Table S1: Prospective Actor and Partner Effects of Agency Motives on Differential Change

Supplemental Materials

When "Together" Means "Too Close": Agency Motives and Relationship Functioning in Co-Resident and Living-Apart-Together Couples

by B. Hagemeyer et al., 2015, Journal of Personality and Social Psychology

Table S1:Prospective Actor and Partner Effects of Agency Motives on Differential Change in Subjective Relationship Quality (4-Item Index) Over One Year in COR and LAT couples (Hypothesis 2)

COR / LAT / Group
differences
Predictors / Effects / b (SE) / p / b (SE) / p / χ²(1) / p
Intercept m / –0.166 (0.072) / .021 / –0.054 (0.198) / .784 / 0.320 / .572
Intercept f / –0.096 (0.077) / .212 / –0.016 (0.220) / .942 / 0.166 / .684
Implicit motive / m → m / –0.100 (0.056) / .073 / 0.087 (0.111) / .433 / 3.370 / .066
f → f / –0.057 (0.053) / .279 / –0.018 (0.129) / .889 / 0.104 / .747
m → f / –0.047 (0.057) / .407 / 0.111 (0.119) / .351 / 2.045 / .153
f → m / –0.024 (0.055) / .668 / 0.012 (0.152) / .939 / 0.094 / .759
Explicit desire / m → m / –0.113 (0.050) / .023 / 0.147 (0.126) / .242 / 6.793 / .009
f → f / –0.027 (0.058) / .648 / 0.024 (0.148) / .873 / 0.173 / .677
m → f / –0.031 (0.053) / .554 / 0.018 (0.117) / .881 / 0.212 / .212
f → m / 0.065 (0.052) / .208 / –0.166 (0.119) / .165 / 3.815 / .051
Implicit  Explicit / m → m / –0.071 (0.053) / .180 / –0.011 (0.131) / .931 / 0.341 / .559
f → f / –0.013 (0.047) / .782 / –0.134 (0.170) / .433 / 1.062 / .303
m → f / –0.051 (0.048) / .295 / –0.067 (0.136) / .624 / 0.021 / .886
f → m / –0.027 (0.061) / .655 / –0.171 (0.166) / .305 / 1.385 / .239

Note. COR = co-resident couples. LAT = living-apart-together couples. b = unstandardized path coefficients. SE = bootstrapped standard errors (10,000 re-samples). m = male. f = female. m → f denotes the partner effect of male predictor on female outcome and vice versa. Significant coefficients (p < .05) are printed in boldface. Not displayed: covariates male and female relationship quality at T1, age, relationship duration, and additional cohabitants.

Table S2: Mediation of Actor and Partner Effects of Motives on Subjective Relationship Quality from Model 1 by Dyadic Conflicts

COR / LAT
Direct effects
Motives → SRQ / Indirect effects (ab)
Motives → Con → SRQ / Direct effects
Motives → SRQ / Indirect effects (ab)
Motives → Con → SRQ
Predictors / Effects / b (SE) / p / b(SE) / p / b (SE) / p / b (SE) / p
Implicit motive / m → m / –0.125 (0.043) / .003 / –0.100 (0.026) / <.001
f → f / –0.152 (0.052) / .003 / –0.082 (0.026) / .001 / –0.235 (0.072) / .001 / –0.019 (0.016) / .258
m → f / –0.083 (0.051) / .104 / –0.111 (0.028) / <.001
f → m / –0.085 (0.045) / .056 / –0.074 (0.024) / .002 / –0.145 (0.076) / .057 / –0.018 (0.016) / .258
Explicit desire / m → m / –0.180(0.047) / <.001 / –0.053 (0.024) / .029 / –0.226 (0.057) / <.001 / –0.005 (0.017) / .789
f → f / –0.144 (0.054) / .008 / 0.010 (0.027) / .721
m → f / –0.116 (0.048) / .016 / –0.059 (0.027) / .028
f → m
Implicit  Explicit / m → m / –0.092 (0.040) / .020 / –0.055 (0.023) / .019
f → f
m → f / –0.107 (0.046) / .022 / –0.061 (0.025) / .015
f → m
Conflicts → SRQ / m / –0.433 (0.047) / <.001 / –0.248 (0.066) / <.001
f / –0.483 (0.054) / <.001 / –0.256 (0.069) / <.001

Note. COR = co-resident couples. LAT = living-apart-together couples. b = unstandardized path coefficients. SE = bootstrapped standard errors (10,000 re-samples). m = male. f = female. m → f denotes the partner effect of male predictor on female outcome and vice versa. Significant coefficients (p < .05) are printed in boldface. Not displayed: covariates age, relationship duration, and additional cohabitants.

