Neuroticism and the Gender Equality Paradox
Setup
The Gender Equality Paradox has also been dubbed the Nordic Paradox.
library(remotes)
remotes::install_github("Jo-Karl/ccpsyc")library(pacman); p_load(psych, lavaan, dplyr, ccpsyc, sjmisc, ggplot2, sandwich, lmtest, umx, tidyr)describe(IPIPA)## Warning in FUN(newX[, i], ...): no non-missing arguments to min; returning Inf## Warning in FUN(newX[, i], ...): no non-missing arguments to max; returning -Infwith(droplevels(IPIPA), list(levels = levels(as.factor(COUNTRY)), nlevels = nlevels(as.factor(COUNTRY))))## $levels
## [1] "" " Greece" " Japan" "Afghanistan" "Albania"
## [6] "Algeria" "Andorra" "Angola" "Anguilla" "Antarctica"
## [11] "Antigua" "Arabian Gul" "Argentina" "Armenia" "Aruba"
## [16] "Australia" "Austria" "Azerbaijan" "Bahamas" "Bahrain"
## [21] "Bangladesh" "Barbados" "Belarus" "Belgium" "Belize"
## [26] "Benin" "Bermuda" "Bhutan" "Bolivia" "Borneo"
## [31] "Bosnia Herz" "Botswana" "Bouvet Isla" "Brazil" "British Ind"
## [36] "British Vir" "Brunei" "Bulgaria" "Burkina Fas" "Burma"
## [41] "Burma(Myanm" "Burundi" "Cambodia" "Cameroon" "Canada"
## [46] "Cape Verde" "Cayman Isla" "Central Afr" "Chad" "Chile"
## [51] "China" "Christmas I" "Cocos (Keel" "Columbia" "Comoros"
## [56] "Congo" "Cook Island" "Costa Rica" "Croatia" "Cuba"
## [61] "Cyprus" "Czech Repub" "Denmark" "Djibouti" "Dominica"
## [66] "Dominican R" "East Timor" "Ecuador" "Egypt" "El Salvador"
## [71] "Equatorial" "Eritrea" "Estonia" "Ethiopia" "Faeroe Isla"
## [76] "Falkland Is" "Fiji" "Finland" "France" "French Guia"
## [81] "French Poly" "Gabon" "Gambia" "Georgia" "Germany"
## [86] "Ghana" "Gibraltar" "Greece" "Greenland" "Grenada"
## [91] "Guadeloupe" "Guam" "Guatemala" "Guinea" "Guinea-Biss"
## [96] "Guyana" "Haiti" "Honduras" "Hong Kong" "Hungary"
## [101] "Iceland" "India" "Indonesia" "Iran" "Iraq"
## [106] "Ireland" "Israel" "Italy" "Ivory Coast" "Jamaica"
## [111] "Japan" "Johnston Is" "Jordan" "Kazakhstan" "Kenya"
## [116] "Kuwait" "Kyrgystan" "Lao P.Dem.R" "Latvia" "Lebanon"
## [121] "Lesotho" "Liberia" "Libyan Arab" "Liechtenste" "Lithuania"
## [126] "Luxembourg" "Macau" "Macedonia" "Madagascar" "Malawi"
## [131] "Malaysia" "Maldives" "Mali" "Malta" "Marshall Is"
## [136] "Martinique" "Mauritania" "Mauritius" "Mexico" "Micronesia"
## [141] "Moldova" "Monaco" "Mongolia" "Montserrat" "Morocco"
## [146] "Mozambique" "Namibia" "Nepal" "Netherlands" "New Caledon"
## [151] "New Zealand" "Nicaragua" "Niger" "Nigeria" "Niue"
## [156] "Norfolk Isl" "North Korea" "Northern Ma" "Norway" "Oman"
## [161] "Pakistan" "Palau" "Panama" "Papua New G" "Paraguay"
## [166] "Peru" "Philippines" "Pitcairn Is" "Poland" "Portugal"
## [171] "Puerto Rico" "Qatar" "Republic of" "Reunion" "Romania"
## [176] "Russian Fed" "Rwanda" "Saint Helen" "Saint Kitts" "Samoa"
## [181] "San Marino" "Saudi Arabi" "Senegal" "Serbia" "Seychelles"
## [186] "Sierra Leon" "Singapore" "Slovakia" "Slovenia" "Solomon Isl"
## [191] "Somalia" "South Afric" "South Korea" "Spain" "Sri Lanka"
## [196] "St Lucia" "St Vincent" "Sudan" "Suriname" "Svalbard &"
## [201] "Swaziland" "Sweden" "Switzerland" "Syria" "Taiwan"
## [206] "Tanzania" "Thailand" "Togo" "Tokelau" "Tonga"
## [211] "Trinidad an" "Tunisia" "Turkey" "Turkmenista" "Turks and C"
## [216] "Tuvalu" "Uganda" "UK" "Ukraine" "United Arab"
## [221] "Uruguay" "USA" "Uzbekistan" "Vanuatu" "Vatican"
## [226] "Vatican Cit" "Venezuela" "Vietnam" "Virgin Isla" "W. Samoa"
## [231] "Wake Island" "Wallis and" "Western Sah" "Yemen" "Yugoslavia"
## [236] "Zaire" "Zambia" "Zimbabwe"
##
## $nlevels
## [1] 238I used the 2015 Gender Gap Index (henceforth GEI) values for available countries with sufficient numbers of participants from both sexes (around 100 and similar group sizes at the lower end, with a GEI entry). Those values are available here: https://www3.weforum.org/docs/GGGR2015/cover.pdf. The data went online in 2015; it is the IPIP300.por file available here: https://osf.io/tbmh5/. Giolla & Kajonius (2018) used a more limited subset of 22 countries and all OCEAN/Big Five scores and data from 2018. They found a positive relationship between the GEI and male-female differences.
IPIPA %>% group_by(COUNTRY, SEX) %>% summarise(Participants = length(COUNTRY))## `summarise()` has grouped output by 'COUNTRY'. You can override using the `.groups` argument.Applicable countries were Albania, Algeria, Argentina, Australia, Austria, Belgium, Brazil, Canada, China, Colombia (misspelled ‘Columbia’ in the dataset), Croatia, Denmark, Egypt, Finland, France, Germany, Greece, India, Indonesia, Ireland, Israel, Italy, Japan, Malaysia, Mexico, the Netherlands, New Zealand, Norway, Pakistan, the Philippines, Poland, Portugal, Romania, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey, UK, United Arab Emirates, and the USA.
#Select countries
IPIPA <- IPIPA[IPIPA$COUNTRY %in% c("Albania", "Algeria", "Argentina", "Australia", "Austria", "Belgium", "Brazil", "Canada", "China", "Columbia", "Croatia", "Denmark", "Egypt", "Finland", "France", "Germany", "Greece", "India", "Indonesia", "Ireland", "Israel", "Italy", "Japan", "Malaysia", "Mexico", "Netherlands", "New Zealand", "Norway", "Pakistan", "Philippines", "Poland", "Portugal", "Romania", "Russian Fed", "Singapore", "South Afric", "South Korea", "Spain", "Sweden", "Switzerland", "Thailand", "Turkey", "UK", "United Arab", "USA"), ]
with(droplevels(IPIPA), list(levels = levels(as.factor(COUNTRY)), nlevels = nlevels(as.factor(COUNTRY))))## $levels
## [1] "Albania" "Algeria" "Argentina" "Australia" "Austria"
## [6] "Belgium" "Brazil" "Canada" "China" "Columbia"
## [11] "Croatia" "Denmark" "Egypt" "Finland" "France"
## [16] "Germany" "Greece" "India" "Indonesia" "Ireland"
## [21] "Israel" "Italy" "Japan" "Malaysia" "Mexico"
## [26] "Netherlands" "New Zealand" "Norway" "Pakistan" "Philippines"
## [31] "Poland" "Portugal" "Romania" "Russian Fed" "Singapore"
## [36] "South Afric" "South Korea" "Spain" "Sweden" "Switzerland"
## [41] "Thailand" "Turkey" "UK" "United Arab" "USA"
##
## $nlevels
## [1] 45#Compute Variables
IPIPA <- mutate(IPIPA,
AchievementStriving = I20 + I50 + I80 + I110 + I140 + I170 + I200 - (6 - I230) - (6 - I260) - (6 - I290),
ActivityLevel = I17 + I47 + I77 + I107 + I137 - (6 - I167) - (6 - I197) - (6 - I227) - (6 - I257) - (6 - I287),
Adventurousness = I18 + I48 + I78 + I108 - (6 - I138) - (6 - I168) - (6 - I198) - (6 - I228) - (6 - I258) - (6 - I288),
Altruism = I14 + I44 + I74 + I104 + I134 - (6 - I164) - (6 - I194) - (6 - I224) - (6 - I254) - (6 - I284),
Anger = I6 + I36 + I66 + I96 + I126 - (6 - I156) - (6 - I187 )- (6 - I216) - (6 - I246) - (6 - I276),
Anxiety = I1 + I31 + I61 + I91 + I121 - (6 - I151) - (6 - I181) - (6 - I211) - (6 - I241) - (6 - I271),
ArtisticInterests = I8 + I38 + I68 + I98 + I128 - (6 - I158) - (6 - I188) - (6 - I218) - (6 - I248) - (6 - I278),
Assertiveness = I12 + I42 + I72 + I102 + I132 - (6 - I162) - (6 - I192) - (6 - I222) - (6 - I252) - (6 - I282),
Cautiousness = I30 + I60 + I90 - (6 - I120) - (6 - I150) - (6 - I180) - (6 - I210) - (6 - I240) - (6 - I270) - (6 - I300),
Cheerfulness = I27 + I57 + I87 + I117 + I147 + I177 + I207 + I237 - (6 - I267) - (6 - I297),
Cooperation = I19 + I49 + I79 - (6 - I109) - (6 - I139) - (6 - I169) - (6 - I199) - (6 - I229) - (6 - I259) - (6 - I289),
Depression = I11 + I41 + I71 + I101 + I131 + I161 + I191 - (6 - I221) - (6 - I251) - (6 - I281),
Dutifulness = I15 + I45 + I75 + I105 + I135 - (6 - I165) - (6 - I195) - (6 - I225) - (6 - I255) - (6 - I285),
Emotionality = I13 + I43 + I73 + I103 + I133 - (6 - I163) - (6 - I193) - (6 - I223) - (6 - I253) - (6 - I283),
