~/Dropbox/Lab Projects/Reputation Study/Data/Payment/2019.08.09 - Part 1 Data.csv")
Part2 <- read.csv("~/Dropbox/Lab Projects/Reputation Study/Data/Check Sample Size/2019.08.05 - Data So Far.csv")#Set theme
theme_set(
theme_classic() + # set the theme
theme(text = element_text(size = 16))
)Prepare both datasets so we can merge Part 1 and Part 2 based on pID/WelcomeCdoe.
#Before joining datasets, remove white spaces in case the extra space prevents participants from matching
Part2$WelcomeCode <- gsub(" ", "", Part2$WelcomeCode) #Remove white spaces in WelcomeCode column (some participants added a space after the code which made it not match with the Part 1 pID)
Part1$email <- gsub(" ", "", Part1$email) #Remove white spaces in Part 1 email column just in case
Part2$Email <- gsub(" ", "", Part2$Email) #Remove white spaces in Part 2 email column just in case
#Note: Participants with pID "k3m2k7twob" and "fllt58q3s15" did not include email but this pID matched with no one's in Part 2, so they must not have completed Part 2
Part1 <- Part1[!(Part1$pID=="k3m2k7twob" | Part1$pID=="fllt58q3s15"),]
#Note: Participant with pID "ypv8o66vtx" did not include email either but matched with email "joshlevine12@comcast.net" in Part 2, so we'll include his email into Part 1 data as well:
Part1$email <- as.character(Part1$email) #Convert emails to character in order to perform operation below
Part1$email[Part1$pID == "ypv8o66vtx"] <- "joshlevine12@comcast.net"
#Note: Participant with pID "t3swzuuwhbp" entered that same code as his email in Part 1 so I must replace Part 1 email with the actual email he provided in Part 2:
Part1$email[Part1$email == "t3swzuuwhbp"] <- Part2$Email[Part2$WelcomeCode == "t3swzuuwhbp"]
#Someone wrote "I do not have one. My regular email is Jackcat00@aol.com" so I will replace that with just his email so that we can match emails in Part 2. His pID was "6o17usgpyei6".
Part1$email[Part1$pID == "6o17usgpyei6"] <- "Jackcat00@aol.com"
#Extract two separate pID for each duplicate, and take out second copy
#Part 1
Duplicate1 <- Part1$pID[Part1$email == "mstrang@wisc.edu"][2]
Duplicate2 <- Part1$pID[Part1$email == "a.swartz@ufl.edu"][2]
Duplicate3 <- Part1$pID[Part1$email == "jrtravagline@gmail.com"][2]
Duplicate4 <- Part1$pID[Part1$email == "houcka@oregonstate.edu"][2]
Part1 <- Part1[!(Part1$pID == Duplicate1 | Part1$pID == Duplicate2 | Part1$pID == Duplicate3 | Part1$pID == Duplicate4) ,] #Take them out
#Part 1
Duplicate1 <- Part2$WelcomeCode[Part2$Email == "houcka@oregonstate.edu"][2]
Part2 <- Part2[!(Part2$WelcomeCode == Duplicate1) ,] #Take them outNow, join participants’ Part 1 and Part 2 data, and call this joined dataframe “Data”
#Turn pID and WelcomeCode into characters
Part1$pID <- as.character(Part1$pID)
Part2$WelcomeCode <- as.character(Part2$WelcomeCode)
#Leftjoin
Data <- left_join(Part1, Part2, by = c("pID" = "WelcomeCode")) #Combine Part1 & Part2 by pID/WelcomeCodeRemove participants who are over 30 years old
Data$Age <- as.numeric(as.character(Data$Age)) #Make Age numeric
Data <- Data %>%
filter(Age <= 30)Set factor levels
#Gender
Data$Gender <- factor(Data$Gender,
levels = c(1,2),
labels = c("Male", "Female"))
#Education
Data$Education <- factor(Data$Education,
levels = c(1,2,3,4,5,6,7,8,9),
labels = c("Less than high school", "High School/GED", "Some college (not currently enrolled)", "Some college (currently enrolled)", "Associates degree", "BA/BS degree", "Master's Degree", "Doctoral Degree", "Professional Degree"))
#SES
Data$FamSES2 <- factor(Data$FamSES2,
levels = c(1,2,3,4,5,6,7,8),
labels = c("<10K", "10-20K", "20-30K", "30-40K", "40-50K", "50-75K","75-100K",">100K"))
#Set order for graphs
Data$reputation <- factor(Data$reputation, levels = c("low", "mod", "high"))
Data$expression <- factor(Data$expression, levels = c("neut", "calm", "exci"))Test for outliers based on the standard discussed in lab meeting (i.e., exclude anything outside of boxplot (for the DV, in this case amount given, or “response”)).
boxplot(Data$response) #No outliers to remove
Create filtering variable so I can easily exclude and include participants who had no variation in the Trust Game (gave a fixed amount the whole time):
Data <- Data %>%
group_by(pID) %>% #Do below by participant
mutate(responseSD = sd(response, na.rm = T)) %>% #First calculate each participants' SD in amount given
mutate(responseNoVar = ifelse(responseSD==0, 1, 0)) #If SD for all trials was 0, make the responseNoVar equal to 1. If not 0, then make it equal to 0
table(Data$responseNoVar)/72 #Count how many participants had no variation in trust game (under the heading "1")##
## 0 1
## 81 6#See what these no variation amounts were
Data %>%
filter(responseNoVar==1 & trial_num == 1) %>% #Filter for people who gave same amount and just select one trial out of 72 since they're all the same
unique() %>%
summarize(response)## # A tibble: 6 x 2
## pID response
## <chr> <int>
## 1 0m6kq8tbb8mp 10
## 2 1sv9s4ko8o9i 5
## 3 2mivlx9sx17 10
## 4 9tpuiq3jo39 10
## 5 m0ce8ox9id 3
## 6 ql9isp4x0hl 10Find out who has done a study like this before and what they thought the purpose of the study was
# If participants answered "Yes" to "Have you ever done a study similar to this one?" see what they described this similar study as:
DidBefore <- Data %>%
group_by(pID, DidBefore, DidBefore_text) %>%
filter(DidBefore==1) %>%
dplyr::select(DidBefore_text) %>%
distinct()## Adding missing grouping variables: `pID`, `DidBefore`#See what people wrote when asked, "What do you think was the purpose of this study?"
Purpose <- Data %>%
filter(!is.na(Purpose_text)) %>%
dplyr::select(Purpose_text) %>%
distinct()## Adding missing grouping variables: `pID`#See what people wrote when asked, "Did you encounter any technical problems?"
