# LuCiD WP8 Study 1: Longitudinal study # LMER-analysis of predictors of Complement-Clause Proficiency at Time 2 (stepwise backwards selection) lucid1comp<-read.csv2(file.choose()) # Choose "LuCiD_WP8_Study1_T2_COMP.csv" library(lme4) library(lmerTest) # First full model including all experimental and control variables: luc1compa.lmer <-lmer(Complements2 ~ EFB+ToM+Age2+Gender+Sibling+STM+WM+DCCS+ICmot+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1071.1 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.24266 -0.78913 0.00425 0.74820 2.42479 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 6.948e-02 0.263583 # Item (Intercept) 7.809e-02 0.279443 # Experimenter (Intercept) 2.704e-02 0.164440 # Set (Intercept) 0.000e+00 0.000000 # Order (Intercept) 4.417e-05 0.006646 # Residual 4.367e-01 0.660863 # Number of obs: 472, groups: ID, 30; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.601992 1.268767 14.192897 -1.263 0.2271 # EFB 0.033707 0.055848 12.812234 0.604 0.5567 # ToM -0.018425 0.122032 13.671117 -0.151 0.8822 # Age2 0.025278 0.034643 13.847917 0.730 0.4777 # GenderM -0.021452 0.144881 12.500873 -0.148 0.8847 # Siblingyounger -0.020875 0.183864 14.157428 -0.114 0.9112 # STM 0.038614 0.033632 12.750904 1.148 0.2720 # WM 0.084930 0.074386 13.737925 1.142 0.2731 # DCCS 0.008307 0.030963 13.308857 0.268 0.7926 # ICmot -0.043135 0.087336 13.397053 -0.494 0.6294 # ICverb -0.009688 0.018296 13.862126 -0.530 0.6048 # Vocab 0.022447 0.009004 13.864490 2.493 0.0260 * # Gram 0.011230 0.021105 13.432552 0.532 0.6033 # Complements 0.116714 0.050071 444.739412 2.331 0.0202 * # MV1 0.262195 0.170248 12.949969 1.540 0.1476 # MV3 -0.045184 0.106152 13.218389 -0.426 0.6772 # Modals -0.286584 0.146689 12.827093 -1.954 0.0729 . # Sibling status discarded: luc1compb.lmer <-lmer(Complements2 ~ EFB+ToM+Age2+Gender+STM+WM+DCCS+ICmot+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1069.6 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.24764 -0.79083 0.00457 0.74398 2.42372 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 6.335e-02 0.251687 # Item (Intercept) 7.794e-02 0.279180 # Experimenter (Intercept) 2.138e-02 0.146210 # Set (Intercept) 0.000e+00 0.000000 # Order (Intercept) 4.805e-05 0.006932 # Residual 4.368e-01 0.660918 # Number of obs: 472, groups: ID, 30; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.574334 1.199680 15.061333 -1.312 0.2091 # EFB 0.034011 0.053965 13.720613 0.630 0.5389 # ToM -0.021220 0.117235 14.625591 -0.181 0.8588 # Age2 0.024818 0.032541 15.053729 0.763 0.4574 # GenderM -0.023628 0.137864 13.300948 -0.171 0.8665 # STM 0.037890 0.030862 14.181635 1.228 0.2395 # WM 0.084148 0.071070 14.774558 1.184 0.2551 # DCCS 0.006646 0.026874 13.221802 0.247 0.8085 # ICmot -0.039459 0.076746 13.946850 -0.514 0.6152 # ICverb -0.009759 0.017098 15.522936 -0.571 0.5763 # Vocab 0.022333 0.008554 14.511270 2.611 0.0201 * # Gram 0.010517 0.019957 14.023156 0.527 0.6065 # Complements 0.118285 0.049825 442.429236 2.374 0.0180 * # MV1 0.265291 0.152655 14.808257 1.738 0.1030 # MV3 -0.045732 0.100039 14.348229 -0.457 0.6544 # Modals -0.289745 0.134492 14.238694 -2.154 0.0488 * # No effect of discarding Sibling status: anova(luc1compa.lmer, luc1compb.