Table S3: Cross-sectional Actor and Partner Effects of Agency Motives on Subjective Relationship Quality (Assessed With Five Measures) in COR and LAT Couples (Hypothesis 2)

COR / LAT / Group
differences
Predictors / Effects / b (SE) / p / b (SE) / p / χ²(1) / p
Intercept m / 0.058 (0.072) / .422 / –0.441 (0.140) / .002 / 11.247 / <.001
Intercept f / 0.044 (0.076) / .563 / –0.348 (0.133) / .009 / 7.230 / .007
Implicit motive / m → m / –0.235 (0.053) / <.001 / 0.038 (0.070) / .589 / 10.900 / .001
f → f / –0.219 (0.055) / <.001 / –0.224(0.072) / .002 / 0.003 / .955
m → f / –0.187 (0.057) / .001 / 0.028 (0.059) / .638 / 6.763 / .009
f → m / –0.175 (0.048) / <.001 / –0.175 (0.081) / .031 / <0.001 / .999
Explicit desire / m → m / –0.253 (0.052) / <.001 / –0.231 (0.065) / <.001 / 0.073 / .787
f → f / –0.146 (0.059) / .013 / –0.051 (0.063) / .424 / 1.173 / .279
m → f / –0.158 (0.054) / .003 / –0.078 (0.073) / .285 / 0.956 / .328
f → m / –0.083 (0.054) / .125 / 0.077 (0.070) / .270 / 3.332 / .068
Implicit  Explicit / m → m / –0.158 (0.047) / .001 / 0.127 (0.076) / .094 / 11.284 / <.001
f → f / 0.011 (0.066) / .868 / 0.012 (0.071) / .870 / <0.001 / .995
m → f / –0.164 (0.055) / .003 / 0.023 (0.069) / .741 / 4.909 / .027
f → m / 0.013 (0.056) / .812 / 0.028 (0.084) / .742 / 0.026 / .873

Note. COR = co-resident couples. LAT = living-apart-together couples. b = unstandardized path coefficients. SE = bootstrapped standard errors (10,000 re-samples). m = male. f = female. m → f denotes the partner effect of male predictor on female outcome and vice versa. Significant coefficients (p < .05) are printed in boldface. Not displayed: covariatesage, relationship duration, and additional cohabitants. Reliability (ω) of the 5-item SRQ index: .73-.75. Multiple-group test of overall moderation: χ²(12) = 28.365, p = .005, CFI = 0.949, RMSEA = .071.

Table S4: Prospective Actor and Partner Effects of Agency Motives on Subjective Relationship Quality (Assessed With Six Measures) Over One Year in COR and LAT Couples (Hypothesis 2)

COR / LAT / Group differences
Predictors / Effects / b (SE) / p / b (SE) / p / χ²(1) / p
Intercept m / –0.003 (0.097) / .977 / –0.474 (0.264) / .073 / 3.616 / .057
Intercept f / –0.031 (0.106) / .768 / –0.308 (0.205) / .133 / 1.566 / .211
Implicit motive / m → m / –0.297 (0.067) / <.001 / 0.099 (0.121) / .411 / 9.335 / .002
f → f / –0.221 (0.067) / .001 / –0.141 (0.143) / .324 / 0.350 / .554
m → f / –0.221 (0.069) / .001 / 0.121 (0.117) / .304 / 7.268 / .007
f → m / –0.145 (0.067) / .030 / –0.035 (0.151) / .817 / 0.586 / .444
Explicit desire / m → m / –0.288 (0.066) / <.001 / 0.023 (0.134) / .862 / 6.190 / .013
f → f / –0.152 (0.073) / .038 / 0.047 (0.143) / .742 / 2.068 / .151
m → f / –0.198 (0.071) / .005 / –0.027 (0.104) / .800 / 2.063 / .151
f → m / –0.007 (0.068) / .920 / –0.083 (0.136) / .542 / 0.258 / .612
Implicit  Explicit / m → m / –0.190 (0.064) / .003 / 0.137 (0.125) / .274 / 6.686 / .010
f → f / –0.062 (0.071) / .383 / –0.145 (0.170) / .394 / 0.377 / .540
m → f / –0.210 (0.068) / .002 / 0.040 (0.116) / .730 / 4.123 / .042
f → m / –0.016 (0.074) / .833 / –0.111 (0.153) / .470 / 0.386 / .535