ExcitementSeeking = I22 + I52 + I82 + I112 + I142 + I172 + I202 - (6 - I262) - (6 - I292),
Friendliness = I2 + I32 + I62 + I92 + I122 - (6 - I152) - (6 - I182) - (6 - I212) - (6 - I242) - (6 - I272),
Gregariousness = I7 + I37 + I67 + I97 + I127 - (6 - I157) - (6 - I187) - (6 - I217) - (6 - I247) - (6 - I277),
Imagination = I3 + I33 + I63 + I93 + I123 + I153 - (6 - I183) - (6 - I213) - (6 - I243) - (6 - I273),
Immoderation = I21 + I51 + I81 + I111 + I141 - (6 - I171) - (6 - I201) - (6 - I231) - (6 - I261) - (6 - I291),
Intellect = I23 + I53 + I83 + I113 + I143 - (6 - I173) - (6 - I203) - (6 - I233) - (6 - I263) - (6 - I293),
Liberalism = I28 + I58 + I88 - (6 - I118) - (6 - I148) - (6 - I178) - (6 - I208) - (6 - I238) - (6 - I268) - (6 - I298),
Modesty = I24 + I54 + I84 + I114 - (6 - I144) - (6 - I174) - (6 - I204) - (6 - I234) - (6 - I264) - (6 - I294),
Morality = I9 + I39 - (6 - I69) - (6 - I99) - (6 - I129) - (6 - I159) - (6 - I189) - (6 - I219) - (6 - I249) - (6 - I279),
Orderliness = I10 + I40 + I70 + I100 + I130 - (6 - I160) - (6 - I190) - (6 - I220) - (6 - I250) - (6 - I280),
SelfConsciousness = I16 + I46 + I76 + I106 + I136 + I166 - (6 - I196) - (6 - I226) - (6 - I256) - (6 - I286),
SelfDiscipline = I25 + I55 + I85 + I115 + I145 - (6 - I175) - (6 - I205) - (6 - I235) - (6 - I265) - (6 - I295),
SelfEfficacy = I5 + I35 + I65 + I95 + I125 + I155 - (6 - I185) - (6 - I215) - (6 - I245) - (6 - I275),
Sympathy = I29 + I59 + I89 + I119 - (6 - I149) - (6 - I179) - (6 - I209) - (6 - I239) - (6 - I269) - (6 - I299),
Trust = I4 + I34 + I64 + I94 + I124 + I154 - (6 - I184) - (6 - I214) - (6 - I244) - (6 - I274),
Vulnerability = I26 + I56 + I86 + I116 + I146 - (6 - I176) - (6 - I206) - (6 - I236) - (6 - I266) - (6 - I296),
Openness = Liberalism + Intellect + Adventurousness + Emotionality + ArtisticInterests + Imagination,
Neuroticism = Vulnerability + Immoderation + SelfConsciousness + Depression + Anger + Anxiety,
Extraversion = Cheerfulness + ExcitementSeeking + ActivityLevel + Assertiveness + Gregariousness + Friendliness,
Conscientiousness = Cautiousness + SelfDiscipline + AchievementStriving + Dutifulness + Orderliness + SelfEfficacy,
Agreeableness = Sympathy + Modesty + Cooperation + Altruism + Morality + Trust)
IPIPA$SEXO <- rec(IPIPA$SEX, rec = "1 = MALE; 2 = FEMALE")
IPIPA$GEI <- rec(IPIPA$COUNTRY, rec = "Albania = 0.701; Algeria = 0.632; Argentina = 0.734; Australia, Austria = 0.733; Belgium = 0.753; Brazil = 0.686; Canada, USA = 0.740; China = 0.682; Columbia = 0.725; Croatia = 0.708; Denmark = 0.767; Egypt = 0.599; Finland, Norway = 0.850; France = 0.761; Germany = 0.779; Greece = 0.685; India = 0.664; Indonesia = 0.681; Ireland = 0.807; Israel = 0.712; Italy = 0.726; Japan = 0.670; Malaysia = 0.655; Mexico = 0.699; Netherlands = 0.776; New Zealand = 0.782; Pakistan = 0.559; Philippines = 0.790; Poland = 0.715; Portugal = 0.731; Romania = 0.693; Russian Fed = 0.694; Singapore = 0.711; South Afric = 0.759; South Korea = 0.651; Spain = 0.742; Sweden = 0.823; Switzerland = 0.785; Thailand = 0.706; Turkey = 0.624; UK = 0.758; United Arab = 0.646")#Create Country-specific DFs
ALBIP <- subset(IPIPA, COUNTRY == "Albania")
ALGIP <- subset(IPIPA, COUNTRY == "Algeria")
ARGIP <- subset(IPIPA, COUNTRY == "Argentina")
AUSIP <- subset(IPIPA, COUNTRY == "Australia")
HABIP <- subset(IPIPA, COUNTRY == "Austria")
BELIP <- subset(IPIPA, COUNTRY == "Belgium")
BRAIP <- subset(IPIPA, COUNTRY == "Brazil")
CANIP <- subset(IPIPA, COUNTRY == "Canada")
CHIIP <- subset(IPIPA, COUNTRY == "China")
COLIP <- subset(IPIPA, COUNTRY == "Columbia")
CROIP <- subset(IPIPA, COUNTRY == "Croatia")
DENIP <- subset(IPIPA, COUNTRY == "Denmark")
EGYIP <- subset(IPIPA, COUNTRY == "Egypt")
FINIP <- subset(IPIPA, COUNTRY == "Finland")
FRAIP <- subset(IPIPA, COUNTRY == "France")
GERIP <- subset(IPIPA, COUNTRY == "Germany")
GREIP <- subset(IPIPA, COUNTRY == "Greece")
INDIP <- subset(IPIPA, COUNTRY == "India")
INOIP <- subset(IPIPA, COUNTRY == "Indonesia")
IREIP <- subset(IPIPA, COUNTRY == "Ireland")
ISRIP <- subset(IPIPA, COUNTRY == "Israel")
ITAIP <- subset(IPIPA, COUNTRY == "Italy")
JAPIP <- subset(IPIPA, COUNTRY == "Japan")
MALIP <- subset(IPIPA, COUNTRY == "Malaysia")
MEXIP <- subset(IPIPA, COUNTRY == "Mexico")
NETIP <- subset(IPIPA, COUNTRY == "Netherlands")
NEWIP <- subset(IPIPA, COUNTRY == "New Zealand")
NORIP <- subset(IPIPA, COUNTRY == "Norway")
PAKIP <- subset(IPIPA, COUNTRY == "Pakistan")
PHIIP <- subset(IPIPA, COUNTRY == "Philippines")
POLIP <- subset(IPIPA, COUNTRY == "Poland")
PORIP <- subset(IPIPA, COUNTRY == "Portugal")
ROMIP <- subset(IPIPA, COUNTRY == "Romania")
RUSIP <- subset(IPIPA, COUNTRY == "Russian Fed")
SINIP <- subset(IPIPA, COUNTRY == "Singapore")
SOUIP <- subset(IPIPA, COUNTRY == "South Afric")
SKOIP <- subset(IPIPA, COUNTRY == "South Korea")
SPAIP <- subset(IPIPA, COUNTRY == "Spain")
SWEIP <- subset(IPIPA, COUNTRY == "Sweden")
SWIIP <- subset(IPIPA, COUNTRY == "Switzerland")
THAIP <- subset(IPIPA, COUNTRY == "Thailand")
TURIP <- subset(IPIPA, COUNTRY == "Turkey")
UKIP <- subset(IPIPA, COUNTRY == "UK")
UAEIP <- subset(IPIPA, COUNTRY == "United Arab")
USAIP <- subset(IPIPA, COUNTRY == "USA")Analysis
Compute the Variables
Averaging is done for comparability to the unbiased d value (i.e., the one based on part-wise averaging to circumvent issues with their correlations; see below).
A reference item had to be chosen to exactly follow the method used by Nye & Drasgow (2011). To that end, I chose the variable with the closest variance between groups, since this is the least likely to be biased because it is the least likely to be impacted by different causes. The logic behind this has been explained by people like Cronbach and Lewbel and (update: Nick Huntington-Klein also noticed it!). In short, say we have a score composed of “reliable” variance and “error” or X = T + E. Error is often treated as if it is just there, not a cause of scoring in its own right, but if we had perfect knowledge of all the causes of a score, that would include an understanding of the causes for error. The residual variance contains causes, and if one group has greater residual variance with the same well-fitting model of the means and covariances, it has additional causes (or moderation in some fashion) acting on it. The smallest SD gap was for Immoderation.
The problem with reference selection via comparing indicator variances is that the sexes could differ for reasons that have nothing to do with bias. We are interested in the residual variances when it comes to the statement that there are different causes at work, and bias could reduce a real difference in the variances. I picked the variable with the closest variance difference to the mean difference. This was Vulnerability.
IPIPA %>% group_by(SEXO) %>% summarise_at(vars(Vulnerability, Immoderation, SelfConsciousness, Depression, Anger, Anxiety), list(mean = mean, sd = sd))#mean(7.417043, 7.055355, 7.552758, 9.419106, 7.770710, 7.557831) - mean(7.281363, 6.958740, 7.660362, 9.605256, 7.979225, 7.760293)
abs(0.13568 - abs(7.417043 - 7.281363))## [1] 1.94289e-16abs(0.13568 - abs(7.055355 - 6.958740))## [1] 0.039065abs(0.13568 - abs(7.552758 - 7.660362))## [1] 0.028076abs(0.13568 - abs(9.419106 - 9.605256))## [1] 0.05047abs(0.13568 - abs(7.770710 - 7.979225))## [1] 0.072835abs(0.13568 - abs(7.557831 - 7.760293))## [1] 0.066782#Single-Factor Model of Neuroticism and Sum of dMACS
NeurotModel <- '
LatentNeuroticism =~ 1*Vulnerability + Immoderation + SelfConsciousness + Depression + Anger + Anxiety'