TechProb <- Data %>%
filter(!is.na(TechProb_text)) %>%
dplyr::select(TechProb_text) %>%
distinct()## Adding missing grouping variables: `pID`Compute ideal and actual affect scores from AVI
# i1_1 enthusiastic
# i1_7 excited
# i1_9 elated
# i2_9 euphoric
# i1_6 relaxed
# i3_3 calm
# i3_9 peaceful
# i3_10 serene
# Compute actual and ideal affect scores
Data <- Data %>%
mutate_at(vars(a1_1:i4_9), list(~ as.numeric(as.character(.)))) %>% #Turn affect variables from factor to numeric
group_by(pID) %>%
rowwise() %>%
mutate(aHAP = mean(c(a1_1, a1_7, a1_9, a2_9), na.rm = T),
aLAP = mean(c(a1_6, a3_3, a3_9, a3_10), na.rm = T),
iHAP = mean(c(i1_1, i1_7, i1_9, i2_9), na.rm = T),
iLAP = mean(c(i1_6, i3_3, i3_9, i3_10), na.rm = T))
#Ipsatize actual and ideal (separately)
Ipsatized_Actual <- Data %>%
dplyr::select(pID, a1_1:a4_9) %>% #Only select actual affect variables
ungroup() %>%
mutate(aSD = apply(.[,2:ncol(.)], na.rm = T, 1, sd)) %>% #Compute rowwise SD
mutate(aMean = rowMeans(dplyr::select(., contains("_")), na.rm = T)) %>% #Compute rowwise means
mutate(a_enthus_i = (a1_1 - aMean)/aSD, #Ipsatize enthusiastic
a_excited_i = (a1_7 - aMean)/aSD, #Ipsatize excited
a_elated_i = (a1_9 - aMean)/aSD, #Ipsatize elated
a_euphoric_i = (a2_9 - aMean)/aSD, #Ipsatize euphoric
a_relaxed_i = (a1_6 - aMean)/aSD, #Ipsatize relaxed
a_calm_i = (a3_3 - aMean)/aSD, #Ipsatize calm
a_peaceful_i = (a3_9 - aMean)/aSD, #Ipsatize peaceful
a_serene_i = (a3_10 - aMean)/aSD) %>% #Ipsatize serene
rowwise %>%
mutate(aHAP_i = mean(c(a_enthus_i, a_excited_i, a_elated_i, a_euphoric_i), na.rm = T), #Compute ipsatized HAP
aLAP_i = mean(c(a_relaxed_i, a_calm_i, a_peaceful_i, a_serene_i), na.rm = T)) %>% #Compute ipsatized LAP
dplyr::select(pID, aHAP_i, aLAP_i) %>% #Keep only the ipsatized HAP and LAP scores
distinct()
Ipsatized_Ideal <- Data %>%
dplyr::select(pID, i1_1:i4_9) %>% #Only select ideal affect variables
ungroup() %>%
mutate(iSD = apply(.[,2:ncol(.)], na.rm = T, 1, sd)) %>% #Compute rowwise SD
mutate(iMean = rowMeans(dplyr::select(., contains("_")), na.rm = T)) %>% #Compute rowwise means
mutate(i_enthus_i = (i1_1 - iMean)/iSD, #Ipsatize enthusiastic
i_excited_i = (i1_7 - iMean)/iSD, #Ipsatize excited
i_elated_i = (i1_9 - iMean)/iSD, #Ipsatize elated
i_euphoric_i = (i2_9 - iMean)/iSD, #Ipsatize euphoric
i_relaxed_i = (i1_6 - iMean)/iSD, #Ipsatize relaxed
i_calm_i = (i3_3 - iMean)/iSD, #Ipsatize calm
i_peaceful_i = (i3_9 - iMean)/iSD, #Ipsatize peaceful
i_serene_i = (i3_10 - iMean)/iSD) %>% #Ipsatize serene
rowwise %>%
mutate(iHAP_i = mean(c(i_enthus_i, i_excited_i, i_elated_i, i_euphoric_i), na.rm = T), #Compute ipsatized HAP
iLAP_i = mean(c(i_relaxed_i, i_calm_i, i_peaceful_i, i_serene_i), na.rm = T)) %>% #Compute ipsatized LAP
dplyr::select(pID, iHAP_i, iLAP_i) %>% #Keep only the ipsatized HAP and LAP scores
distinct()
Data <- left_join(Data, Ipsatized_Actual, by = "pID") #Add ipsatized actual affect scores to original dataframe
Data <- left_join(Data, Ipsatized_Ideal, by = "pID") #Add ipsatized ideal affect scores to original dataframeDemographics
#Create dataframe with just demographics vars (so that each participant doesn't get duplicated 72 times for each trial)
Demographics <- Data %>%
dplyr::select(pID, Age, Gender, Education, FamSES1, FamSES2, SelfSES1, SelfSES2, iHAP, iLAP, iHAP_i, iLAP_i, aHAP, aLAP) %>%
distinct()
#Age
Demographics %>%
ungroup() %>%
summarize(Age_Mean = mean(Age, na.rm = T),
Age_SD = sd(Age, na.rm = T))## # A tibble: 1 x 2
## Age_Mean Age_SD
## <dbl> <dbl>
## 1 24.9 3.46#Gender
Demographics %>%
ungroup() %>%
count(Gender)## Warning: Factor `Gender` contains implicit NA, consider using
## `forcats::fct_explicit_na`## # A tibble: 3 x 2
## Gender n
## <fct> <int>
## 1 Male 25
## 2 Female 58
## 3 <NA> 4#Education
Demographics %>%
ungroup() %>%
count(Education)## # A tibble: 6 x 2
## Education n
## <fct> <int>
## 1 High School/GED 2
## 2 Some college (not currently enrolled) 9
## 3 Some college (currently enrolled) 17
## 4 Associates degree 3
## 5 BA/BS degree 38
## 6 Master's Degree 18#Plot
ggplot(Demographics, aes(x = Education)) +
geom_bar() +
xlab("Education")
#SES
Demographics %>%
ungroup() %>%
count(FamSES2)## Warning: Factor `FamSES2` contains implicit NA, consider using
## `forcats::fct_explicit_na`## # A tibble: 8 x 2
## FamSES2 n
## <fct> <int>
## 1 <10K 1
## 2 20-30K 3
## 3 30-40K 4
## 4 40-50K 8
## 5 50-75K 18
## 6 75-100K 20
## 7 >100K 32
## 8 <NA> 1#Plot
ggplot(Demographics, aes(x = FamSES2)) +
geom_bar() +
xlab("SES")
#AVI
Demographics %>%
ungroup() %>%
summarize(iHAP_Mean = mean(iHAP, na.rm = T),
iHAP_SD = sd(iHAP, na.rm = T),
iLAP_Mean = mean(iLAP, na.rm = T),
iLAP_SD = sd(iLAP, na.rm = T),
iHAPi_Mean = mean(iHAP_i, na.rm = T),
iHAPi_SD = sd(iHAP_i, na.rm = T),
iLAPi_Mean = mean(iLAP_i, na.rm = T),
iLAPi_SD = sd(iLAP_i, na.rm = T),
aHAP_Mean = mean(aHAP, na.rm = T),
aHAP_SD = sd(aHAP, na.rm = T),
aLAP_Mean = mean(aLAP, na.rm = T),
aLAP_SD = sd(aLAP, na.rm = T))## # A tibble: 1 x 12
## iHAP_Mean iHAP_SD iLAP_Mean iLAP_SD iHAPi_Mean iHAPi_SD iLAPi_Mean
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 3.34 0.793 3.96 0.795 0.636 0.382 1.09
## # … with 5 more variables: iLAPi_SD <dbl>, aHAP_Mean <dbl>, aHAP_SD <dbl>,
## # aLAP_Mean <dbl>, aLAP_SD <dbl>#To create a bar graph, I have to tidy up the dataframe so that iHAP, iLAP, iHAP_i, iLAP_i, aHAP and aLAP (which are currently individual columns) are now rows under IdealAffect, IpsatizedIdealAffect, and ActualAffect, respectively:
TidyAVI <- Data %>%
gather('iHAP', 'iLAP', key = "IdealAffect", value = "Valuation") %>%
gather('iHAP_i', 'iLAP_i', key = "IpsatizedIdealAffect", value = "Valuation_i") %>%
gather('aHAP', 'aLAP', key = "ActualAffect", value = "Reported")
#Ideal affect (not ipsatized)
ggplot(TidyAVI, aes(x = IdealAffect, y = Valuation)) +
geom_bar(position="dodge", stat = "summary", fun.y = "mean", fill="light grey") +
stat_summary(geom = "errorbar", fun.data = mean_se, position=position_dodge(.9), width = 0.5) +
scale_y_continuous(limits=c(1,5),oob = rescale_none) +
theme(text = element_text(size=20, family="serif")) +
xlab("Ideal Affect")
#Ideal affect (ipsatized)
ggplot(TidyAVI, aes(x = IpsatizedIdealAffect, y = Valuation_i)) +
geom_bar(position="dodge", stat = "summary", fun.y = "mean", fill="light grey") +
stat_summary(geom = "errorbar", fun.data = mean_se, position=position_dodge(.9), width = 0.5) +
theme(text = element_text(size=20, family="serif")) +
xlab("Ideal Affect (Ipsatized)")## Warning: Removed 576 rows containing non-finite values (stat_summary).## Warning: Removed 576 rows containing non-finite values (stat_summary).