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compb.lmer 22 1037.2 1128.7 -496.61 993.22 # luc1compa.lmer 23 1039.2 1134.8 -496.61 993.22 0 1 0.9995 # Gender discarded: luc1compc.lmer <-lmer(Complements2 ~ EFB+ToM+Age2+STM+WM+DCCS+ICmot+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1067.4 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.25974 -0.78199 0.00381 0.73877 2.42447 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 5.775e-02 2.403e-01 # Item (Intercept) 7.795e-02 2.792e-01 # Experimenter (Intercept) 2.163e-02 1.471e-01 # Set (Intercept) 3.899e-09 6.244e-05 # Order (Intercept) 5.911e-06 2.431e-03 # Residual 4.368e-01 6.609e-01 # Number of obs: 472, groups: ID, 30; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.559844 1.163636 16.153604 -1.340 0.1986 # EFB 0.033874 0.052248 14.664622 0.648 0.5268 # ToM -0.023581 0.113527 15.583952 -0.208 0.8381 # Age2 0.023950 0.031497 16.120916 0.760 0.4580 # STM 0.037576 0.029889 15.146524 1.257 0.2277 # WM 0.085159 0.068972 15.703701 1.235 0.2351 # DCCS 0.007380 0.025553 14.674263 0.289 0.7768 # ICmot -0.039596 0.074356 14.925616 -0.533 0.6022 # ICverb -0.008721 0.015293 16.642856 -0.570 0.5761 # Vocab 0.022336 0.008269 15.552127 2.701 0.0160 * # Gram 0.010842 0.019309 14.989957 0.562 0.5827 # Complements 0.118826 0.049705 442.651748 2.391 0.0172 * # MV1 0.269540 0.147205 15.752338 1.831 0.0861 . # MV3 -0.052892 0.091594 15.701355 -0.577 0.5718 # Modals -0.289496 0.130133 15.171139 -2.225 0.0417 * # No effect of discarding Gender: anova(luc1compb.lmer, luc1compc.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compc.lmer 21 1035.3 1122.6 -496.64 993.29 # luc1compb.lmer 22 1037.2 1128.7 -496.61 993.22 0.0684 1 0.7937 # Theory of Mind Precursors discarded: luc1compd.lmer <-lmer(Complements2 ~ EFB+Age2+STM+WM+DCCS+ICmot+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1064.9 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.25495 -0.77774 0.00414 0.74013 2.41244 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 5.300e-02 2.302e-01 # Item (Intercept) 7.755e-02 2.785e-01 # Experimenter (Intercept) 2.192e-02 1.481e-01 # Set (Intercept) 6.049e-09 7.777e-05 # Order (Intercept) 1.312e-05 3.622e-03 # Residual 4.368e-01 6.609e-01 # Number of obs: 472, groups: ID, 30; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.556801 1.131440 17.258344 -1.376 0.1864 # EFB 0.034192 0.050770 15.660896 0.673 0.5105 # Age2 0.023109 0.030607 17.165286 0.755 0.4605 # STM 0.036470 0.028718 16.181053 1.270 0.2221 # WM 0.086642 0.067102 16.659098 1.291 0.2143 # DCCS 0.008997 0.023620 15.641167 0.381 0.7084 # ICmot -0.032969 0.065571 15.710700 -0.503 0.6221 # ICverb -0.010112 0.013728 17.538213 -0.737 0.4711 # Vocab 0.021897 0.007651 16.196885 2.862 0.0112 * # Gram 0.010867 0.018764 15.968551 0.579 0.5706 # Complements 0.120646 0.049486 439.344914 2.438 0.0152 * # MV1 0.274277 0.142394 16.759321 1.926 0.0712 . # MV3 -0.049628 0.086670 16.617315 -0.573 0.5746 # Modals -0.288919 0.126344 16.150212 -2.287 0.0360 * # No effect of discarding Theory of Mind Precursors anova(luc1compc.lmer, luc1compd.