NeurAlb <- cfa(NeurotModel, data = ALBIP, group = "SEXO"); dAlb <- dMACS_True(NeurAlb, group1 = "MALE", group2 = "FEMALE")/6; sum(dAlb)## [1] -0.2605NeurAlg <- cfa(NeurotModel, data = ALGIP, group = "SEXO"); dAlg <- dMACS_True(NeurAlg, group1 = "MALE", group2 = "FEMALE")/6; sum(dAlg)## [1] 0.189NeurArg <- cfa(NeurotModel, data = ARGIP, group = "SEXO"); dArg <- dMACS_True(NeurArg, group1 = "MALE", group2 = "FEMALE")/6; sum(dArg)## [1] 0.1858333NeurAus <- cfa(NeurotModel, data = AUSIP, group = "SEXO"); dAus <- dMACS_True(NeurAus, group1 = "MALE", group2 = "FEMALE")/6; sum(dAus)## [1] -0.3865NeurHab <- cfa(NeurotModel, data = HABIP, group = "SEXO"); dHab <- dMACS_True(NeurHab, group1 = "MALE", group2 = "FEMALE")/6; sum(dHab)## [1] -0.3345NeurBel <- cfa(NeurotModel, data = BELIP, group = "SEXO"); dBel <- dMACS_True(NeurBel, group1 = "MALE", group2 = "FEMALE")/6; sum(dBel)## [1] -0.2925NeurBra <- cfa(NeurotModel, data = BRAIP, group = "SEXO"); dBra <- dMACS_True(NeurAus, group1 = "MALE", group2 = "FEMALE")/6; sum(dBra)## [1] -0.3865NeurCan <- cfa(NeurotModel, data = CANIP, group = "SEXO"); dCan <- dMACS_True(NeurCan, group1 = "MALE", group2 = "FEMALE")/6; sum(dCan)## [1] -0.3181667NeurChi <- cfa(NeurotModel, data = CHIIP, group = "SEXO"); dChi <- dMACS_True(NeurChi, group1 = "MALE", group2 = "FEMALE")/6; sum(dChi)## [1] -0.2453333NeurCol <- cfa(NeurotModel, data = COLIP, group = "SEXO"); dCol <- dMACS_True(NeurCol, group1 = "MALE", group2 = "FEMALE")/6; sum(dCol)## [1] 0.3338333NeurCro <- cfa(NeurotModel, data = CROIP, group = "SEXO"); dCro <- dMACS_True(NeurCro, group1 = "MALE", group2 = "FEMALE")/6; sum(dCro)## [1] 0.3813333NeurDen <- cfa(NeurotModel, data = DENIP, group = "SEXO"); dDen <- dMACS_True(NeurDen, group1 = "MALE", group2 = "FEMALE")/6; sum(dDen)## [1] 0.3936667NeurEgy <- cfa(NeurotModel, data = EGYIP, group = "SEXO"); dEgy <- dMACS_True(NeurEgy, group1 = "MALE", group2 = "FEMALE")/6; sum(dEgy)## [1] 0.5061667NeurFin <- cfa(NeurotModel, data = FINIP, group = "SEXO"); dFin <- dMACS_True(NeurFin, group1 = "MALE", group2 = "FEMALE")/6; sum(dFin)## [1] -0.3306667NeurFra <- cfa(NeurotModel, data = FRAIP, group = "SEXO"); dFra <- dMACS_True(NeurFra, group1 = "MALE", group2 = "FEMALE")/6; sum(dFra)## [1] 0.4063333NeurGer <- cfa(NeurotModel, data = GERIP, group = "SEXO"); dGer <- dMACS_True(NeurGer, group1 = "MALE", group2 = "FEMALE")/6; sum(dGer)## [1] -0.2878333NeurGre <- cfa(NeurotModel, data = GREIP, group = "SEXO"); dGre <- dMACS_True(NeurGre, group1 = "MALE", group2 = "FEMALE")/6; sum(dGre)## [1] 0.2106667NeurInd <- cfa(NeurotModel, data = INDIP, group = "SEXO"); dInd <- dMACS_True(NeurInd, group1 = "MALE", group2 = "FEMALE")/6; sum(dInd)## [1] -0.357NeurIno <- cfa(NeurotModel, data = INOIP, group = "SEXO"); dIno <- dMACS_True(NeurIno, group1 = "MALE", group2 = "FEMALE")/6; sum(dIno)## [1] -0.1233333NeurIre <- cfa(NeurotModel, data = IREIP, group = "SEXO"); dIre <- dMACS_True(NeurIre, group1 = "MALE", group2 = "FEMALE")/6; sum(dIre)## [1] -0.4068333NeurIsr <- cfa(NeurotModel, data = ISRIP, group = "SEXO"); dIsr <- dMACS_True(NeurIsr, group1 = "MALE", group2 = "FEMALE")/6; sum(dIsr)## [1] 0.5388333NeurIta <- cfa(NeurotModel, data = ITAIP, group = "SEXO"); dIta <- dMACS_True(NeurIta, group1 = "MALE", group2 = "FEMALE")/6; sum(dIta)## [1] 0.4876667NeurJap <- cfa(NeurotModel, data = JAPIP, group = "SEXO"); dJap <- dMACS_True(NeurJap, group1 = "MALE", group2 = "FEMALE")/6; sum(dJap)## [1] 0.2751667NeurMal <- cfa(NeurotModel, data = MALIP, group = "SEXO"); dMal <- dMACS_True(NeurMal, group1 = "MALE", group2 = "FEMALE")/6; sum(dMal)## [1] 0.2463333NeurMex <- cfa(NeurotModel, data = MEXIP, group = "SEXO"); dMex <- dMACS_True(NeurMex, group1 = "MALE", group2 = "FEMALE")/6; sum(dMex)## [1] -0.1115NeurNet <- cfa(NeurotModel, data = NETIP, group = "SEXO"); dNet <- dMACS_True(NeurNet, group1 = "MALE", group2 = "FEMALE")/6; sum(dNet)## [1] 0.3741667NeurNew <- cfa(NeurotModel, data = NEWIP, group = "SEXO"); dNew <- dMACS_True(NeurNew, group1 = "MALE", group2 = "FEMALE")/6; sum(dNew)## [1] 0.3656667NeurNor <- cfa(NeurotModel, data = NORIP, group = "SEXO"); dNor <- dMACS_True(NeurNor, group1 = "MALE", group2 = "FEMALE")/6; sum(dNor)## [1] -0.2996667NeurPak <- cfa(NeurotModel, data = PAKIP, group = "SEXO"); dPak <- dMACS_True(NeurPak, group1 = "MALE", group2 = "FEMALE")/6; sum(dPak)## [1] -0.3146667NeurPhi <- cfa(NeurotModel, data = PHIIP, group = "SEXO"); dPhi <- dMACS_True(NeurPhi, group1 = "MALE", group2 = "FEMALE")/6; sum(dPhi)## [1] -0.2783333NeurPol <- cfa(NeurotModel, data = POLIP, group = "SEXO"); dPol <- dMACS_True(NeurPol, group1 = "MALE", group2 = "FEMALE")/6; sum(dPol)## [1] 0.1726667NeurPor <- cfa(NeurotModel, data = PORIP, group = "SEXO"); dPor <- dMACS_True(NeurPor, group1 = "MALE", group2 = "FEMALE")/6; sum(dPor)## [1] -0.1993333NeurRom <- cfa(NeurotModel, data = ROMIP, group = "SEXO"); dRom <- dMACS_True(NeurRom, group1 = "MALE", group2 = "FEMALE")/6; sum(dRom)## [1] -0.3101667NeurRus <- cfa(NeurotModel, data = RUSIP, group = "SEXO"); dRus <- dMACS_True(NeurRus, group1 = "MALE", group2 = "FEMALE")/6; sum(dRus)## [1] 0.2561667NeurSin <- cfa(NeurotModel, data = SINIP, group = "SEXO"); dSin <- dMACS_True(NeurSin, group1 = "MALE", group2 = "FEMALE")/6; sum(dSin)## [1] -0.3403333NeurSou <- cfa(NeurotModel, data = SOUIP, group = "SEXO"); dSou <- dMACS_True(NeurSou, group1 = "MALE", group2 = "FEMALE")/6; sum(dSou)## [1] -0.3425NeurSKo <- cfa(NeurotModel, data = SKOIP, group = "SEXO"); dSKo <- dMACS_True(NeurSKo, group1 = "MALE", group2 = "FEMALE")/6; sum(dSKo)## [1] 0.2763333NeurSpa <- cfa(NeurotModel, data = SPAIP, group = "SEXO"); dSpa <- dMACS_True(NeurSpa, group1 = "MALE", group2 = "FEMALE")/6; sum(dSpa)## [1] -0.34NeurSwe <- cfa(NeurotModel, data = SWEIP, group = "SEXO"); dSwe <- dMACS_True(NeurSwe, group1 = "MALE", group2 = "FEMALE")/6; sum(dSwe)## [1] 0.3095NeurSwi <- cfa(NeurotModel, data = SWIIP, group = "SEXO"); dSwi <- dMACS_True(NeurSwi, group1 = "MALE", group2 = "FEMALE")/6; sum(dSwi)## [1] -0.529NeurTha <- cfa(NeurotModel, data = THAIP, group = "SEXO"); dTha <- dMACS_True(NeurTha, group1 = "MALE", group2 = "FEMALE")/6; sum(dTha)## [1] 0.05033333NeurTur <- cfa(NeurotModel, data = TURIP, group = "SEXO"); dTur <- dMACS_True(NeurTur, group1 = "MALE", group2 = "FEMALE")/6; sum(dTur)## [1] -0.1413333NeurUK <- cfa(NeurotModel, data = UKIP, group = "SEXO"); dUK <- dMACS_True(NeurUK, group1 = "MALE", group2 = "FEMALE")/6; sum(dUK)## [1] 0.3188333NeurUAE <- cfa(NeurotModel, data = UAEIP, group = "SEXO"); dUAE <- dMACS_True(NeurUAE, group1 = "MALE", group2 = "FEMALE")/6; sum(dUAE)## [1] -0.6493333NeurUSA <- cfa(NeurotModel, data = USAIP, group = "SEXO"); dUSA <- dMACS_True(NeurUSA, group1 = "MALE", group2 = "FEMALE")/6; sum(dUSA)## [1] 0.3023333As should be done due to the distortion in sumscores as a result of the imperfect correlations of their parts (i.e., the composite score extremity effect), we can compute and compare the sum of d’s for facets and the d for the sumscore. The former is the desired quantity until we have an actual aggregated dMACS. We could figure it out with the loadings, but it’s needlessly time-consuming when everything should be approximately equal.
#d for Sumscore
dOAlb = cohen.d(ALBIP$Neuroticism, ALBIP$SEX); dOAlb$cohen.d[,2]## effect