#Actual affect
ggplot(TidyAVI, aes(x = ActualAffect, y = Reported)) +
geom_bar(position="dodge", stat = "summary", fun.y = "mean", fill="light grey") +
stat_summary(geom = "errorbar", fun.data = mean_se, position=position_dodge(.9), width = 0.5) +
scale_y_continuous(limits=c(1,5),oob = rescale_none) +
theme(text = element_text(size=20, family="serif")) +
xlab("Actual Affect")
Compute Cronbach’s alpha to make sure ideal and actual affect items hang together
#Actual HAP
aHAP_alpha <- Data %>%
dplyr::select(a1_1, a1_7, a1_9, a2_9) %>%
drop_na()
Alpha <- reliability(cov(data.matrix(aHAP_alpha))) #Convert to matrix and calculate alpha
Alpha$alpha #Output just alpha## alpha
## 0.8259208#Actual LAP
aLAP_alpha <- Data %>%
dplyr::select(a1_6, a3_3, a3_9, a3_10) %>%
drop_na()
Alpha <- reliability(cov(data.matrix(aLAP_alpha))) #Convert to matrix and calculate alpha
Alpha$alpha #Output just alpha## alpha
## 0.8339752#Ideal HAP
iHAP_alpha <- Data %>%
dplyr::select(i1_1, i1_7, i1_9, i2_9) %>%
drop_na()
Alpha <- reliability(cov(data.matrix(iHAP_alpha))) #Convert to matrix and calculate alpha
Alpha$alpha #Output just alpha## alpha
## 0.7781464#Ideal LAP
iLAP_alpha <- Data %>%
dplyr::select(i1_6, i3_3, i3_9, i3_10) %>%
drop_na()
Alpha <- reliability(cov(data.matrix(iLAP_alpha))) #Convert to matrix and calculate alpha
Alpha$alpha #Output just alpha## alpha
## 0.8261ANALYZE: Confirmatory analysis
#Extract just variables from Data that are relevant for analyses
Data <- Data %>%
dplyr::select(pID, email, trial_num, sex, race, identity, expression, reputation, exact_value, rt, response, Age, Gender, aHAP, aLAP, iHAP, iLAP, aHAP_i, aLAP_i, iHAP_i, iLAP_i, responseNoVar, FamSES2)
#Subset data to just expression levels "calm" vs. "excited" (without "neutral")
Data <- Data %>%
filter(expression != "neut")
#Refactor expression so that neut is no longer a level
Data$expression <- factor(Data$expression)
#Check levels
contrasts(Data$expression)## exci
## calm 0
## exci 1#Optional: Take out participants who gave the same amount across all trials using: filter(responseNoVar==0)Is there an interaction between target expression and target reputation on giving?
Data_noNA <- Data %>%
filter(!is.na(FamSES2))
#Model 1: entering Player 2's reputation and expression as predictors for amount given by participants (entering participant (pID) as random effect)
model1 <- lmer(data = Data_noNA, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_noNA
##
## AIC BIC logLik deviance df.resid
## 15644.0 15694.6 -7814.0 15628.0 4120
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8351 -0.5696 -0.0118 0.5860 4.6617
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 3.712 1.927
## Residual 2.357 1.535
## Number of obs: 4128, groups: pID, 86
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 2.555e+00 2.158e-01 9.759e+01 11.839
## reputationmod 2.307e+00 8.278e-02 4.042e+03 27.864
## reputationhigh 4.581e+00 8.278e-02 4.042e+03 55.342
## expressionexci -1.453e-03 8.278e-02 4.042e+03 -0.018
## reputationmod:expressionexci -1.730e-01 1.171e-01 4.042e+03 -1.477
## reputationhigh:expressionexci -5.087e-02 1.171e-01 4.042e+03 -0.435
## Pr(>|t|)
## (Intercept) <2e-16 ***
## reputationmod <2e-16 ***
## reputationhigh <2e-16 ***
## expressionexci 0.986
## reputationmod:expressionexci 0.140
## reputationhigh:expressionexci 0.664
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnm rpttnh exprss rpttnm:
## reputatinmd -0.192
## reputatnhgh -0.192 0.500
## expressinxc -0.192 0.500 0.500
## rpttnmd:xpr 0.136 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.136 -0.354 -0.707 -0.707 0.500#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_noNA$reputation <- relevel(Data_noNA$reputation, ref = "mod")
model1 <- lmer(data = Data_noNA, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_noNA
##
## AIC BIC logLik deviance df.resid
## 15644.0 15694.6 -7814.0 15628.0 4120
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8351 -0.5696 -0.0118 0.5860 4.6617
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 3.712 1.927
## Residual 2.357 1.535
## Number of obs: 4128, groups: pID, 86
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.86192 0.21584 97.58856 22.526
## reputationlow -2.30669 0.08278 4042.00000 -27.864
## reputationhigh 2.27471 0.08278 4042.00000 27.478
## expressionexci -0.17442 0.08278 4042.00000 -2.107
## reputationlow:expressionexci 0.17297 0.11707 4042.00000 1.477
## reputationhigh:expressionexci 0.12209 0.11707 4042.00000 1.043
## Pr(>|t|)
## (Intercept) <2e-16 ***
## reputationlow <2e-16 ***
## reputationhigh <2e-16 ***
## expressionexci 0.0352 *
## reputationlow:expressionexci 0.1396
## reputationhigh:expressionexci 0.2971
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnl rpttnh exprss rpttnl:
## reputatinlw -0.192
## reputatnhgh -0.192 0.500
## expressinxc -0.192 0.500 0.500
## rpttnlw:xpr 0.136 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.136 -0.354 -0.707 -0.707 0.500#Model 2: adding Player 2 race and sex and participant SES as predictors for amount given by participants
Data_noNA$reputation <- relevel(Data_noNA$reputation, ref = "low") #relevel back to low as reference group
model2 <- lmer(data = Data_noNA, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_noNA
##
## AIC BIC logLik deviance df.resid
## 15622.1 15824.6 -7779.1 15558.1 4096
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8290 -0.5752 -0.0065 0.5584 4.6085