Note. COR = co-resident couples. LAT = living-apart-together couples. b = unstandardized path coefficients. SE = bootstrapped standard errors (10,000 re-samples).m = male. f = female. m → f denotes the partner effect of male predictor on female outcome and vice versa. Significant coefficients (p < .05) are printed in boldface. Not displayed: covariates age, relationship duration, and additional cohabitants. Reliability (ω) of the 6-item T2-SRQ index: .81-.84. Multiple-group test of overall moderation: χ²(12) = 31.412, p = .002 , CFI = 0.868, RMSEA = .098.

Table S5: Cross-Sectional Effects of Communal Motives on the Amount of Dyadic Conflicts in COR and LAT Couples

COR / LAT / Group differences
Predictors / Effects / b (SE) / p / b (SE) / p / χ²(1) / p
Intercept / –0.035 (0.066) / .593 / –0.039 (0.132) / .767 / 0.001 / .977
Implicit motive / m / –0.096 (0.045) / .031 / –0.017 (0.064) / .787 / 0.885 / .347
f / –0.055 (0.047) / .242 / 0.063 (0.060) / .300 / 2.366 / .124
Explicit desire / m / –0.347 (0.057) / <.001 / –0.362 (0.074) / <.001 / 0.032 / .858
f / –0.296 (0.064) / <.001 / –0.112 (0.081) / .167 / 4.506 / .034
Implicit  Explicit / m / 0.028 (0.041) / .506 / –0.118 (0.070) / .091 / 3.018 / .082
f / –0.095 (0.045) / .035 / –0.048 (0.064) / .448 / 0.430 / .512

Note. COR = co-resident couples. LAT = living-apart-together couples. b = unstandardized path coefficients. SE = bootstrapped standard errors (10,000 re-samples).m = male. f = female. Significant coefficients (p < .05) are printed in boldface. Not displayed: covariates age, relationship duration, and additional cohabitants. Implicit communal motives were assessed with the PACT (Hagemeyer & Neyer, 2012). Explicit communal motives were assessed with the desire for closeness scale (Hagemeyer, Neyer, et al., 2013)

Table S6: Cross-Sectional Multilevel Regression of Subjective Relationship Quality (4-Item Index) on Implicit and Explicit Agency Motives, Living Arrangement, and Sex (Random Intercept Model; Hypothesis 2)

Predictors / Actor effects / Partner effects
b (SE) / p / b (SE) / p
Intercept / –0.022 (0.037) / .556
Sex (m = –1, f = 1) / 0.002 (0.022) / .919
LA (LAT = –1, COR = 1) / 0.057 (0.041) / .163
Implicit Agency / –0.167 (0.029) / <.001 / –0.127 (0.029) / <.001
Explicit Agency / –0.183 (0.029) / <.001 / –0.080 (0.029) / .006
Implicit  Explicit / –0.012 (0.031) / .696 / –0.021 (0.031) / .498
LA  Implicit Agency / –0.065 (0.029) / .024 / –0.053 (0.029) / .069
LA  Explicit Agency / –0.006 (0.029) / .830 / –0.042 (0.029) / .149
LA  Implicit  Explicit / –0.058 (0.031) / .062 / –0.057 (0.031) / .068
Sex  LA / –0.024 (0.022) / .276
Sex  Implicit Agency / –0.092 (0.031) / .003 / –0.046 (0.031) / .136
Sex  Explicit Agency / 0.060 (0.031) / .057 / 0.057 (0.031) / .071
Sex  Implicit  Explicit / 0.008 (0.031) / .788 / 0.024 (0.031) / .437
Sex  LA  Implicit Agency / 0.087 (0.031) / .005 / 0.062 (0.031) / .044
Sex  LA  Explicit Agency / –0.012 (0.031) / .710 / –0.001 (0.031) / .977
Sex  LA  Implicit  Explicit / 0.066 (0.031) / .037 / 0.063 (0.031) / .043

Note. LA = living arrangement. b = unstandardized regression coefficients. SE = standard errors of regression coefficients. Not displayed: covariates age, sex, and additional cohabitants.