## 0.3104432dOAlg = cohen.d(ALGIP$Neuroticism, ALGIP$SEX); dOAlg$cohen.d[,2]## effect
## 0.2048881dOArg = cohen.d(ARGIP$Neuroticism, ARGIP$SEX); dOArg$cohen.d[,2]## effect
## 0.2497773dOAus = cohen.d(AUSIP$Neuroticism, AUSIP$SEX); dOAus$cohen.d[,2]## effect
## 0.4841326dOHab = cohen.d(HABIP$Neuroticism, HABIP$SEX); dOHab$cohen.d[,2]## effect
## 0.3710997dOBel = cohen.d(BELIP$Neuroticism, BELIP$SEX); dOBel$cohen.d[,2]## effect
## 0.3196483dOBra = cohen.d(BRAIP$Neuroticism, BRAIP$SEX); dOBra$cohen.d[,2]## effect
## 0.3714844dOCan = cohen.d(CANIP$Neuroticism, CANIP$SEX); dOCan$cohen.d[,2]## effect
## 0.4035774dOChi = cohen.d(CHIIP$Neuroticism, CHIIP$SEX); dOChi$cohen.d[,2]## effect
## 0.3071515dOCol = cohen.d(COLIP$Neuroticism, COLIP$SEX); dOCol$cohen.d[,2]## effect
## 0.3449715dOCro = cohen.d(CROIP$Neuroticism, CROIP$SEX); dOCro$cohen.d[,2]## effect
## 0.4613239dODen = cohen.d(DENIP$Neuroticism, DENIP$SEX); dODen$cohen.d[,2]## effect
## 0.5104887dOEgy = cohen.d(EGYIP$Neuroticism, EGYIP$SEX); dOEgy$cohen.d[,2]## effect
## 0.632099dOFin = cohen.d(FINIP$Neuroticism, FINIP$SEX); dOFin$cohen.d[,2]## effect
## 0.415431dOFra = cohen.d(FRAIP$Neuroticism, FRAIP$SEX); dOFra$cohen.d[,2]## effect
## 0.5181189dOGer = cohen.d(GERIP$Neuroticism, GERIP$SEX); dOGer$cohen.d[,2]## effect
## 0.348082dOGre = cohen.d(GREIP$Neuroticism, GREIP$SEX); dOGre$cohen.d[,2]## effect
## 0.2533055dOInd = cohen.d(INDIP$Neuroticism, INDIP$SEX); dOInd$cohen.d[,2]## effect
## 0.456674dOIno = cohen.d(INOIP$Neuroticism, INOIP$SEX); dOIno$cohen.d[,2]## effect
## 0.1149468dOIre = cohen.d(IREIP$Neuroticism, IREIP$SEX); dOIre$cohen.d[,2]## effect
## 0.5070944dOIsr = cohen.d(ISRIP$Neuroticism, ISRIP$SEX); dOIsr$cohen.d[,2]## effect
## 0.6722948dOIta = cohen.d(ITAIP$Neuroticism, ITAIP$SEX); dOIta$cohen.d[,2]## effect
## 0.6069695dOJap = cohen.d(JAPIP$Neuroticism, JAPIP$SEX); dOJap$cohen.d[,2]## effect
## 0.333087dOMal = cohen.d(MALIP$Neuroticism, MALIP$SEX); dOMal$cohen.d[,2]## effect
## 0.3330984dOMex = cohen.d(MEXIP$Neuroticism, MEXIP$SEX); dOMex$cohen.d[,2]## effect
## 0.1634371dONet = cohen.d(NETIP$Neuroticism, NETIP$SEX); dONet$cohen.d[,2]## effect
## 0.5011561dONew = cohen.d(NEWIP$Neuroticism, NEWIP$SEX); dONew$cohen.d[,2]## effect
## 0.4721986dONor = cohen.d(NORIP$Neuroticism, NORIP$SEX); dONor$cohen.d[,2]## effect
## 0.3629768dOPak = cohen.d(PAKIP$Neuroticism, PAKIP$SEX); dOPak$cohen.d[,2]## effect
## 0.3726361dOPhi = cohen.d(PHIIP$Neuroticism, PHIIP$SEX); dOPhi$cohen.d[,2]## effect
## 0.3642182dOPol = cohen.d(POLIP$Neuroticism, POLIP$SEX); dOPol$cohen.d[,2]## effect
## 0.2228362dOPor = cohen.d(PORIP$Neuroticism, PORIP$SEX); dOPor$cohen.d[,2]## effect
## 0.2269227dORom = cohen.d(ROMIP$Neuroticism, ROMIP$SEX); dORom$cohen.d[,2]## effect
## 0.3822293dORus = cohen.d(RUSIP$Neuroticism, RUSIP$SEX); dORus$cohen.d[,2]## effect
## 0.3126842dOSin = cohen.d(SINIP$Neuroticism, SINIP$SEX); dOSin$cohen.d[,2]## effect
## 0.4418945dOSou = cohen.d(SOUIP$Neuroticism, SOUIP$SEX); dOSou$cohen.d[,2]## effect
## 0.4364684dOSKo = cohen.d(SKOIP$Neuroticism, SKOIP$SEX); dOSKo$cohen.d[,2]## effect
## 0.3582619dOSpa = cohen.d(SPAIP$Neuroticism, SPAIP$SEX); dOSpa$cohen.d[,2]## effect
## 0.4074406dOSwe = cohen.d(SWEIP$Neuroticism, SWEIP$SEX); dOSwe$cohen.d[,2]## effect
## 0.3883253dOSwi = cohen.d(SWIIP$Neuroticism, SWIIP$SEX); dOSwi$cohen.d[,2]## effect
## 0.650499dOTha = cohen.d(THAIP$Neuroticism, THAIP$SEX); dOTha$cohen.d[,2]## effect
## 0.04522867dOTur = cohen.d(TURIP$Neuroticism, TURIP$SEX); dOTur$cohen.d[,2]## effect
## 0.1520373dOUK = cohen.d(UKIP$Neuroticism, UKIP$SEX); dOUK$cohen.d[,2]## effect
## 0.408826dOUAE = cohen.d(UAEIP$Neuroticism, UAEIP$SEX); dOUAE$cohen.d[,2]## effect
## 0.7909179dOUSA = cohen.d(USAIP$Neuroticism, USAIP$SEX); dOUSA$cohen.d[,2]## effect
## 0.4178913#d for Parts of Sumscores
dOPAlb = (1/6)*(cohen.d(ALBIP$Vulnerability, ALBIP$SEX)$cohen.d[,2] + cohen.d(ALBIP$Immoderation, ALBIP$SEX)$cohen.d[,2] + cohen.d(ALBIP$SelfConsciousness, ALBIP$SEX)$cohen.d[,2] + cohen.d(ALBIP$Depression, ALBIP$SEX)$cohen.d[,2] + cohen.d(ALBIP$Anger, ALBIP$SEX)$cohen.d[,2] + cohen.d(ALBIP$Anxiety, ALBIP$SEX)$cohen.d[,2]); dOPAlb## effect
## 0.2435067dOPAlg = (1/6)*(cohen.d(ALGIP$Vulnerability, ALGIP$SEX)$cohen.d[,2] + cohen.d(ALGIP$Immoderation, ALGIP$SEX)$cohen.d[,2] + cohen.d(ALGIP$SelfConsciousness, ALGIP$SEX)$cohen.d[,2] + cohen.d(ALGIP$Depression, ALGIP$SEX)$cohen.d[,2] + cohen.d(ALGIP$Anger, ALGIP$SEX)$cohen.d[,2] + cohen.d(ALGIP$Anxiety, ALGIP$SEX)$cohen.d[,2]); dOPAlg## effect
## 0.1643438dOPArg = (1/6)*(cohen.d(ARGIP$Vulnerability, ARGIP$SEX)$cohen.d[,2] + cohen.d(ARGIP$Immoderation, ARGIP$SEX)$cohen.d[,2] + cohen.d(ARGIP$SelfConsciousness, ARGIP$SEX)$cohen.d[,2] + cohen.d(ARGIP$Depression, ARGIP$SEX)$cohen.d[,2] + cohen.d(ARGIP$Anger, ARGIP$SEX)$cohen.d[,2] + cohen.d(ARGIP$Anxiety, ARGIP$SEX)$cohen.d[,2]); dOPArg## effect
## 0.1978211dOPAus = (1/6)*(cohen.d(AUSIP$Vulnerability, AUSIP$SEX)$cohen.d[,2] + cohen.d(AUSIP$Immoderation, AUSIP$SEX)$cohen.d[,2] + cohen.d(AUSIP$SelfConsciousness, AUSIP$SEX)$cohen.d[,2] + cohen.d(AUSIP$Depression, AUSIP$SEX)$cohen.d[,2] + cohen.d(AUSIP$Anger, AUSIP$SEX)$cohen.d[,2] + cohen.d(AUSIP$Anxiety, AUSIP$SEX)$cohen.d[,2]); dOPAus## effect
## 0.3820402dOPHab = (1/6)*(cohen.d(HABIP$Vulnerability, HABIP$SEX)$cohen.d[,2] + cohen.d(HABIP$Immoderation, HABIP$SEX)$cohen.d[,2] + cohen.d(HABIP$SelfConsciousness, HABIP$SEX)$cohen.d[,2] + cohen.d(HABIP$Depression, HABIP$SEX)$cohen.d[,2] + cohen.d(HABIP$Anger, HABIP$SEX)$cohen.d[,2] + cohen.d(HABIP$Anxiety, HABIP$SEX)$cohen.d[,2]); dOPHab## effect
## 0.3094754dOPBel = (1/6)*(cohen.d(BELIP$Vulnerability, BELIP$SEX)$cohen.d[,2] + cohen.d(BELIP$Immoderation, BELIP$SEX)$cohen.d[,2] + cohen.d(BELIP$SelfConsciousness, BELIP$SEX)$cohen.d[,2] + cohen.d(BELIP$Depression, BELIP$SEX)$cohen.d[,2] + cohen.d(BELIP$Anger, BELIP$SEX)$cohen.d[,2] + cohen.d(BELIP$Anxiety, BELIP$SEX)$cohen.d[,2]); dOPBel## effect
## 0.2618641dOPBra = (1/6)*(cohen.d(BRAIP$Vulnerability, BRAIP$SEX)$cohen.d[,2] + cohen.d(BRAIP$Immoderation, BRAIP$SEX)$cohen.d[,2] + cohen.d(BRAIP$SelfConsciousness, BRAIP$SEX)$cohen.d[,2] + cohen.d(BRAIP$Depression, BRAIP$SEX)$cohen.d[,2] + cohen.d(BRAIP$Anger, BRAIP$SEX)$cohen.d[,2] + cohen.d(BRAIP$Anxiety, BRAIP$SEX)$cohen.d[,2]); dOPBra## effect
## 0.31262dOPCan = (1/6)*(cohen.d(CANIP$Vulnerability, CANIP$SEX)$cohen.d[,2] + cohen.d(CANIP$Immoderation, CANIP$SEX)$cohen.d[,2] + cohen.d(CANIP$SelfConsciousness, CANIP$SEX)$cohen.d[,2] + cohen.d(CANIP$Depression, CANIP$SEX)$cohen.d[,2] + cohen.d(CANIP$Anger, CANIP$SEX)$cohen.d[,2] + cohen.d(CANIP$Anxiety, CANIP$SEX)$cohen.d[,2]); dOPCan## effect
## 0.3153231dOPChi = (1/6)*(cohen.d(CHIIP$Vulnerability, CHIIP$SEX)$cohen.d[,2] + cohen.d(CHIIP$Immoderation, CHIIP$SEX)$cohen.d[,2] + cohen.d(CHIIP$SelfConsciousness, CHIIP$SEX)$cohen.d[,2] + cohen.d(CHIIP$Depression, CHIIP$SEX)$cohen.d[,2] + cohen.d(CHIIP$Anger, CHIIP$SEX)$cohen.d[,2] + cohen.d(CHIIP$Anxiety, CHIIP$SEX)$cohen.d[,2]); dOPChi## effect
## 0.2357846dOPCol = (1/6)*(cohen.d(COLIP$Vulnerability, COLIP$SEX)$cohen.d[,2] + cohen.d(COLIP$Immoderation, COLIP$SEX)$cohen.d[,2] + cohen.d(COLIP$SelfConsciousness, COLIP$SEX)$cohen.d[,2] + cohen.d(COLIP$Depression, COLIP$SEX)$cohen.d[,2] + cohen.d(COLIP$Anger, COLIP$SEX)$cohen.d[,2] + cohen.d(COLIP$Anxiety, COLIP$SEX)$cohen.d[,2]); dOPCol## effect
## 0.2788671dOPCro = (1/6)*(cohen.d(CROIP$Vulnerability, CROIP$SEX)$cohen.d[,2] + cohen.d(CROIP$Immoderation, CROIP$SEX)$cohen.d[,2] + cohen.d(CROIP$SelfConsciousness, CROIP$SEX)$cohen.d[,2] + cohen.d(CROIP$Depression, CROIP$SEX)$cohen.d[,2] + cohen.d(CROIP$Anger, CROIP$SEX)$cohen.d[,2] + cohen.d(CROIP$Anxiety, CROIP$SEX)$cohen.d[,2]); dOPCro## effect
## 0.369925dOPDen = (1/6)*(cohen.d(DENIP$Vulnerability, DENIP$SEX)$cohen.d[,2] + cohen.d(DENIP$Immoderation, DENIP$SEX)$cohen.d[,2] + cohen.d(DENIP$SelfConsciousness, DENIP$SEX)$cohen.d[,2] + cohen.d(DENIP$Depression, DENIP$SEX)$cohen.d[,2] + cohen.d(DENIP$Anger, DENIP$SEX)$cohen.d[,2] + cohen.d(DENIP$Anxiety, DENIP$SEX)$cohen.d[,2]); dOPDen## effect
## 0.4013048dOPEgy = (1/6)*(cohen.d(EGYIP$Vulnerability, EGYIP$SEX)$cohen.d[,2] + cohen.d(EGYIP$Immoderation, EGYIP$SEX)$cohen.d[,2] + cohen.d(EGYIP$SelfConsciousness, EGYIP$SEX)$cohen.d[,2] + cohen.d(EGYIP$Depression, EGYIP$SEX)$cohen.d[,2] + cohen.d(EGYIP$Anger, EGYIP$SEX)$cohen.d[,2] + cohen.d(EGYIP$Anxiety, EGYIP$SEX)$cohen.d[,2]); dOPEgy## effect
## 0.4913482dOPFin = (1/6)*(cohen.d(FINIP$Vulnerability, FINIP$SEX)$cohen.d[,2] + cohen.d(FINIP$Immoderation, FINIP$SEX)$cohen.d[,2] + cohen.d(FINIP$SelfConsciousness, FINIP$SEX)$cohen.d[,2] + cohen.d(FINIP$Depression, FINIP$SEX)$cohen.d[,2] + cohen.d(FINIP$Anger, FINIP$SEX)$cohen.d[,2] + cohen.d(FINIP$Anxiety, FINIP$SEX)$cohen.d[,2]); dOPFin## effect
## 0.3260756dOPFra = (1/6)*(cohen.d(FRAIP$Vulnerability, FRAIP$SEX)$cohen.d[,2] + cohen.d(FRAIP$Immoderation, FRAIP$SEX)$cohen.d[,2] + cohen.d(FRAIP$SelfConsciousness, FRAIP$SEX)$cohen.d[,2] + cohen.d(FRAIP$Depression, FRAIP$SEX)$cohen.d[,2] + cohen.d(FRAIP$Anger, FRAIP$SEX)$cohen.d[,2] + cohen.d(FRAIP$Anxiety, FRAIP$SEX)$cohen.d[,2]); dOPFra## effect
## 0.4018442dOPGer = (1/6)*(cohen.d(GERIP$Vulnerability, GERIP$SEX)$cohen.d[,2] + cohen.d(GERIP$Immoderation, GERIP$SEX)$cohen.d[,2] + cohen.d(GERIP$SelfConsciousness, GERIP$SEX)$cohen.d[,2] + cohen.d(GERIP$Depression, GERIP$SEX)$cohen.d[,2] + cohen.d(GERIP$Anger, GERIP$SEX)$cohen.d[,2] + cohen.d(GERIP$Anxiety, GERIP$SEX)$cohen.d[,2]); dOPGer## effect
## 0.2797789dOPGre = (1/6)*(cohen.d(GREIP$Vulnerability, GREIP$SEX)$cohen.d[,2] + cohen.d(GREIP$Immoderation, GREIP$SEX)$cohen.d[,2] + cohen.d(GREIP$SelfConsciousness, GREIP$SEX)$cohen.d[,2] + cohen.d(GREIP$Depression, GREIP$SEX)$cohen.d[,2] + cohen.d(GREIP$Anger, GREIP$SEX)$cohen.d[,2] + cohen.d(GREIP$Anxiety, GREIP$SEX)$cohen.d[,2]); dOPGre## effect
## 0.2011276dOPInd = (1/6)*(cohen.d(INDIP$Vulnerability, INDIP$SEX)$cohen.d[,2] + cohen.d(INDIP$Immoderation, INDIP$SEX)$cohen.d[,2] + cohen.d(INDIP$SelfConsciousness, INDIP$SEX)$cohen.d[,2] + cohen.d(INDIP$Depression, INDIP$SEX)$cohen.d[,2] + cohen.d(INDIP$Anger, INDIP$SEX)$cohen.d[,2] + cohen.d(INDIP$Anxiety, INDIP$SEX)$cohen.d[,2]); dOPInd## effect
## 0.3551415dOPIno = (1/6)*(cohen.d(INOIP$Vulnerability, INOIP$SEX)$cohen.d[,2] + cohen.d(INOIP$Immoderation, INOIP$SEX)$cohen.d[,2] + cohen.d(INOIP$SelfConsciousness, INOIP$SEX)$cohen.d[,2] + cohen.d(INOIP$Depression, INOIP$SEX)$cohen.d[,2] + cohen.d(INOIP$Anger, INOIP$SEX)$cohen.d[,2] + cohen.d(INOIP$Anxiety, INOIP$SEX)$cohen.d[,2]); dOPIno## effect
## 0.09796545dOPIre = (1/6)*(cohen.d(IREIP$Vulnerability, IREIP$SEX)$cohen.d[,2] + cohen.d(IREIP$Immoderation, IREIP$SEX)$cohen.d[,2] + cohen.d(IREIP$SelfConsciousness, IREIP$SEX)$cohen.d[,2] + cohen.d(IREIP$Depression, IREIP$SEX)$cohen.d[,2] + cohen.d(IREIP$Anger, IREIP$SEX)$cohen.d[,2] + cohen.d(IREIP$Anxiety, IREIP$SEX)$cohen.d[,2]); dOPIre## effect
## 0.4007762dOPIsr = (1/6)*(cohen.d(ISRIP$Vulnerability, ISRIP$SEX)$cohen.d[,2] + cohen.d(ISRIP$Immoderation, ISRIP$SEX)$cohen.d[,2] + cohen.d(ISRIP$SelfConsciousness, ISRIP$SEX)$cohen.d[,2] + cohen.d(ISRIP$Depression, ISRIP$SEX)$cohen.d[,2] + cohen.d(ISRIP$Anger, ISRIP$SEX)$cohen.d[,2] + cohen.d(ISRIP$Anxiety, ISRIP$SEX)$cohen.d[,2]); dOPIsr## effect
## 0.5344007dOPIta = (1/6)*(cohen.d(ITAIP$Vulnerability, ITAIP$SEX)$cohen.d[,2] + cohen.d(ITAIP$Immoderation, ITAIP$SEX)$cohen.d[,2] + cohen.d(ITAIP$SelfConsciousness, ITAIP$SEX)$cohen.d[,2] + cohen.d(ITAIP$Depression, ITAIP$SEX)$cohen.d[,2] + cohen.d(ITAIP$Anger, ITAIP$SEX)$cohen.d[,2] + cohen.d(ITAIP$Anxiety, ITAIP$SEX)$cohen.d[,2]); dOPIta## effect
## 0.4798951dOPJap = (1/6)*(cohen.d(JAPIP$Vulnerability, JAPIP$SEX)$cohen.d[,2] + cohen.d(JAPIP$Immoderation, JAPIP$SEX)$cohen.d[,2] + cohen.d(JAPIP$SelfConsciousness, JAPIP$SEX)$cohen.d[,2] + cohen.d(JAPIP$Depression, JAPIP$SEX)$cohen.d[,2] + cohen.d(JAPIP$Anger, JAPIP$SEX)$cohen.d[,2] + cohen.d(JAPIP$Anxiety, JAPIP$SEX)$cohen.d[,2]); dOPJap## effect
## 0.2621303dOPMal = (1/6)*(cohen.d(MALIP$Vulnerability, MALIP$SEX)$cohen.d[,2] + cohen.d(MALIP$Immoderation, MALIP$SEX)$cohen.d[,2] + cohen.d(MALIP$SelfConsciousness, MALIP$SEX)$cohen.d[,2] + cohen.d(MALIP$Depression, MALIP$SEX)$cohen.d[,2] + cohen.d(MALIP$Anger, MALIP$SEX)$cohen.d[,2] + cohen.d(MALIP$Anxiety, MALIP$SEX)$cohen.d[,2]); dOPMal## effect
## 0.2536049dOPMex = (1/6)*(cohen.d(MEXIP$Vulnerability, MEXIP$SEX)$cohen.d[,2] + cohen.d(MEXIP$Immoderation, MEXIP$SEX)$cohen.d[,2] + cohen.d(MEXIP$SelfConsciousness, MEXIP$SEX)$cohen.d[,2] + cohen.d(MEXIP$Depression, MEXIP$SEX)$cohen.d[,2] + cohen.d(MEXIP$Anger, MEXIP$SEX)$cohen.d[,2] + cohen.d(MEXIP$Anxiety, MEXIP$SEX)$cohen.d[,2]); dOPMex## effect
## 0.1252991dOPNet = (1/6)*(cohen.d(NETIP$Vulnerability, NETIP$SEX)$cohen.d[,2] + cohen.d(NETIP$Immoderation, NETIP$SEX)$cohen.d[,2] + cohen.d(NETIP$SelfConsciousness, NETIP$SEX)$cohen.d[,2] + cohen.d(NETIP$Depression, NETIP$SEX)$cohen.d[,2] + cohen.d(NETIP$Anger, NETIP$SEX)$cohen.d[,2] + cohen.d(NETIP$Anxiety, NETIP$SEX)$cohen.d[,2]); dOPNet## effect
## 0.3906305dOPNew = (1/6)*(cohen.d(NEWIP$Vulnerability, NEWIP$SEX)$cohen.d[,2] + cohen.d(NEWIP$Immoderation, NEWIP$SEX)$cohen.d[,2] + cohen.d(NEWIP$SelfConsciousness, NEWIP$SEX)$cohen.d[,2] + cohen.d(NEWIP$Depression, NEWIP$SEX)$cohen.d[,2] + cohen.d(NEWIP$Anger, NEWIP$SEX)$cohen.d[,2] + cohen.d(NEWIP$Anxiety, NEWIP$SEX)$cohen.d[,2]); dOPNew## effect
## 0.3672418dOPNor = (1/6)*(cohen.d(NORIP$Vulnerability, NORIP$SEX)$cohen.d[,2] + cohen.d(NORIP$Immoderation, NORIP$SEX)$cohen.d[,2] + cohen.d(NORIP$SelfConsciousness, NORIP$SEX)$cohen.d[,2] + cohen.d(NORIP$Depression, NORIP$SEX)$cohen.d[,2] + cohen.d(NORIP$Anger, NORIP$SEX)$cohen.d[,2] + cohen.d(NORIP$Anxiety, NORIP$SEX)$cohen.d[,2]); dOPNor## effect
## 0.2912541dOPPak = (1/6)*(cohen.d(PAKIP$Vulnerability, PAKIP$SEX)$cohen.d[,2] + cohen.d(PAKIP$Immoderation, PAKIP$SEX)$cohen.d[,2] + cohen.d(PAKIP$SelfConsciousness, PAKIP$SEX)$cohen.d[,2] + cohen.d(PAKIP$Depression, PAKIP$SEX)$cohen.d[,2] + cohen.d(PAKIP$Anger, PAKIP$SEX)$cohen.d[,2] + cohen.d(PAKIP$Anxiety, PAKIP$SEX)$cohen.d[,2]); dOPPak## effect
## 0.2888258dOPPhi = (1/6)*(cohen.d(PHIIP$Vulnerability, PHIIP$SEX)$cohen.d[,2] + cohen.d(PHIIP$Immoderation, PHIIP$SEX)$cohen.d[,2] + cohen.d(PHIIP$SelfConsciousness, PHIIP$SEX)$cohen.d[,2] + cohen.d(PHIIP$Depression, PHIIP$SEX)$cohen.d[,2] + cohen.d(PHIIP$Anger, PHIIP$SEX)$cohen.d[,2] + cohen.d(PHIIP$Anxiety, PHIIP$SEX)$cohen.d[,2]); dOPPhi## effect
## 0.276129dOPPol = (1/6)*(cohen.d(POLIP$Vulnerability, POLIP$SEX)$cohen.d[,2] + cohen.d(POLIP$Immoderation, POLIP$SEX)$cohen.d[,2] + cohen.d(POLIP$SelfConsciousness, POLIP$SEX)$cohen.d[,2] + cohen.d(POLIP$Depression, POLIP$SEX)$cohen.d[,2] + cohen.d(POLIP$Anger, POLIP$SEX)$cohen.d[,2] + cohen.d(POLIP$Anxiety, POLIP$SEX)$cohen.d[,2]); dOPPol## effect
## 0.1811741dOPPor = (1/6)*(cohen.d(PORIP$Vulnerability, PORIP$SEX)$cohen.d[,2] + cohen.d(PORIP$Immoderation, PORIP$SEX)$cohen.d[,2] + cohen.d(PORIP$SelfConsciousness, PORIP$SEX)$cohen.d[,2] + cohen.d(PORIP$Depression, PORIP$SEX)$cohen.d[,2] + cohen.d(PORIP$Anger, PORIP$SEX)$cohen.d[,2] + cohen.d(PORIP$Anxiety, PORIP$SEX)$cohen.d[,2]); dOPPor## effect
## 0.1819325dOPRom = (1/6)*(cohen.d(ROMIP$Vulnerability, ROMIP$SEX)$cohen.d[,2] + cohen.d(ROMIP$Immoderation, ROMIP$SEX)$cohen.d[,2] + cohen.d(ROMIP$SelfConsciousness, ROMIP$SEX)$cohen.d[,2] + cohen.d(ROMIP$Depression, ROMIP$SEX)$cohen.d[,2] + cohen.d(ROMIP$Anger, ROMIP$SEX)$cohen.d[,2] + cohen.d(ROMIP$Anxiety, ROMIP$SEX)$cohen.d[,2]); dOPRom## effect
## 0.3062691dOPRus = (1/6)*(cohen.d(RUSIP$Vulnerability, RUSIP$SEX)$cohen.d[,2] + cohen.d(RUSIP$Immoderation, RUSIP$SEX)$cohen.d[,2] + cohen.d(RUSIP$SelfConsciousness, RUSIP$SEX)$cohen.d[,2] + cohen.d(RUSIP$Depression, RUSIP$SEX)$cohen.d[,2] + cohen.d(RUSIP$Anger, RUSIP$SEX)$cohen.d[,2] + cohen.d(RUSIP$Anxiety, RUSIP$SEX)$cohen.d[,2]); dOPRus## effect