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.935 1.713
## Residual 2.329 1.526
## Number of obs: 4128, groups: pID, 86
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 7.52980 1.73109
## reputationmod 2.72674 0.16455
## reputationhigh 4.62209 0.16455
## expressionexci 0.09884 0.16455
## racewhite 0.29651 0.16455
## sexmale 0.20349 0.16455
## FamSES220-30K -7.36111 1.99456
## FamSES230-40K -6.10937 1.93123
## FamSES240-50K -4.43490 1.83212
## FamSES250-75K -5.43981 1.77468
## FamSES275-100K -5.82917 1.77000
## FamSES2>100K -4.67383 1.75412
## reputationmod:expressionexci -0.49419 0.23271
## reputationhigh:expressionexci 0.26163 0.23271
## reputationmod:racewhite -0.61047 0.23271
## reputationhigh:racewhite -0.17442 0.23271
## expressionexci:racewhite -0.40116 0.23271
## reputationmod:sexmale -0.34884 0.23271
## reputationhigh:sexmale -0.07558 0.23271
## expressionexci:sexmale 0.10465 0.23271
## racewhite:sexmale -0.15116 0.23271
## reputationmod:expressionexci:racewhite 0.65698 0.32910
## reputationhigh:expressionexci:racewhite -0.25581 0.32910
## reputationmod:expressionexci:sexmale -0.13953 0.32910
## reputationhigh:expressionexci:sexmale -0.30814 0.32910
## reputationmod:racewhite:sexmale 0.23837 0.32910
## reputationhigh:racewhite:sexmale 0.33721 0.32910
## expressionexci:racewhite:sexmale 0.19186 0.32910
## reputationmod:expressionexci:racewhite:sexmale 0.25000 0.46541
## reputationhigh:expressionexci:racewhite:sexmale -0.12209 0.46541
## df t value
## (Intercept) 86.74949 4.350
## reputationmod 4041.99999 16.571
## reputationhigh 4041.99999 28.090
## expressionexci 4041.99998 0.601
## racewhite 4041.99998 1.802
## sexmale 4041.99998 1.237
## FamSES220-30K 85.99999 -3.691
## FamSES230-40K 86.00000 -3.163
## FamSES240-50K 86.00000 -2.421
## FamSES250-75K 86.00000 -3.065
## FamSES275-100K 86.00000 -3.293
## FamSES2>100K 86.00000 -2.664
## reputationmod:expressionexci 4042.00000 -2.124
## reputationhigh:expressionexci 4042.00000 1.124
## reputationmod:racewhite 4042.00000 -2.623
## reputationhigh:racewhite 4042.00000 -0.750
## expressionexci:racewhite 4042.00000 -1.724
## reputationmod:sexmale 4042.00000 -1.499
## reputationhigh:sexmale 4042.00000 -0.325
## expressionexci:sexmale 4042.00000 0.450
## racewhite:sexmale 4042.00000 -0.650
## reputationmod:expressionexci:racewhite 4042.00001 1.996
## reputationhigh:expressionexci:racewhite 4042.00001 -0.777
## reputationmod:expressionexci:sexmale 4042.00001 -0.424
## reputationhigh:expressionexci:sexmale 4042.00001 -0.936
## reputationmod:racewhite:sexmale 4042.00001 0.724
## reputationhigh:racewhite:sexmale 4042.00001 1.025
## expressionexci:racewhite:sexmale 4042.00002 0.583
## reputationmod:expressionexci:racewhite:sexmale 4042.00002 0.537
## reputationhigh:expressionexci:racewhite:sexmale 4042.00002 -0.262
## Pr(>|t|)
## (Intercept) 3.7e-05 ***
## reputationmod < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.548100
## racewhite 0.071623 .
## sexmale 0.216288
## FamSES220-30K 0.000392 ***
## FamSES230-40K 0.002155 **
## FamSES240-50K 0.017597 *
## FamSES250-75K 0.002906 **
## FamSES275-100K 0.001438 **
## FamSES2>100K 0.009207 **
## reputationmod:expressionexci 0.033760 *
## reputationhigh:expressionexci 0.260958
## reputationmod:racewhite 0.008740 **
## reputationhigh:racewhite 0.453585
## expressionexci:racewhite 0.084801 .
## reputationmod:sexmale 0.133939
## reputationhigh:sexmale 0.745354
## expressionexci:sexmale 0.652940
## racewhite:sexmale 0.515995
## reputationmod:expressionexci:racewhite 0.045967 *
## reputationhigh:expressionexci:racewhite 0.437013
## reputationmod:expressionexci:sexmale 0.671592
## reputationhigh:expressionexci:sexmale 0.349163
## reputationmod:racewhite:sexmale 0.468908
## reputationhigh:racewhite:sexmale 0.305587
## expressionexci:racewhite:sexmale 0.559930
## reputationmod:expressionexci:racewhite:sexmale 0.591187
## reputationhigh:expressionexci:racewhite:sexmale 0.793078
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 30 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_noNA$reputation <- relevel(Data_noNA$reputation, ref = "mod")
model2 <- lmer(data = Data_noNA, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_noNA
##
## AIC BIC logLik deviance df.resid
## 15622.1 15824.6 -7779.1 15558.1 4096
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8290 -0.5752 -0.0065 0.5584 4.6085
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.935 1.713
## Residual 2.329 1.526
## Number of obs: 4128, groups: pID, 86
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 10.25654 1.73109
## reputationlow -2.72674 0.16455
## reputationhigh 1.89535 0.16455
## expressionexci -0.39535 0.16455
## racewhite -0.31395 0.16455
## sexmale -0.14535 0.16455
## FamSES220-30K -7.36111 1.99456
## FamSES230-40K -6.10938 1.93123
## FamSES240-50K -4.43490 1.83212
## FamSES250-75K -5.43981 1.77468
## FamSES275-100K -5.82917 1.77000
## FamSES2>100K -4.67383 1.75412
## reputationlow:expressionexci 0.49419 0.23271
## reputationhigh:expressionexci 0.75581 0.23271
## reputationlow:racewhite 0.61047 0.23271
## reputationhigh:racewhite 0.43605 0.23271
## expressionexci:racewhite 0.25581 0.23271
## reputationlow:sexmale 0.34884 0.23271
## reputationhigh:sexmale 0.27326 0.23271
## expressionexci:sexmale -0.03488 0.23271
## racewhite:sexmale 0.08721 0.23271
## reputationlow:expressionexci:racewhite -0.65698 0.32910
## reputationhigh:expressionexci:racewhite -0.91279 0.32910
## reputationlow:expressionexci:sexmale 0.13953 0.32910
## reputationhigh:expressionexci:sexmale -0.16860 0.32910
## reputationlow:racewhite:sexmale -0.23837 0.32910
## reputationhigh:racewhite:sexmale 0.09884 0.32910
## expressionexci:racewhite:sexmale 0.44186 0.32910
## reputationlow:expressionexci:racewhite:sexmale -0.25000 0.46541
## reputationhigh:expressionexci:racewhite:sexmale -0.37209 0.46541
## df t value
## (Intercept) 86.74956 5.925
## reputationlow 4042.00001 -16.571
## reputationhigh 4042.00001 11.519
## expressionexci 4042.00001 -2.403
## racewhite 4042.00001 -1.908
## sexmale 4042.00001 -0.883
## FamSES220-30K 86.00005 -3.691
## FamSES230-40K 86.00005 -3.163
## FamSES240-50K 86.00006 -2.421
## FamSES250-75K 86.00007 -3.065
## FamSES275-100K 86.00007 -3.293
## FamSES2>100K 86.00007 -2.664
## reputationlow:expressionexci 4042.00001 2.124
## reputationhigh:expressionexci 4042.00001 3.248
## reputationlow:racewhite 4042.00001 2.623
## reputationhigh:racewhite 4042.00001 1.874
## expressionexci:racewhite 4042.00001 1.099
## reputationlow:sexmale 4042.00001 1.499
## reputationhigh:sexmale 4042.00001 1.174
## expressionexci:sexmale 4042.00001 -0.150
## racewhite:sexmale 4042.00001 0.375
## reputationlow:expressionexci:racewhite 4042.00001 -1.996
## reputationhigh:expressionexci:racewhite 4042.00001 -2.774
## reputationlow:expressionexci:sexmale 4042.00001 0.424
## reputationhigh:expressionexci:sexmale 4042.00001 -0.512
## reputationlow:racewhite:sexmale 4042.00001 -0.724
## reputationhigh:racewhite:sexmale 4042.00001 0.300
## expressionexci:racewhite:sexmale 4042.00001 1.343
## reputationlow:expressionexci:racewhite:sexmale 4042.00001 -0.537
## reputationhigh:expressionexci:racewhite:sexmale 4042.00001 -0.799
## Pr(>|t|)
## (Intercept) 6.16e-08 ***
## reputationlow < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.016322 *
## racewhite 0.056465 .
## sexmale 0.377114
## FamSES220-30K 0.000392 ***
## FamSES230-40K 0.002155 **
## FamSES240-50K 0.017597 *
## FamSES250-75K 0.002906 **
## FamSES275-100K 0.001438 **
## FamSES2>100K 0.009207 **
## reputationlow:expressionexci 0.033760 *
## reputationhigh:expressionexci 0.001172 **
## reputationlow:racewhite 0.008740 **
## reputationhigh:racewhite 0.061029 .
## expressionexci:racewhite 0.271702
## reputationlow:sexmale 0.133939
## reputationhigh:sexmale 0.240362
## expressionexci:sexmale 0.880847
## racewhite:sexmale 0.707857
## reputationlow:expressionexci:racewhite 0.045967 *
## reputationhigh:expressionexci:racewhite 0.005569 **
## reputationlow:expressionexci:sexmale 0.671592
## reputationhigh:expressionexci:sexmale 0.608450
## reputationlow:racewhite:sexmale 0.468908
## reputationhigh:racewhite:sexmale 0.763941
## expressionexci:racewhite:sexmale 0.179460
## reputationlow:expressionexci:racewhite:sexmale 0.591187
## reputationhigh:expressionexci:racewhite:sexmale 0.424052
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 30 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#Model comparison
anova(model1, model2)## Data: Data_noNA
## Models:
## model1: response ~ reputation * expression + (1 | pID)
## model2: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## model2: pID)
## Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
## model1 8 15644 15695 -7814.0 15628
## model2 32 15622 15825 -7779.1 15558 69.813 24 2.333e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plot the above
Data %>%
ggplot(aes(x = reputation, y = response, fill = expression)) +
geom_bar(position="dodge", stat = "summary", fun.y = "mean") +
stat_summary(geom = "errorbar", fun.data = mean_se, position=position_dodge(.9), width = 0.2) +
scale_x_discrete(limits=c("low","mod", "high"),
labels=c("Low","Moderate", "High")) +
guides(fill=guide_legend(title="Expression")) +
scale_fill_manual(values=c("#94B3D7", "#D99594")) +
xlab("Player 2 Reputation") +
ylab("Amount Given ($)") +
ylim(c(0,10))
Individual Differences in Ideal Affect –> Ideal Affect Match –> Giving?