Table S7: Prospective Multilevel Regression of Subjective Relationship Quality (4-Item Index) at T2 on Implicit and Explicit Agency Motives, Living Arrangement, and Sex (Random Intercept Model; Hypothesis 2)

Predictors / Actor effects / Partner effects
b (SE) / p / b (SE) / p
Intercept / –0.097 (0.059) / .101
Sex (m = –1, f = 1) / –0.025 (0.037) / .503
LA (LAT = –1, COR = 1) / 0.103 (0.064) / .107
Implicit Agency / –0.149 (0.044) / .001 / –0.069 (0.046) / .137
Explicit Agency / –0.106 (0.044) / .015 / –0.110 (0.047) / .019
Implicit  Explicit / –0.052 (0.047) / .266 / –0.056 (0.051) / .270
LA  Implicit Agency / –0.111 (0.043) / .011 / –0.103 (0.046) / .027
LA  Explicit Agency / –0.095 (0.043) / .028 / 0.018 (0.046) / .697
LA  Implicit  Explicit / –0.067 (0.047) / .152 / –0.046 (0.051) / .367
Sex  LA / 0.025 (0.037) / .496
Sex  Implicit Agency / –0.032 (0.049) / .511 / –0.018 (0.052) / .730
Sex  Explicit Agency / 0.082 (0.047) / .084 / 0.034 (0.051) / .504
Sex  Implicit  Explicit / 0.044 (0.047) / .343 / –0.002 (0.051) / .967
Sex  LA  Implicit Agency / 0.086 (0.049) / .081 / 0.052 (0.052) / .319
Sex  LA  Explicit Agency / 0.003 (0.047) / .947 / 0.054 (0.050) / .280
Sex  LA  Implicit  Explicit / 0.140 (0.047) / .003 / 0.085 (0.051) / .094

Note. LA = living arrangement. b = unstandardized regression coefficients. SE = standard errors of regression coefficients. Not displayed: covariates age, sex, and additional cohabitants.

Table S8: Cross-Sectional Actor and Partner Effects of Agency Motives on Subjective Relationship Quality (4-Item Index) in COR and LAT Couples Balanced by Propensity Score Matching (Hypothesis2)

COR / LAT
Predictors / Effects / b (SE) / p / b (SE) / p
Intercept m / 0.074 (0.078) / .342 / –0.245 (0.148) / .098
Intercept f / 0.027 (0.080) / .731 / –0.285 (0.158) / .071
Implicit motive / m → m / –0.225 (0.051) / <.001 / 0.181 (0.101) / .072
f → f / –0.212 (0.056) / <.001 / –0.465 (0.150) / .002
m → f / –0.156 (0.054) / .004 / 0.039 (0.111) / .727
f → m / –0.149 (0.052) / .005 / –0.205 (0.151) / .174
Explicit desire / m → m / –0.226 (0.054) / <.001 / –0.215 (0.112) / .055
f → f / –0.095 (0.051) / .061 / 0.137 (0.132) / .301
m → f / –0.173 (0.055) / .002 / –0.260 (0.141) / .066
f → m / –0.089 (0.053) / .089 / 0.169 (0.102) / .096
Implicit  Explicit / m → m / –0.164 (0.050) / .001 / –0.004 (0.116) / .975
f → f / 0.054 (0.057) / .339 / 0.166 (0.120) / .168
m → f / –0.157 (0.056) / .005 / 0.054 (0.156) / .729
f → m / 0.046 (0.052) / .380 / 0.128 (0.154) / .405

Note. Results from 305 COR and 175 LAT couples that were weighted based on propensity scores derived from logistic regression with age, relationship duration, and additional cohabitants as predictors of living arrangement. Couples were matched by a 1:5 nearest-neighbor algorithm with replacement. To avoid bad matches, a caliper of .10 (proportion of the standard deviation of the propensity score logit) was specified. Multivariate imbalance (L1) was reduced from .655 before matching to .483 after matching. Mean differences between the matched groups were d < .18. Propensity score matching was done using the SPPS-to-R extension by Thoemmes (2012). Weighted path analysis was carried out using MLR estimation in Mplus.

1