## 0.2643918dOPSin = (1/6)*(cohen.d(SINIP$Vulnerability, SINIP$SEX)$cohen.d[,2] + cohen.d(SINIP$Immoderation, SINIP$SEX)$cohen.d[,2] + cohen.d(SINIP$SelfConsciousness, SINIP$SEX)$cohen.d[,2] + cohen.d(SINIP$Depression, SINIP$SEX)$cohen.d[,2] + cohen.d(SINIP$Anger, SINIP$SEX)$cohen.d[,2] + cohen.d(SINIP$Anxiety, SINIP$SEX)$cohen.d[,2]); dOPSin## effect
## 0.3376697dOPSou = (1/6)*(cohen.d(SOUIP$Vulnerability, SOUIP$SEX)$cohen.d[,2] + cohen.d(SOUIP$Immoderation, SOUIP$SEX)$cohen.d[,2] + cohen.d(SOUIP$SelfConsciousness, SOUIP$SEX)$cohen.d[,2] + cohen.d(SOUIP$Depression, SOUIP$SEX)$cohen.d[,2] + cohen.d(SOUIP$Anger, SOUIP$SEX)$cohen.d[,2] + cohen.d(SOUIP$Anxiety, SOUIP$SEX)$cohen.d[,2]); dOPSou## effect
## 0.3381776dOPSKo = (1/6)*(cohen.d(SKOIP$Vulnerability, SKOIP$SEX)$cohen.d[,2] + cohen.d(SKOIP$Immoderation, SKOIP$SEX)$cohen.d[,2] + cohen.d(SKOIP$SelfConsciousness, SKOIP$SEX)$cohen.d[,2] + cohen.d(SKOIP$Depression, SKOIP$SEX)$cohen.d[,2] + cohen.d(SKOIP$Anger, SKOIP$SEX)$cohen.d[,2] + cohen.d(SKOIP$Anxiety, SKOIP$SEX)$cohen.d[,2]); dOPSKo## effect
## 0.2618332dOPSpa = (1/6)*(cohen.d(SPAIP$Vulnerability, SPAIP$SEX)$cohen.d[,2] + cohen.d(SPAIP$Immoderation, SPAIP$SEX)$cohen.d[,2] + cohen.d(SPAIP$SelfConsciousness, SPAIP$SEX)$cohen.d[,2] + cohen.d(SPAIP$Depression, SPAIP$SEX)$cohen.d[,2] + cohen.d(SPAIP$Anger, SPAIP$SEX)$cohen.d[,2] + cohen.d(SPAIP$Anxiety, SPAIP$SEX)$cohen.d[,2]); dOPSpa## effect
## 0.3229537dOPSwe = (1/6)*(cohen.d(SWEIP$Vulnerability, SWEIP$SEX)$cohen.d[,2] + cohen.d(SWEIP$Immoderation, SWEIP$SEX)$cohen.d[,2] + cohen.d(SWEIP$SelfConsciousness, SWEIP$SEX)$cohen.d[,2] + cohen.d(SWEIP$Depression, SWEIP$SEX)$cohen.d[,2] + cohen.d(SWEIP$Anger, SWEIP$SEX)$cohen.d[,2] + cohen.d(SWEIP$Anxiety, SWEIP$SEX)$cohen.d[,2]); dOPSwe## effect
## 0.3079226dOPSwi = (1/6)*(cohen.d(SWIIP$Vulnerability, SWIIP$SEX)$cohen.d[,2] + cohen.d(SWIIP$Immoderation, SWIIP$SEX)$cohen.d[,2] + cohen.d(SWIIP$SelfConsciousness, SWIIP$SEX)$cohen.d[,2] + cohen.d(SWIIP$Depression, SWIIP$SEX)$cohen.d[,2] + cohen.d(SWIIP$Anger, SWIIP$SEX)$cohen.d[,2] + cohen.d(SWIIP$Anxiety, SWIIP$SEX)$cohen.d[,2]); dOPSwi## effect
## 0.5228171dOPTha = (1/6)*(cohen.d(THAIP$Vulnerability, THAIP$SEX)$cohen.d[,2] + cohen.d(THAIP$Immoderation, THAIP$SEX)$cohen.d[,2] + cohen.d(THAIP$SelfConsciousness, THAIP$SEX)$cohen.d[,2] + cohen.d(THAIP$Depression, THAIP$SEX)$cohen.d[,2] + cohen.d(THAIP$Anger, THAIP$SEX)$cohen.d[,2] + cohen.d(THAIP$Anxiety, THAIP$SEX)$cohen.d[,2]); dOPTha## effect
## 0.0339517dOPTur = (1/6)*(cohen.d(TURIP$Vulnerability, TURIP$SEX)$cohen.d[,2] + cohen.d(TURIP$Immoderation, TURIP$SEX)$cohen.d[,2] + cohen.d(TURIP$SelfConsciousness, TURIP$SEX)$cohen.d[,2] + cohen.d(TURIP$Depression, TURIP$SEX)$cohen.d[,2] + cohen.d(TURIP$Anger, TURIP$SEX)$cohen.d[,2] + cohen.d(TURIP$Anxiety, TURIP$SEX)$cohen.d[,2]); dOPTur## effect
## 0.1213023dOPUK = (1/6)*(cohen.d(UKIP$Vulnerability, UKIP$SEX)$cohen.d[,2] + cohen.d(UKIP$Immoderation, UKIP$SEX)$cohen.d[,2] + cohen.d(UKIP$SelfConsciousness, UKIP$SEX)$cohen.d[,2] + cohen.d(UKIP$Depression, UKIP$SEX)$cohen.d[,2] + cohen.d(UKIP$Anger, UKIP$SEX)$cohen.d[,2] + cohen.d(UKIP$Anxiety, UKIP$SEX)$cohen.d[,2]); dOPUK## effect
## 0.3188531dOPUAE = (1/6)*(cohen.d(UAEIP$Vulnerability, UAEIP$SEX)$cohen.d[,2] + cohen.d(UAEIP$Immoderation, UAEIP$SEX)$cohen.d[,2] + cohen.d(UAEIP$SelfConsciousness, UAEIP$SEX)$cohen.d[,2] + cohen.d(UAEIP$Depression, UAEIP$SEX)$cohen.d[,2] + cohen.d(UAEIP$Anger, UAEIP$SEX)$cohen.d[,2] + cohen.d(UAEIP$Anxiety, UAEIP$SEX)$cohen.d[,2]); dOPUAE## effect
## 0.6239353dOPUSA = (1/6)*(cohen.d(USAIP$Vulnerability, USAIP$SEX)$cohen.d[,2] + cohen.d(USAIP$Immoderation, USAIP$SEX)$cohen.d[,2] + cohen.d(USAIP$SelfConsciousness, USAIP$SEX)$cohen.d[,2] + cohen.d(USAIP$Depression, USAIP$SEX)$cohen.d[,2] + cohen.d(USAIP$Anger, USAIP$SEX)$cohen.d[,2] + cohen.d(USAIP$Anxiety, USAIP$SEX)$cohen.d[,2]); dOPUSA## effect
## 0.3275854NeuCou <- data.frame("Country" = c("Albania", "Algeria", "Argentina", "Australia", "Austria", "Belgium", "Brazil", "Canada", "China", "Colombia", "Croatia", "Denmark", "Egypt", "Finland", "France", "Germany", "Greece", "India", "Indonesia", "Ireland", "Israel", "Italy", "Japan", "Malaysia", "Mexico", "The Netherlands", "New Zealand", "Norway", "Pakistan", "The Philippines", "Poland", "Portugal", "Romania", "Russian Federation", "Singapore", "South Africa", "South Korea", "Spain", "Sweden", "Switzerland", "Thailand", "Turkey", "UK", "United Arab Emirates", "USA"),
"GEI" = c(0.701, 0.632, 0.734, 0.733, 0.733, 0.753, 0.686, 0.740, 0.682, 0.725, 0.708, 0.767, 0.599, 0.850, 0.761, 0.779, 0.685, 0.664, 0.681, 0.807, 0.712, 0.726, 0.670, 0.655, 0.699, 0.776, 0.782, 0.850, 0.559, 0.790, 0.715, 0.731, 0.693, 0.694, 0.711, 0.759, 0.651, 0.742, 0.823, 0.785, 0.706, 0.624, 0.758, 0.646, 0.740),
"dBias" = c(sum(dAlb), sum(dAlg), sum(dArg), sum(dAus), sum(dHab), sum(dBel), sum(dBra), sum(dCan), sum(dChi), sum(dCol), sum(dCro), sum(dDen), sum(dEgy), sum(dFin), sum(dFra), sum(dGer), sum(dGre), sum(dInd), sum(dIno), sum(dIre), sum(dIsr), sum(dIta), sum(dJap), sum(dMal), sum(dMex), sum(dNet), sum(dNew), sum(dNor), sum(dPak), sum(dPhi), sum(dPol), sum(dPor), sum(dRom), sum(dRus), sum(dSin), sum(dSou), sum(dSKo), sum(dSpa), sum(dSwe), sum(dSwi), sum(dTha), sum(dTur), sum(dUK), sum(dUAE), sum(dUSA)),
"dReal" = c(dOAlb$cohen.d[,2], dOAlg$cohen.d[,2], dOArg$cohen.d[,2], dOAus$cohen.d[,2], dOHab$cohen.d[,2], dOBel$cohen.d[,2], dOBra$cohen.d[,2], dOCan$cohen.d[,2], dOChi$cohen.d[,2], dOCol$cohen.d[,2], dOCro$cohen.d[,2], dODen$cohen.d[,2], dOEgy$cohen.d[,2], dOFin$cohen.d[,2], dOFra$cohen.d[,2], dOGer$cohen.d[,2], dOGre$cohen.d[,2], dOInd$cohen.d[,2], dOIno$cohen.d[,2], dOIre$cohen.d[,2], dOIsr$cohen.d[,2], dOIta$cohen.d[,2], dOJap$cohen.d[,2], dOMal$cohen.d[,2], dOMex$cohen.d[,2], dONet$cohen.d[,2], dONew$cohen.d[,2], dONor$cohen.d[,2], dOPak$cohen.d[,2], dOPhi$cohen.d[,2], dOPol$cohen.d[,2], dOPor$cohen.d[,2], dORom$cohen.d[,2], dORus$cohen.d[,2], dOSin$cohen.d[,2], dOSou$cohen.d[,2], dOSKo$cohen.d[,2], dOSpa$cohen.d[,2], dOSwe$cohen.d[,2], dOSwi$cohen.d[,2], dOTha$cohen.d[,2], dOTur$cohen.d[,2], dOUK$cohen.d[,2], dOUAE$cohen.d[,2], dOUSA$cohen.d[,2]), #Ignore
"dRealP" = c(dOPAlb, dOPAlg, dOPArg, dOPAus, dOPHab, dOPBel, dOPBra, dOPCan, dOPChi, dOPCol, dOPCro, dOPDen, dOPEgy, dOPFin, dOPFra, dOPGer, dOPGre, dOPInd, dOPIno, dOPIre, dOPIsr, dOPIta, dOPJap, dOPMal, dOPMex, dOPNet, dOPNew, dOPNor, dOPPak, dOPPhi, dOPPol, dOPPor, dOPRom, dOPRus, dOPSin, dOPSou, dOPSKo, dOPSpa, dOPSwe, dOPSwi, dOPTha, dOPTur, dOPUK, dOPUAE, dOPUSA),
"N" = c(nrow(ALBIP), nrow(ALGIP), nrow(ARGIP), nrow(AUSIP), nrow(HABIP), nrow(BELIP), nrow(BRAIP), nrow(CANIP), nrow(CHIIP), nrow(COLIP), nrow(CROIP), nrow(DENIP), nrow(EGYIP), nrow(FINIP), nrow(FRAIP), nrow(GERIP), nrow(GREIP), nrow(INDIP), nrow(INOIP), nrow(IREIP), nrow(ISRIP), nrow(ITAIP), nrow(JAPIP), nrow(MALIP), nrow(MEXIP), nrow(NETIP), nrow(NEWIP), nrow(NORIP), nrow(PAKIP), nrow(PHIIP), nrow(POLIP), nrow(PORIP), nrow(ROMIP), nrow(RUSIP), nrow(SINIP), nrow(SOUIP), nrow(SKOIP), nrow(SPAIP), nrow(SWEIP), nrow(SWIIP), nrow(THAIP), nrow(TURIP), nrow(UKIP), nrow(UAEIP), nrow(USAIP)))
NeuCou$dDiff = NeuCou$dReal - NeuCou$dBias #Ignore
NeuCou$dDiffP = NeuCou$dRealP - NeuCou$dBiasSummary Statistics
NeuCoumean(NeuCou$dBias); weighted.mean(NeuCou$dBias, sqrt(NeuCou$N))## [1] -0.02233333## [1] 0.02641572mean(NeuCou$dRealP); weighted.mean(NeuCou$dRealP, sqrt(NeuCou$N))## [1] 0.3054914## [1] 0.3144421mean(NeuCou$dDiffP); weighted.mean(NeuCou$dDiffP, sqrt(NeuCou$N))## [1] 0.3278248## [1] 0.2880264Correlations
One could use the correlation with N as a negative control of sorts.