##For now, we only have European American data, so we will test whether there are individual differences in ideal affect:
#First, filter for just moderate reputation ("mod") because that's where it seems ideal affect matters
Data_ModRepOnly <- Data %>%
filter(reputation == "mod")
#Model 3: adding participant ideal HAP and Player 2 expression as predictors for amount given by participants, controlling for actual HAP
model3 <- lmer(data = Data_ModRepOnly, response ~ expression * scale(iHAP, scale = F) + scale(aHAP, scale = F) + (1|pID), REML=FALSE)
summary(model3)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ expression * scale(iHAP, scale = F) + scale(aHAP,
## scale = F) + (1 | pID)
## Data: Data_ModRepOnly
##
## AIC BIC logLik deviance df.resid
## 4348.3 4384.9 -2167.1 4334.3 1385
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -6.3364 -0.3401 0.0207 0.3579 8.8579
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 5.1771 2.2753
## Residual 0.9989 0.9995
## Number of obs: 1392, groups: pID, 87
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 4.83908 0.24687 89.08487
## expressionexci -0.17816 0.05358 1305.00000
## scale(iHAP, scale = F) 0.05755 0.35722 88.59362
## scale(aHAP, scale = F) 0.18962 0.40562 87.00000
## expressionexci:scale(iHAP, scale = F) -0.03890 0.06792 1305.00000
## t value Pr(>|t|)
## (Intercept) 19.602 < 2e-16 ***
## expressionexci -3.325 0.000908 ***
## scale(iHAP, scale = F) 0.161 0.872387
## scale(aHAP, scale = F) 0.467 0.641331
## expressionexci:scale(iHAP, scale = F) -0.573 0.566924
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) exprss scale(iHAP,scl=F) scale(aHAP,scl=F)
## expressinxc -0.109
## scale(iHAP,scl=F) 0.000 0.000
## scale(aHAP,scl=F) 0.000 0.000 -0.482
## e:(HAP,s=F) 0.000 0.000 -0.095 0.000#Without adding participant ID as a random variable
model3 <- lm(data = Data_ModRepOnly, response ~ expression * scale(iHAP, scale = F) + scale(aHAP, scale = F))
summary(model3)##
## Call:
## lm(formula = response ~ expression * scale(iHAP, scale = F) +
## scale(aHAP, scale = F), data = Data_ModRepOnly)
##
## Residuals:
## Min 1Q Median 3Q Max
## -4.884 -1.666 0.046 1.203 5.675
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.83908 0.09437 51.278 <2e-16
## expressionexci -0.17816 0.13346 -1.335 0.1821
## scale(iHAP, scale = F) 0.05755 0.12847 0.448 0.6543
## scale(aHAP, scale = F) 0.18962 0.11029 1.719 0.0858
## expressionexci:scale(iHAP, scale = F) -0.03890 0.16918 -0.230 0.8182
##
## (Intercept) ***
## expressionexci
## scale(iHAP, scale = F)
## scale(aHAP, scale = F) .
## expressionexci:scale(iHAP, scale = F)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.49 on 1387 degrees of freedom
## Multiple R-squared: 0.004848, Adjusted R-squared: 0.001978
## F-statistic: 1.689 on 4 and 1387 DF, p-value: 0.15#Model 4: adding participant ideal LAP and Player 2 expression as predictors for amount given by participants, controlling for actual LAP
model4 <- lmer(data = Data_ModRepOnly, response ~ expression * scale(iLAP, scale = F) + scale(aLAP, scale = F) + (1|pID), REML=FALSE)
summary(model4)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ expression * scale(iLAP, scale = F) + scale(aLAP,
## scale = F) + (1 | pID)
## Data: Data_ModRepOnly
##
## AIC BIC logLik deviance df.resid
## 4341.9 4378.5 -2163.9 4327.9 1385
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -6.3350 -0.3465 0.0289 0.3530 8.8789
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 4.7954 2.1898
## Residual 0.9991 0.9996
## Number of obs: 1392, groups: pID, 87
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 4.83908 0.23781 89.25003
## expressionexci -0.17816 0.05358 1305.00000
## scale(iLAP, scale = F) -0.72887 0.30867 89.13547
## scale(aLAP, scale = F) 0.57662 0.31642 87.00000
## expressionexci:scale(iLAP, scale = F) 0.02417 0.06779 1305.00000
## t value Pr(>|t|)
## (Intercept) 20.348 < 2e-16 ***
## expressionexci -3.325 0.000909 ***
## scale(iLAP, scale = F) -2.361 0.020391 *
## scale(aLAP, scale = F) 1.822 0.071840 .
## expressionexci:scale(iLAP, scale = F) 0.357 0.721460
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) exprss scale(iLAP,scl=F) scale(aLAP,scl=F)
## expressinxc -0.113
## scale(iLAP,scl=F) 0.000 0.000
## scale(aLAP,scl=F) 0.000 0.000 -0.224
## e:(LAP,s=F) 0.000 0.000 -0.110 0.000#Without adding participant ID as a random variable
model4 <- lm(data = Data_ModRepOnly, response ~ expression * scale(iLAP, scale = F) + scale(aLAP, scale = F))
summary(model4)##
## Call:
## lm(formula = response ~ expression * scale(iLAP, scale = F) +
## scale(aLAP, scale = F), data = Data_ModRepOnly)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5.1427 -1.4899 -0.1555 0.9565 5.7253
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.83908 0.09141 52.939 < 2e-16
## expressionexci -0.17816 0.12927 -1.378 0.168
## scale(iLAP, scale = F) -0.72887 0.11717 -6.221 6.55e-10
## scale(aLAP, scale = F) 0.57662 0.08655 6.662 3.88e-11
## expressionexci:scale(iLAP, scale = F) 0.02417 0.16354 0.148 0.883
##
## (Intercept) ***
## expressionexci
## scale(iLAP, scale = F) ***
## scale(aLAP, scale = F) ***
## expressionexci:scale(iLAP, scale = F)
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.412 on 1387 degrees of freedom
## Multiple R-squared: 0.06633, Adjusted R-squared: 0.06364
## F-statistic: 24.64 on 4 and 1387 DF, p-value: < 2.2e-16#Plot
expr_colors=c("#94B3D7","#D99594") #Set correct colors for "expression" variable for the jittered points in the graphs below
#iHAP
Data_ModRepOnly %>%
ggplot(aes(x = iHAP, y = response, fill = expression)) +
geom_jitter(aes(colour = expression, alpha=0.01), width = 0.2, height = 0.4) +
geom_smooth(method=lm, color='#2C3E50') +
guides(fill=guide_legend(title="Player 2 Expression")) +
scale_fill_manual(values=c("#94B3D7", "#D99594")) + #Set correct ribbon colors
scale_color_manual(values = expr_colors) + #Set correct jitter (point) colors
xlab("Participant ideal HAP") +
ylab("Amount Given ($)") +
ylim(c(0,10))## Warning: Removed 93 rows containing missing values (geom_point).
#iLAP
Data_ModRepOnly %>%
ggplot(aes(x = iLAP, y = response, fill = expression)) +
geom_jitter(aes(colour = expression, alpha=0.01), width = 0.2, height = 0.4) +
geom_smooth(method=lm, color='#2C3E50') +
guides(fill=guide_legend(title="Player 2 Expression")) +
scale_fill_manual(values=c("#94B3D7", "#D99594")) + #Set correct ribbon colors
scale_color_manual(values = expr_colors) + #Set correct jitter (point) colors
xlab("Participant ideal LAP") +
ylab("Amount Given ($)") +
ylim(c(0,10))## Warning: Removed 88 rows containing missing values (geom_point).
Gender analysis:
#Separate dataframe into male only vs. female only (not enough males to use as covariate)
Data_Males <- Data_noNA %>%
filter(Gender == "Male")
Data_Females <- Data_noNA %>%
filter(Gender == "Female")
#For males only:
#Model 1: entering Player 2's reputation and expression as predictors for amount given by participants (entering participant (pID) as random effect)
model1 <- lmer(data = Data_Males, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_Males
##
## AIC BIC logLik deviance df.resid
## 4688.3 4729.0 -2336.1 4672.3 1192
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2927 -0.7028 -0.0222 0.5561 4.1566
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 4.462 2.112
## Residual 2.621 1.619
## Number of obs: 1200, groups: pID, 25
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.9050 0.4377 28.1121 11.206
## reputationlow -1.9600 0.1619 1175.0000 -12.106
## reputationhigh 1.7200 0.1619 1175.0000 10.623
## expressionexci -0.2250 0.1619 1175.0000 -1.390
## reputationlow:expressionexci 0.2700 0.2290 1175.0000 1.179
## reputationhigh:expressionexci 0.1050 0.2290 1175.0000 0.459
## Pr(>|t|)
## (Intercept) 6.99e-12 ***
## reputationlow < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.165
## reputationlow:expressionexci 0.239
## reputationhigh:expressionexci 0.647
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnl rpttnh exprss rpttnl:
## reputatinlw -0.185
## reputatnhgh -0.185 0.500
## expressinxc -0.185 0.500 0.500
## rpttnlw:xpr 0.131 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.131 -0.354 -0.707 -0.707 0.500#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_Males$reputation <- relevel(Data_Males$reputation, ref = "mod")
model1 <- lmer(data = Data_Males, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_Males
##
## AIC BIC logLik deviance df.resid
## 4688.3 4729.0 -2336.1 4672.3 1192
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2927 -0.7028 -0.0222 0.5561 4.1566
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 4.462 2.112
## Residual 2.621 1.619
## Number of obs: 1200, groups: pID, 25
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.9050 0.4377 28.1121 11.206
## reputationlow -1.9600 0.1619 1175.0000 -12.106
## reputationhigh 1.7200 0.1619 1175.0000 10.623
## expressionexci -0.2250 0.1619 1175.0000 -1.390
## reputationlow:expressionexci 0.2700 0.2290 1175.0000 1.179
## reputationhigh:expressionexci 0.1050 0.2290 1175.0000 0.459
## Pr(>|t|)
## (Intercept) 6.99e-12 ***
## reputationlow < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.165
## reputationlow:expressionexci 0.239
## reputationhigh:expressionexci 0.647
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnl rpttnh exprss rpttnl:
## reputatinlw -0.185
## reputatnhgh -0.185 0.500
## expressinxc -0.185 0.500 0.500
## rpttnlw:xpr 0.131 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.131 -0.354 -0.707 -0.707 0.500#Model 2: adding Player 2 race and sex and participant SES as predictors for amount given by participants
Data_Males$reputation <- relevel(Data_Males$reputation, ref = "low") #relevel back to low as reference group
model2 <- lmer(data = Data_Males, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_Males
##
## AIC BIC logLik deviance df.resid
## 4709.3 4867.1 -2323.6 4647.3 1169
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2211 -0.6970 -0.0109 0.5791 4.0851
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 3.501 1.871
## Residual 2.579 1.606
## Number of obs: 1200, groups: pID, 25
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 0.7439 1.1111
## reputationmod 2.2000 0.3212
## reputationhigh 3.7200 0.3212
## expressionexci 0.1000 0.3212
## racewhite 0.0800 0.3212
## sexmale -0.0600 0.3212
## FamSES230-40K 2.5903 1.7212
## FamSES240-50K 2.1736 2.1772
## FamSES250-75K 2.1944 1.3332
## FamSES275-100K 1.8075 1.3011
## FamSES2>100K 3.3715 1.3332
## reputationmod:expressionexci -0.4000 0.4542
## reputationhigh:expressionexci 0.0800 0.4542
## reputationmod:racewhite -0.3800 0.4542
## reputationhigh:racewhite -0.1400 0.4542
## expressionexci:racewhite -0.3800 0.4542
## reputationmod:sexmale 0.0200 0.4542
## reputationhigh:sexmale -0.0800 0.4542
## expressionexci:sexmale 0.2800 0.4542
## racewhite:sexmale 0.2200 0.4542
## reputationmod:expressionexci:racewhite 0.4400 0.6424
## reputationhigh:expressionexci:racewhite -0.1000 0.6424
## reputationmod:expressionexci:sexmale -0.7200 0.6424
## reputationhigh:expressionexci:sexmale -0.2200 0.6424
## reputationmod:racewhite:sexmale -0.2400 0.6424
## reputationhigh:racewhite:sexmale 0.2800 0.6424
## expressionexci:racewhite:sexmale -0.0200 0.6424
## reputationmod:expressionexci:racewhite:sexmale 1.0800 0.9085
## reputationhigh:expressionexci:racewhite:sexmale -0.3400 0.9085
## df t value Pr(>|t|)
## (Intercept) 27.1284 0.670 0.5088
## reputationmod 1175.0000 6.849 1.19e-11
## reputationhigh 1175.0000 11.581 < 2e-16
## expressionexci 1175.0000 0.311 0.7556
## racewhite 1175.0000 0.249 0.8034
## sexmale 1175.0000 -0.187 0.8519
## FamSES230-40K 25.0000 1.505 0.1449
## FamSES240-50K 25.0000 0.998 0.3277
## FamSES250-75K 25.0000 1.646 0.1123
## FamSES275-100K 25.0000 1.389 0.1770
## FamSES2>100K 25.0000 2.529 0.0181
## reputationmod:expressionexci 1175.0000 -0.881 0.3787
## reputationhigh:expressionexci 1175.0000 0.176 0.8602
## reputationmod:racewhite 1175.0000 -0.837 0.4030
## reputationhigh:racewhite 1175.0000 -0.308 0.7580
## expressionexci:racewhite 1175.0000 -0.837 0.4030
## reputationmod:sexmale 1175.0000 0.044 0.9649
## reputationhigh:sexmale 1175.0000 -0.176 0.8602
## expressionexci:sexmale 1175.0000 0.616 0.5378
## racewhite:sexmale 1175.0000 0.484 0.6283
## reputationmod:expressionexci:racewhite 1175.0000 0.685 0.4935
## reputationhigh:expressionexci:racewhite 1175.0000 -0.156 0.8763
## reputationmod:expressionexci:sexmale 1175.0000 -1.121 0.2626
## reputationhigh:expressionexci:sexmale 1175.0000 -0.342 0.7321
## reputationmod:racewhite:sexmale 1175.0000 -0.374 0.7088
## reputationhigh:racewhite:sexmale 1175.0000 0.436 0.6630
## expressionexci:racewhite:sexmale 1175.0000 -0.031 0.9752
## reputationmod:expressionexci:racewhite:sexmale 1175.0000 1.189 0.2348
## reputationhigh:expressionexci:racewhite:sexmale 1175.0000 -0.374 0.7083
##
## (Intercept)
## reputationmod ***
## reputationhigh ***
## expressionexci
## racewhite
## sexmale
## FamSES230-40K
## FamSES240-50K
## FamSES250-75K
## FamSES275-100K
## FamSES2>100K *
## reputationmod:expressionexci
## reputationhigh:expressionexci
## reputationmod:racewhite
## reputationhigh:racewhite
## expressionexci:racewhite
## reputationmod:sexmale
## reputationhigh:sexmale
## expressionexci:sexmale
## racewhite:sexmale
## reputationmod:expressionexci:racewhite
## reputationhigh:expressionexci:racewhite
## reputationmod:expressionexci:sexmale
## reputationhigh:expressionexci:sexmale
## reputationmod:racewhite:sexmale
## reputationhigh:racewhite:sexmale
## expressionexci:racewhite:sexmale
## reputationmod:expressionexci:racewhite:sexmale
## reputationhigh:expressionexci:racewhite:sexmale
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 29 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_Males$reputation <- relevel(Data_Males$reputation, ref = "mod")
model2 <- lmer(data = Data_Males, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_Males
##
## AIC BIC logLik deviance df.resid
## 4709.3 4867.1 -2323.6 4647.3 1169
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.2211 -0.6970 -0.0109 0.5791 4.0851
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 3.501 1.871