cor(NeuCou[2:8])## GEI dBias dReal dRealP N dDiff
## GEI 1.00000000 -0.07969057 0.15467620 0.16319557 0.07128976 0.1364493
## dBias -0.07969057 1.00000000 0.02608998 0.01573145 0.13428636 -0.9140270
## dReal 0.15467620 0.02608998 1.00000000 0.99790172 0.04101710 0.3816683
## dRealP 0.16319557 0.01573145 0.99790172 1.00000000 0.03665477 0.3903945
## N 0.07128976 0.13428636 0.04101710 0.03665477 1.00000000 -0.1075187
## dDiff 0.13644931 -0.91402700 0.38166833 0.39039450 -0.10751867 1.0000000
## dDiffP 0.12915084 -0.94487317 0.30200381 0.31253181 -0.11557178 0.9961936
## dDiffP
## GEI 0.1291508
## dBias -0.9448732
## dReal 0.3020038
## dRealP 0.3125318
## N -0.1155718
## dDiff 0.9961936
## dDiffP 1.0000000Score by Sex Plots
IPIPA_Means <- IPIPA %>% group_by(GEI, SEX) %>% summarise(Neuroticism = mean(Neuroticism), N = n())## `summarise()` has grouped output by 'GEI'. You can override using the `.groups` argument.ggplot(IPIPA_Means, aes(x = GEI, y = Neuroticism, color = as.factor(SEX), group = as.factor(SEX))) +
geom_point() +
geom_smooth(method = "lm") +
labs(title = "", x = "Gender Equality Index", y = "Neuroticism Mean") +
theme_bw() +
scale_color_manual(values = c("steelblue", "magenta"), name = "Sex", labels = c("Male", "Female")) +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5),
legend.background = element_blank(), legend.key = element_blank(), legend.position = c(0.1, 0.1)) +
guides(color=guide_legend(override.aes=list(fill=NA)))## `geom_smooth()` using formula 'y ~ x'
ggplot(IPIPA, aes(x = GEI, y = Neuroticism, color = as.factor(SEX), group = as.factor(SEX))) +
geom_point() +
geom_smooth(method = "lm") +
labs(title = "", x = "Gender Equality Index", y = "Neuroticism Score") + theme_bw() +
scale_color_manual(values = c("steelblue", "magenta"), name = "Sex", labels = c("Male", "Female")) +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5),
legend.position = c(0.86, 0.92), legend.background = element_blank(), legend.key = element_blank()) +
guides(color=guide_legend(override.aes=list(fill=NA)))## `geom_smooth()` using formula 'y ~ x'
Regression Plots
ggplot(NeuCou, aes(x = dBias, y = dRealP)) +
geom_point() +
geom_smooth(method = lm, color = "gold", formula = "y ~ x") +
labs(title = "Sex Differences vs Test Bias", x = "Mean Bias (dMACS)", y = "Mean Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = dBias, y = dRealP)) +
geom_point() +
geom_smooth(method = lm, color = "gold", formula = "y ~ x", mapping = aes(weight = sqrt(N))) +
labs(title = "Sex Differences vs Test Bias (Square Root-N-Weighted)", x = "Mean Bias (dMACS)", y = "Mean Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
Linear regression was obviously not appropriate for the relationship between bias and differences. Circularity/ipsitivity/ipsativity/uniform bias likely played a considerable role here.
ggplot(NeuCou, aes(x = GEI, y = dRealP)) +
geom_point() +
geom_smooth(method = lm, color = "steelblue", formula = "y ~ x") +
labs(title = "Neuroticism Differences vs Gender Equality", x = "Gender Equality Index", y = "Neuroticism Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = GEI, y = dRealP)) +
geom_point() +
geom_smooth(method = lm, color = "steelblue", formula = "y ~ x", mapping = aes(weight = sqrt(N))) +
labs(title = "Neuroticism Differences vs Gender Equality (Square Root-N-Weighted)", x = "Gender Equality Index", y = "Neuroticism Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = GEI, y = dBias)) +
geom_point() +
geom_smooth(method = lm, color = "darkgreen", formula = "y ~ x") +
labs(title = "Test Bias vs Gender Equality", x = "Gender Equality Index", y = "Mean Bias (dMACS)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = GEI, y = dBias)) +
geom_point() +
geom_smooth(method = lm, color = "darkgreen", formula = "y ~ x", mapping = aes(weight = sqrt(N))) +
labs(title = "Test Bias vs Gender Equality (Square Root-N-Weighted)", x = "Gender Equality Index", y = "Mean Bias (dMACS)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = GEI, y = dDiffP)) +
geom_point() +
geom_smooth(method = lm, color = "orangered", formula = "y ~ x") +
labs(title = "Sex Differences in Neuroticism sans Bias vs the Gender Equality Index", x = "Gender Equality Index", y = "Unbiased Neuroticism Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
ggplot(NeuCou, aes(x = GEI, y = dDiffP)) +
geom_point() +
geom_smooth(method = lm, color = "orangered", formula = "y ~ x", mapping = aes(weight = sqrt(N))) +
labs(title = "Sex Differences in Neuroticism sans Bias vs the Gender Equality Index (Square Root-N-Weighted)", x = "Gender Equality Index", y = "Unbiased Neuroticism Difference (Cohen's d)") +
theme_bw() +
theme(text = element_text(size = 12, family = "serif"), plot.title = element_text(hjust = 0.5)) 
Regressions
G1 <- lm(dRealP ~ GEI, NeuCou); summary(G1)##
## Call:
## lm(formula = dRealP ~ GEI, data = NeuCou)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.26710 -0.05426 0.00009 0.06706 0.34141
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.08305 0.20581 0.404 0.689
## GEI 0.30879 0.28468 1.085 0.284
##
## Residual standard error: 0.1164 on 43 degrees of freedom
## Multiple R-squared: 0.02663, Adjusted R-squared: 0.003996
## F-statistic: 1.177 on 1 and 43 DF, p-value: 0.2841G2 <- lm(dBias ~ GEI, NeuCou); summary(G2)##
## Call:
## lm(formula = dBias ~ GEI, data = NeuCou)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.6595 -0.2998 -0.1182 0.3332 0.5575
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.2928 0.6033 0.485 0.630
## GEI -0.4374 0.8344 -0.524 0.603
##
## Residual standard error: 0.3412 on 43 degrees of freedom
## Multiple R-squared: 0.006351, Adjusted R-squared: -0.01676
## F-statistic: 0.2748 on 1 and 43 DF, p-value: 0.6028G3 <- lm(dDiffP ~ GEI, NeuCou); summary(G3)##
## Call:
## lm(formula = dDiffP ~ GEI, data = NeuCou)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.40598 -0.32600 -0.07507 0.30673 1.00095
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.2097 0.6317 -0.332 0.741
## GEI 0.7462 0.8737 0.854 0.398
##
## Residual standard error: 0.3573 on 43 degrees of freedom
## Multiple R-squared: 0.01668, Adjusted R-squared: -0.006188
## F-statistic: 0.7294 on 1 and 43 DF, p-value: 0.3978NC <- lm(N ~ GEI, NeuCou); summary(NC)##
## Call:
## lm(formula = N ~ GEI, data = NeuCou)
##
## Residuals:
## Min 1Q Median 3Q Max
## -10326 -6900 -5304 -3715 205270
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -19753 56508 -0.350 0.728
## GEI 36633 78163 0.469 0.642
##
## Residual standard error: 31960 on 43 degrees of freedom
## Multiple R-squared: 0.005082, Adjusted R-squared: -0.01806
## F-statistic: 0.2197 on 1 and 43 DF, p-value: 0.6417Robust Regressions
HC5 beats MM-type estimators like that in the lmrob() function and collapses to HC4 in scenarios without high-leverage, where HC4 beats HC3, even at small sample sizes.
VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.083046 0.220190 0.3772 0.7079
## GEI 0.308790 0.292947 1.0541 0.2977VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.29279 0.65092 0.4498 0.6551
## GEI -0.43744 0.89743 -0.4874 0.6284VG3 <- vcovHC(G3, type = "HC5"); VM3 <- coeftest(G3, vcov = VG3); VM3##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.20974 0.68796 -0.3049 0.7619
## GEI 0.74623 0.94280 0.7915 0.4330VNC <- vcovHC(NC, type = "HC5"); VN1 <- coeftest(NC, vcov = VNC); VN1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -19753 17179 -1.1499 0.2566
## GEI 36633 29301 1.2502 0.2180Weighted Robust Regressions
I have no justification for these, I just ran them because I could, and to assess if the correlation between N and other variables could confound this association. This should not be considered a particularly relevant robustness check/analysis, since N is obviously biased towards Anglophone countries for reasons unrelated to the constructs of interest (i.e., country-level wealth/internet access and proficiency in the default questionnaire language). The results may, just, be more reliable due to their greater sample sizes.
G1 <- lm(dRealP ~ GEI, NeuCou, weights = sqrt(N)); VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.029619 0.178325 0.1661 0.8689
## GEI 0.387378 0.235605 1.6442 0.1074G2 <- lm(dBias ~ GEI, NeuCou, weights = sqrt(N)); VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.038229 0.682698 0.0560 0.9556
## GEI -0.016067 0.954112 -0.0168 0.9866G3 <- lm(dDiffP ~ GEI, NeuCou, weights = sqrt(N)); VG3 <- vcovHC(G3, type = "HC5"); VM3 <- coeftest(G3, vcov = VG3); VM3##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.00861 0.69667 -0.0124 0.9902
## GEI 0.40344 0.96573 0.4178 0.6782N-Adjusted Robust Regressions
I also have no justification for these except as a check on the prior result, as it is another way to correct for sample size-related discrepancies. It obviously does not do the same thing as weighting by N, since it corrects the coefficients for them rather than giving them additional leverage by N.