## Residual 2.579 1.606
## Number of obs: 1200, groups: pID, 25
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 2.9439 1.1111
## reputationlow -2.2000 0.3212
## reputationhigh 1.5200 0.3212
## expressionexci -0.3000 0.3212
## racewhite -0.3000 0.3212
## sexmale -0.0400 0.3212
## FamSES230-40K 2.5903 1.7212
## FamSES240-50K 2.1736 2.1772
## FamSES250-75K 2.1944 1.3332
## FamSES275-100K 1.8075 1.3011
## FamSES2>100K 3.3715 1.3332
## reputationlow:expressionexci 0.4000 0.4542
## reputationhigh:expressionexci 0.4800 0.4542
## reputationlow:racewhite 0.3800 0.4542
## reputationhigh:racewhite 0.2400 0.4542
## expressionexci:racewhite 0.0600 0.4542
## reputationlow:sexmale -0.0200 0.4542
## reputationhigh:sexmale -0.1000 0.4542
## expressionexci:sexmale -0.4400 0.4542
## racewhite:sexmale -0.0200 0.4542
## reputationlow:expressionexci:racewhite -0.4400 0.6424
## reputationhigh:expressionexci:racewhite -0.5400 0.6424
## reputationlow:expressionexci:sexmale 0.7200 0.6424
## reputationhigh:expressionexci:sexmale 0.5000 0.6424
## reputationlow:racewhite:sexmale 0.2400 0.6424
## reputationhigh:racewhite:sexmale 0.5200 0.6424
## expressionexci:racewhite:sexmale 1.0600 0.6424
## reputationlow:expressionexci:racewhite:sexmale -1.0800 0.9085
## reputationhigh:expressionexci:racewhite:sexmale -1.4200 0.9085
## df t value Pr(>|t|)
## (Intercept) 27.1284 2.650 0.0133
## reputationlow 1175.0000 -6.849 1.19e-11
## reputationhigh 1175.0000 4.732 2.49e-06
## expressionexci 1175.0000 -0.934 0.3505
## racewhite 1175.0000 -0.934 0.3505
## sexmale 1175.0000 -0.125 0.9009
## FamSES230-40K 25.0000 1.505 0.1449
## FamSES240-50K 25.0000 0.998 0.3277
## FamSES250-75K 25.0000 1.646 0.1123
## FamSES275-100K 25.0000 1.389 0.1770
## FamSES2>100K 25.0000 2.529 0.0181
## reputationlow:expressionexci 1175.0000 0.881 0.3787
## reputationhigh:expressionexci 1175.0000 1.057 0.2909
## reputationlow:racewhite 1175.0000 0.837 0.4030
## reputationhigh:racewhite 1175.0000 0.528 0.5974
## expressionexci:racewhite 1175.0000 0.132 0.8949
## reputationlow:sexmale 1175.0000 -0.044 0.9649
## reputationhigh:sexmale 1175.0000 -0.220 0.8258
## expressionexci:sexmale 1175.0000 -0.969 0.3329
## racewhite:sexmale 1175.0000 -0.044 0.9649
## reputationlow:expressionexci:racewhite 1175.0000 -0.685 0.4935
## reputationhigh:expressionexci:racewhite 1175.0000 -0.841 0.4008
## reputationlow:expressionexci:sexmale 1175.0000 1.121 0.2626
## reputationhigh:expressionexci:sexmale 1175.0000 0.778 0.4365
## reputationlow:racewhite:sexmale 1175.0000 0.374 0.7088
## reputationhigh:racewhite:sexmale 1175.0000 0.809 0.4184
## expressionexci:racewhite:sexmale 1175.0000 1.650 0.0992
## reputationlow:expressionexci:racewhite:sexmale 1175.0000 -1.189 0.2348
## reputationhigh:expressionexci:racewhite:sexmale 1175.0000 -1.563 0.1183
##
## (Intercept) *
## reputationlow ***
## reputationhigh ***
## expressionexci
## racewhite
## sexmale
## FamSES230-40K
## FamSES240-50K
## FamSES250-75K
## FamSES275-100K
## FamSES2>100K *
## reputationlow:expressionexci
## reputationhigh:expressionexci
## reputationlow:racewhite
## reputationhigh:racewhite
## expressionexci:racewhite
## reputationlow:sexmale
## reputationhigh:sexmale
## expressionexci:sexmale
## racewhite:sexmale
## reputationlow:expressionexci:racewhite
## reputationhigh:expressionexci:racewhite
## reputationlow:expressionexci:sexmale
## reputationhigh:expressionexci:sexmale
## reputationlow:racewhite:sexmale
## reputationhigh:racewhite:sexmale
## expressionexci:racewhite:sexmale .
## reputationlow:expressionexci:racewhite:sexmale
## reputationhigh:expressionexci:racewhite:sexmale
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 29 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#Model comparison
anova(model1, model2)## Data: Data_Males
## Models:
## model1: response ~ reputation * expression + (1 | pID)
## model2: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## model2: pID)
## Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
## model1 8 4688.3 4729.0 -2336.1 4672.3
## model2 31 4709.3 4867.1 -2323.6 4647.3 24.997 23 0.3504#######################################################################
#For females only:
#Model 1: entering Player 2's reputation and expression as predictors for amount given by participants (entering participant (pID) as random effect)
model1 <- lmer(data = Data_Females, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_Females
##
## AIC BIC logLik deviance df.resid
## 10017.3 10064.6 -5000.7 10001.3 2728
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.0953 -0.5015 -0.0309 0.5217 4.7204
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.630 1.622
## Residual 2.079 1.442
## Number of obs: 2736, groups: pID, 57
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.66447 0.22515 66.59952 20.718
## reputationlow -2.53509 0.09548 2679.00001 -26.551
## reputationhigh 2.52851 0.09548 2679.00001 26.482
## expressionexci -0.17544 0.09548 2679.00001 -1.837
## reputationlow:expressionexci 0.17763 0.13503 2679.00001 1.315
## reputationhigh:expressionexci 0.14254 0.13503 2679.00001 1.056
## Pr(>|t|)
## (Intercept) <2e-16 ***
## reputationlow <2e-16 ***
## reputationhigh <2e-16 ***
## expressionexci 0.0663 .
## reputationlow:expressionexci 0.1885
## reputationhigh:expressionexci 0.2912
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnl rpttnh exprss rpttnl:
## reputatinlw -0.212
## reputatnhgh -0.212 0.500
## expressinxc -0.212 0.500 0.500
## rpttnlw:xpr 0.150 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.150 -0.354 -0.707 -0.707 0.500#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_Females$reputation <- relevel(Data_Females$reputation, ref = "mod")
model1 <- lmer(data = Data_Females, response ~ reputation * expression + (1|pID), REML=FALSE)
summary(model1)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression + (1 | pID)
## Data: Data_Females
##
## AIC BIC logLik deviance df.resid
## 10017.3 10064.6 -5000.7 10001.3 2728
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.0953 -0.5015 -0.0309 0.5217 4.7204
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.630 1.622
## Residual 2.079 1.442
## Number of obs: 2736, groups: pID, 57
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 4.66447 0.22515 66.59952 20.718
## reputationlow -2.53509 0.09548 2679.00001 -26.551
## reputationhigh 2.52851 0.09548 2679.00001 26.482
## expressionexci -0.17544 0.09548 2679.00001 -1.837
## reputationlow:expressionexci 0.17763 0.13503 2679.00001 1.315
## reputationhigh:expressionexci 0.14254 0.13503 2679.00001 1.056
## Pr(>|t|)
## (Intercept) <2e-16 ***
## reputationlow <2e-16 ***
## reputationhigh <2e-16 ***
## expressionexci 0.0663 .
## reputationlow:expressionexci 0.1885
## reputationhigh:expressionexci 0.2912
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) rpttnl rpttnh exprss rpttnl:
## reputatinlw -0.212
## reputatnhgh -0.212 0.500
## expressinxc -0.212 0.500 0.500
## rpttnlw:xpr 0.150 -0.707 -0.354 -0.707
## rpttnhgh:xp 0.150 -0.354 -0.707 -0.707 0.500#Model 2: adding Player 2 race and sex and participant SES as predictors for amount given by participants
Data_Females$reputation <- relevel(Data_Females$reputation, ref = "low") #relevel back to low as reference group
model2 <- lmer(data = Data_Females, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_Females
##
## AIC BIC logLik deviance df.resid
## 10011.6 10189.0 -4975.8 9951.6 2706
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.9127 -0.5205 -0.0172 0.5160 4.6117