RNeuCou <- umx_residualize(c("dRealP", "dBias", "dDiffP"), "N", data = NeuCou)## (Intercept) B = 0.305 [0.268, 0.341], t = 16.954, p < 0.001
## N B = 0 [0, 0], t = 0.241, p = 0.811
## R² = 0.001 (adj = -0.022)(Intercept) B = -0.032 [-0.136, 0.072], t = -0.616, p = 0.541
## N B = 0 [0, 0], t = 0.889, p = 0.379
## R² = 0.018 (adj = -0.005)(Intercept) B = 0.336 [0.226, 0.446], t = 6.17, p < 0.001
## N B = 0 [0, 0], t = -0.763, p = 0.450
## R² = 0.013 (adj = -0.01)G1 <- lm(dRealP ~ GEI, RNeuCou); VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.21888 0.22018 -0.9941 0.3257
## GEI 0.30385 0.29294 1.0372 0.3054G2 <- lm(dBias ~ GEI, RNeuCou); VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.35298 0.64760 0.5451 0.5885
## GEI -0.48999 0.89202 -0.5493 0.5856G3 <- lm(dDiffP ~ GEI, RNeuCou); VG3 <- vcovHC(G3, type = "HC5"); VM3 <- coeftest(G3, vcov = VG3); VM3##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.57186 0.68516 -0.8346 0.4085
## GEI 0.79383 0.93817 0.8462 0.4022Is there a Sex Interaction in General?
Visually, this would seem impossible, at least for the means. I will test both for the means and at the individual level.
G1 <- lm(Neuroticism ~ GEI, IPIPA_Means); summary(G1)##
## Call:
## lm(formula = Neuroticism ~ GEI, data = IPIPA_Means)
##
## Residuals:
## Min 1Q Median 3Q Max
## -27.9567 -6.4022 0.2246 7.6008 18.4631
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 29.30 12.48 2.348 0.0213 *
## GEI -30.85 17.36 -1.777 0.0792 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 9.482 on 82 degrees of freedom
## Multiple R-squared: 0.03709, Adjusted R-squared: 0.02535
## F-statistic: 3.159 on 1 and 82 DF, p-value: 0.07924G2 <- lm(Neuroticism ~ GEI * SEX, IPIPA_Means); summary(G2)##
## Call:
## lm(formula = Neuroticism ~ GEI * SEX, data = IPIPA_Means)
##
## Residuals:
## Min 1Q Median 3Q Max
## -21.448 -3.089 1.372 3.939 11.319
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 27.9471 27.4904 1.017 0.312
## GEI -57.3387 38.2343 -1.500 0.138
## SEX 0.9025 17.3864 0.052 0.959
## GEI:SEX 17.6602 24.1815 0.730 0.467
##
## Residual standard error: 6.605 on 80 degrees of freedom
## Multiple R-squared: 0.5442, Adjusted R-squared: 0.5271
## F-statistic: 31.84 on 3 and 80 DF, p-value: 1.207e-13VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 29.301 12.408 2.3615 0.02057 *
## GEI -30.848 17.327 -1.7804 0.07872 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 27.94705 24.71111 1.1310 0.26145
## GEI -57.33872 34.22302 -1.6754 0.09775 .
## SEX 0.90245 16.12206 0.0560 0.95550
## GEI:SEX 17.66022 22.23082 0.7944 0.42931
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1G1 <- lm(Neuroticism ~ GEI, IPIPA); summary(G1)##
## Call:
## lm(formula = Neuroticism ~ GEI, data = IPIPA)
##
## Residuals:
## Min 1Q Median 3Q Max
## -109.582 -26.546 -1.577 25.418 122.418
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10.0979 2.2265 4.535 5.75e-06 ***
## GEI 0.6538 3.0038 0.218 0.828
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 36.79 on 298609 degrees of freedom
## Multiple R-squared: 1.586e-07, Adjusted R-squared: -3.19e-06
## F-statistic: 0.04737 on 1 and 298609 DF, p-value: 0.8277G2 <- lm(Neuroticism ~ GEI * SEX, IPIPA); summary(G2)##
## Call:
## lm(formula = Neuroticism ~ GEI * SEX, data = IPIPA)
##
## Residuals:
## Min 1Q Median 3Q Max
## -111.501 -25.501 -1.501 24.499 129.451
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10.306 6.857 1.503 0.132826
## GEI -32.039 9.250 -3.464 0.000533 ***
## SEX -1.084 4.365 -0.248 0.803790
## GEI:SEX 21.671 5.889 3.680 0.000233 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 36.05 on 298607 degrees of freedom
## Multiple R-squared: 0.03966, Adjusted R-squared: 0.03965
## F-statistic: 4111 on 3 and 298607 DF, p-value: < 2.2e-16VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10.09790 2.16131 4.6721 2.982e-06 ***
## GEI 0.65379 2.91710 0.2241 0.8227
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10.3057 6.7027 1.5376 0.1241585
## GEI -32.0388 9.0461 -3.5417 0.0003976 ***
## SEX -1.0845 4.2292 -0.2564 0.7976276
## GEI:SEX 21.6710 5.7078 3.7967 0.0001466 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1There is not a significant direct effect or interaction at the level of the means but there are moderately significant effects in the total dataset, although the main effect only shows up with the interaction included. Graphically, it would seem there are interactions in both cases.
For Robustness’ Sake
G1 <- lm(Neuroticism ~ GEI, weight = sqrt(N), IPIPA_Means); summary(G1)##
## Call:
## lm(formula = Neuroticism ~ GEI, data = IPIPA_Means, weights = sqrt(N))
##
## Weighted Residuals:
## Min 1Q Median 3Q Max
## -130.993 -35.533 -5.772 26.689 157.615
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 17.55 15.16 1.158 0.25
## GEI -12.37 20.61 -0.600 0.55
##
## Residual standard error: 51.52 on 82 degrees of freedom
## Multiple R-squared: 0.004375, Adjusted R-squared: -0.007767
## F-statistic: 0.3603 on 1 and 82 DF, p-value: 0.55G2 <- lm(Neuroticism ~ GEI * SEX, weight = sqrt(N), IPIPA_Means); summary(G2)##
## Call:
## lm(formula = Neuroticism ~ GEI * SEX, data = IPIPA_Means, weights = sqrt(N))
##
## Weighted Residuals:
## Min 1Q Median 3Q Max
## -99.470 -20.291 0.165 14.360 88.613
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 22.04 30.21 0.729 0.468
## GEI -48.41 41.09 -1.178 0.242
## SEX -2.68 19.16 -0.140 0.889
## GEI:SEX 23.28 26.05 0.894 0.374
##
## Residual standard error: 32.55 on 80 degrees of freedom
## Multiple R-squared: 0.6122, Adjusted R-squared: 0.5977
## F-statistic: 42.1 on 3 and 80 DF, p-value: < 2.2e-16VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 17.555 13.712 1.2803 0.2041
## GEI -12.374 19.459 -0.6359 0.5266VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 22.0364 25.9188 0.8502 0.3977
## GEI -48.4094 35.9222 -1.3476 0.1816
## SEX -2.6804 17.0872 -0.1569 0.8757
## GEI:SEX 23.2775 23.6581 0.9839 0.3281RIPIPA_Means <- umx_residualize("Neuroticism", "N", data = IPIPA_Means)## (Intercept) B = 7.07 [4.939, 9.201], t = 6.599, p < 0.001
## N B = 0 [0, 0], t = 0.623, p = 0.535
## R² = 0.005 (adj = -0.007)G1 <- lm(Neuroticism ~ GEI, RIPIPA_Means); summary(G1)##
## Call:
## lm(formula = Neuroticism ~ GEI, data = RIPIPA_Means)
##
## Residuals:
## Min 1Q Median 3Q Max
## -27.8714 -6.5153 0.3128 7.4764 18.3064
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 22.73 12.44 1.828 0.0712 .
## GEI -31.72 17.30 -1.834 0.0703 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 9.448 on 82 degrees of freedom
## Multiple R-squared: 0.0394, Adjusted R-squared: 0.02769
## F-statistic: 3.363 on 1 and 82 DF, p-value: 0.07029G2 <- lm(Neuroticism ~ GEI * SEX, RIPIPA_Means); summary(G2)##
## Call:
## lm(formula = Neuroticism ~ GEI * SEX, data = RIPIPA_Means)
##
## Residuals:
## Min 1Q Median 3Q Max
## -21.3835 -2.9424 0.8854 4.0520 11.1959
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 21.062 27.406 0.769 0.444
## GEI -57.663 38.117 -1.513 0.134
## SEX 1.111 17.333 0.064 0.949
## GEI:SEX 17.296 24.107 0.717 0.475
##
## Residual standard error: 6.585 on 80 degrees of freedom
## Multiple R-squared: 0.5448, Adjusted R-squared: 0.5278
## F-statistic: 31.92 on 3 and 80 DF, p-value: 1.141e-13VG1 <- vcovHC(G1, type = "HC5"); VM1 <- coeftest(G1, vcov = VG1); VM1##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 22.728 12.411 1.8313 0.07068 .
## GEI -31.720 17.329 -1.8304 0.07082 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1VG2 <- vcovHC(G2, type = "HC5"); VM2 <- coeftest(G2, vcov = VG2); VM2##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 21.0619 24.7327 0.8516 0.39699
## GEI -57.6635 34.2536 -1.6834 0.09619 .
## SEX 1.1107 16.1409 0.0688 0.94531
## GEI:SEX 17.2959 22.2596 0.7770 0.43945
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Conclusions
The GEI used by Giolla & Kajonius (2018) was positively related to the extent of sex differences in neuroticism. More equal countries yielded some some scant evidence for larger differences in more gender-equal countries in this dataset. A major problem with the analyses here is the lack of granularity in the gender inequality index. This could be, perhaps, solved by using age and a suite of assumptions: that the past was less gender equal in a granular fashion, that sampling bias is not bias with respect to personality in a way that’s relevant to the hypothesis, and that the effect is one that persists from early life and there aren’t aging effects, unless we can get data actually given in earlier years. Since that is not a possibility for a meaningful era of time because of recent invention of the tests and the low likelihood of invariance for even more recent periods, it seems the an age-based analysis ought to be done, and it ought to showcase invariance by age, with differences theoretically due to the (shared by age) effect of gender equality, though excluding sampling and other sources of difference would be difficult. It could also be done, with sufficient sample sizes, using GEI values for years of administration if we have enough variation in the years of administration. However, the fact that gender equality indices are made based on a combination of socially relevant and irrelevant things may make analyzing the covariance between change in indices and personality a fraught endeavor. To some degree, gender equality indices may reflect characteristics of populations, signaling again the need to address spatial autocorrelation. In any case, there is much more to say and far more to do to properly investigate the idea that gender equality is relevant to personality differences at all, let alone in a systematic fashion.
Other interesting findings included:
- There was no relationship between test bias and the extent of sex differences across countries.
- Test bias is present and systematic with respect to the scale of mean differences, but its direction was consistently such that it raised female scores, in a v-shaped pattern.
- It is likely that bias was artefactual, as the pattern of bias was to covary with mean differences. If resampling a pseudo-female group from the male group or vice-versa led to similar results, the artefactual claim could be verified. If it was not, this would be an area ripe for sociological explanation. Unfortunately, this requires large sample sizes, which may, incidentally, be the explanation for this problematic pattern in the first place (different means with insufficient n can produce artefactual intercept differences).
- Because this is probably the case, these results probably do not mean much for interpretation except that patterns of relationships between personality and other variables are unlikely to have much inferential value. The reason for this is simple: we cannot say if the differences are “real” or to what extent they represent measurement bias. As I see it, we cannot even tell if they are partly or wholly biased.
Limitations
References
Gunn, H. J., Grimm, K. J., & Edwards, M. C. (2019). Evaluation of Six Effect Size Measures of Measurement Non-Invariance for Continuous Outcomes. Structural Equation Modeling: A Multidisciplinary Journal, 0(0), 1-12. https://doi.org/10.1080/10705511.2019.1689507
Giolla, E. M., & Kajonius, P. J. (2019). Sex differences in personality are larger in gender equal countries: Replicating and extending a surprising finding. International Journal of Psychology, 54(6), 705–711. https://doi.org/10.1002/ijop.12529