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.181 1.477
## Residual 2.048 1.431
## Number of obs: 2736, groups: pID, 57
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) -0.22496 1.06254
## reputationmod 3.03509 0.18957
## reputationhigh 5.09649 0.18957
## expressionexci 0.11404 0.18957
## racewhite 0.39474 0.18957
## sexmale 0.31579 0.18957
## FamSES240-50K 3.12054 1.19559
## FamSES250-75K 1.67917 1.15505
## FamSES275-100K 1.47035 1.13262
## FamSES2>100K 2.42292 1.09577
## reputationmod:expressionexci -0.61404 0.26809
## reputationhigh:expressionexci 0.33333 0.26809
## reputationmod:racewhite -0.71053 0.26809
## reputationhigh:racewhite -0.18421 0.26809
## expressionexci:racewhite -0.40351 0.26809
## reputationmod:sexmale -0.50000 0.26809
## reputationhigh:sexmale -0.05263 0.26809
## expressionexci:sexmale 0.05263 0.26809
## racewhite:sexmale -0.30702 0.26809
## reputationmod:expressionexci:racewhite 0.82456 0.37914
## reputationhigh:expressionexci:racewhite -0.35965 0.37914
## reputationmod:expressionexci:sexmale 0.10526 0.37914
## reputationhigh:expressionexci:sexmale -0.39474 0.37914
## reputationmod:racewhite:sexmale 0.42105 0.37914
## reputationhigh:racewhite:sexmale 0.34211 0.37914
## expressionexci:racewhite:sexmale 0.25439 0.37914
## reputationmod:expressionexci:racewhite:sexmale -0.11404 0.53618
## reputationhigh:expressionexci:racewhite:sexmale 0.03509 0.53618
## df t value
## (Intercept) 58.77902 -0.212
## reputationmod 2678.99999 16.011
## reputationhigh 2678.99999 26.885
## expressionexci 2678.99999 0.602
## racewhite 2678.99999 2.082
## sexmale 2678.99999 1.666
## FamSES240-50K 57.00002 2.610
## FamSES250-75K 57.00002 1.454
## FamSES275-100K 57.00002 1.298
## FamSES2>100K 57.00002 2.211
## reputationmod:expressionexci 2678.99999 -2.290
## reputationhigh:expressionexci 2678.99999 1.243
## reputationmod:racewhite 2678.99999 -2.650
## reputationhigh:racewhite 2678.99999 -0.687
## expressionexci:racewhite 2678.99999 -1.505
## reputationmod:sexmale 2678.99999 -1.865
## reputationhigh:sexmale 2678.99999 -0.196
## expressionexci:sexmale 2678.99999 0.196
## racewhite:sexmale 2678.99999 -1.145
## reputationmod:expressionexci:racewhite 2678.99999 2.175
## reputationhigh:expressionexci:racewhite 2678.99999 -0.949
## reputationmod:expressionexci:sexmale 2678.99999 0.278
## reputationhigh:expressionexci:sexmale 2678.99999 -1.041
## reputationmod:racewhite:sexmale 2678.99999 1.111
## reputationhigh:racewhite:sexmale 2678.99999 0.902
## expressionexci:racewhite:sexmale 2678.99999 0.671
## reputationmod:expressionexci:racewhite:sexmale 2679.00000 -0.213
## reputationhigh:expressionexci:racewhite:sexmale 2678.99999 0.065
## Pr(>|t|)
## (Intercept) 0.83306
## reputationmod < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.54752
## racewhite 0.03741 *
## sexmale 0.09586 .
## FamSES240-50K 0.01155 *
## FamSES250-75K 0.15150
## FamSES275-100K 0.19945
## FamSES2>100K 0.03105 *
## reputationmod:expressionexci 0.02207 *
## reputationhigh:expressionexci 0.21384
## reputationmod:racewhite 0.00809 **
## reputationhigh:racewhite 0.49206
## expressionexci:racewhite 0.13241
## reputationmod:sexmale 0.06228 .
## reputationhigh:sexmale 0.84437
## expressionexci:sexmale 0.84437
## racewhite:sexmale 0.25223
## reputationmod:expressionexci:racewhite 0.02973 *
## reputationhigh:expressionexci:racewhite 0.34291
## reputationmod:expressionexci:sexmale 0.78131
## reputationhigh:expressionexci:sexmale 0.29790
## reputationmod:racewhite:sexmale 0.26686
## reputationhigh:racewhite:sexmale 0.36696
## expressionexci:racewhite:sexmale 0.50230
## reputationmod:expressionexci:racewhite:sexmale 0.83159
## reputationhigh:expressionexci:racewhite:sexmale 0.94783
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 28 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#To see all contrasts (not just with reference category), re-level reputation and run again:
Data_Females$reputation <- relevel(Data_Females$reputation, ref = "mod")
model2 <- lmer(data = Data_Females, response ~ reputation * expression * race * sex + FamSES2 + (1|pID), REML=FALSE)
summary(model2)## Linear mixed model fit by maximum likelihood . t-tests use
## Satterthwaite's method [lmerModLmerTest]
## Formula: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## pID)
## Data: Data_Females
##
## AIC BIC logLik deviance df.resid
## 10011.6 10189.0 -4975.8 9951.6 2706
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.9127 -0.5205 -0.0172 0.5160 4.6117
##
## Random effects:
## Groups Name Variance Std.Dev.
## pID (Intercept) 2.181 1.477
## Residual 2.048 1.431
## Number of obs: 2736, groups: pID, 57
##
## Fixed effects:
## Estimate Std. Error
## (Intercept) 2.81012 1.06254
## reputationlow -3.03509 0.18957
## reputationhigh 2.06140 0.18957
## expressionexci -0.50000 0.18957
## racewhite -0.31579 0.18957
## sexmale -0.18421 0.18957
## FamSES240-50K 3.12054 1.19559
## FamSES250-75K 1.67917 1.15505
## FamSES275-100K 1.47035 1.13262
## FamSES2>100K 2.42292 1.09577
## reputationlow:expressionexci 0.61404 0.26809
## reputationhigh:expressionexci 0.94737 0.26809
## reputationlow:racewhite 0.71053 0.26809
## reputationhigh:racewhite 0.52632 0.26809
## expressionexci:racewhite 0.42105 0.26809
## reputationlow:sexmale 0.50000 0.26809
## reputationhigh:sexmale 0.44737 0.26809
## expressionexci:sexmale 0.15789 0.26809
## racewhite:sexmale 0.11404 0.26809
## reputationlow:expressionexci:racewhite -0.82456 0.37914
## reputationhigh:expressionexci:racewhite -1.18421 0.37914
## reputationlow:expressionexci:sexmale -0.10526 0.37914
## reputationhigh:expressionexci:sexmale -0.50000 0.37914
## reputationlow:racewhite:sexmale -0.42105 0.37914
## reputationhigh:racewhite:sexmale -0.07895 0.37914
## expressionexci:racewhite:sexmale 0.14035 0.37914
## reputationlow:expressionexci:racewhite:sexmale 0.11404 0.53618
## reputationhigh:expressionexci:racewhite:sexmale 0.14912 0.53618
## df t value
## (Intercept) 58.77901 2.645
## reputationlow 2679.00001 -16.011
## reputationhigh 2679.00001 10.874
## expressionexci 2679.00001 -2.638
## racewhite 2679.00001 -1.666
## sexmale 2679.00001 -0.972
## FamSES240-50K 57.00000 2.610
## FamSES250-75K 57.00000 1.454
## FamSES275-100K 57.00000 1.298
## FamSES2>100K 57.00000 2.211
## reputationlow:expressionexci 2679.00001 2.290
## reputationhigh:expressionexci 2679.00001 3.534
## reputationlow:racewhite 2679.00001 2.650
## reputationhigh:racewhite 2679.00000 1.963
## expressionexci:racewhite 2679.00001 1.571
## reputationlow:sexmale 2679.00001 1.865
## reputationhigh:sexmale 2679.00000 1.669
## expressionexci:sexmale 2679.00001 0.589
## racewhite:sexmale 2679.00001 0.425
## reputationlow:expressionexci:racewhite 2679.00000 -2.175
## reputationhigh:expressionexci:racewhite 2679.00000 -3.123
## reputationlow:expressionexci:sexmale 2679.00000 -0.278
## reputationhigh:expressionexci:sexmale 2679.00000 -1.319
## reputationlow:racewhite:sexmale 2679.00000 -1.111
## reputationhigh:racewhite:sexmale 2679.00000 -0.208
## expressionexci:racewhite:sexmale 2679.00000 0.370
## reputationlow:expressionexci:racewhite:sexmale 2679.00000 0.213
## reputationhigh:expressionexci:racewhite:sexmale 2679.00000 0.278
## Pr(>|t|)
## (Intercept) 0.010469 *
## reputationlow < 2e-16 ***
## reputationhigh < 2e-16 ***
## expressionexci 0.008398 **
## racewhite 0.095862 .
## sexmale 0.331267
## FamSES240-50K 0.011546 *
## FamSES250-75K 0.151496
## FamSES275-100K 0.199451
## FamSES2>100K 0.031052 *
## reputationlow:expressionexci 0.022075 *
## reputationhigh:expressionexci 0.000417 ***
## reputationlow:racewhite 0.008088 **
## reputationhigh:racewhite 0.049725 *
## expressionexci:racewhite 0.116400
## reputationlow:sexmale 0.062283 .
## reputationhigh:sexmale 0.095287 .
## expressionexci:sexmale 0.555935
## racewhite:sexmale 0.670606
## reputationlow:expressionexci:racewhite 0.029729 *
## reputationhigh:expressionexci:racewhite 0.001807 **
## reputationlow:expressionexci:sexmale 0.781310
## reputationhigh:expressionexci:sexmale 0.187352
## reputationlow:racewhite:sexmale 0.266857
## reputationhigh:racewhite:sexmale 0.835065
## expressionexci:racewhite:sexmale 0.711272
## reputationlow:expressionexci:racewhite:sexmale 0.831592
## reputationhigh:expressionexci:racewhite:sexmale 0.780940
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 28 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it#Model comparison
anova(model1, model2)## Data: Data_Females
## Models:
## model1: response ~ reputation * expression + (1 | pID)
## model2: response ~ reputation * expression * race * sex + FamSES2 + (1 |
## model2: pID)
## Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
## model1 8 10017 10065 -5000.7 10001.3
## model2 30 10012 10189 -4975.8 9951.6 49.749 22 0.000634 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1