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compd.lmer 20 1033.4 1116.6 -496.71 993.42 # luc1compc.lmer 21 1035.3 1122.6 -496.64 993.29 0.1294 1 0.719 # Cognitive Flexibility (DCCS) discarded: luc1compe.lmer <-lmer(Complements2 ~ EFB+Age2+STM+WM+ICmot+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1132.5 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.1939 -0.8080 -0.0056 0.7513 2.3630 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 3.982e-02 1.996e-01 # Item (Intercept) 7.867e-02 2.805e-01 # Experimenter (Intercept) 2.026e-02 1.424e-01 # Set (Intercept) 8.829e-10 2.971e-05 # Order (Intercept) 0.000e+00 0.000e+00 # Residual 4.434e-01 6.659e-01 # Number of obs: 504, groups: ID, 32; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.636915 0.955321 20.729318 -1.713 0.10154 # EFB 0.032499 0.038896 18.890105 0.836 0.41387 # Age2 0.024377 0.024639 19.905367 0.989 0.33436 # STM 0.036704 0.024503 18.996612 1.498 0.15059 # WM 0.088845 0.051773 19.672563 1.716 0.10186 # ICmot -0.025353 0.057833 18.845386 -0.438 0.66609 # ICverb -0.011362 0.012364 20.523890 -0.919 0.36882 # Vocab 0.022611 0.006814 18.883922 3.318 0.00364 ** # Gram 0.009202 0.014229 18.489526 0.647 0.52578 # Complements 0.139083 0.047553 464.642047 2.925 0.00362 ** # MV1 0.284918 0.123917 19.761512 2.299 0.03254 * # MV3 -0.054827 0.078117 19.552833 -0.702 0.49104 # Modals -0.304198 0.111337 19.309374 -2.732 0.01311 * anova(luc1compd.lmer, luc1compe.lmer) # Error in anova.merMod(luc1compd.lmer, luc1compe.lmer) : # models were not all fitted to the same size of dataset # [Error in comparison due to three children having NA's on DCCS] # Motoric Inhibitory Control discarded: luc1compf.lmer <-lmer(Complements2 ~ EFB+Age2+STM+WM+ICverb+Vocab+Gram+Complements+MV1+MV3+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1464.9 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.19033 -0.76456 -0.03915 0.72445 2.57813 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 4.523e-02 2.127e-01 # Item (Intercept) 8.414e-02 2.901e-01 # Experimenter (Intercept) 4.017e-06 2.004e-03 # Set (Intercept) 1.212e-09 3.482e-05 # Order (Intercept) 0.000e+00 0.000e+00 # Residual 4.389e-01 6.625e-01 # Number of obs: 664, groups: ID, 42; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.176099 0.818670 32.544219 -1.437 0.16038 # EFB 0.074840 0.034433 30.482383 2.173 0.03764 * # Age2 0.017870 0.021227 31.689632 0.842 0.40618 # STM 0.025219 0.017873 30.611406 1.411 0.16832 # WM 0.067200 0.039815 30.621184 1.688 0.10161 # ICverb -0.003171 0.009876 31.507452 -0.321 0.75030 # Vocab 0.015882 0.005220 30.485117 3.043 0.00479 ** # Gram 0.010920 0.011794 30.414836 0.926 0.36179 # Complements 0.116187 0.041396 634.295317 2.807 0.00516 ** # MV1 0.261931 0.082099 30.646172 3.190 0.00327 ** # MV3 -0.010528 0.071765 30.292158 -0.147 0.88434 # Modals -0.261951 0.086411 30.395006 -3.031 0.00494 ** # Model comparison anova(luc1compe.lmer, luc1compf.lmer) # Error in anova.merMod(luc1compe.lmer, luc1compf.lmer) : # models were not all fitted to the same size of dataset # [Error in comparison due to 11 children having NA's on ICmot] # Mental verbs in Third-person complements discarded: luc1compg.lmer <-lmer(Complements2 ~ EFB+Age2+STM+WM+ICverb+Vocab+Gram+Complements+MV1+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1461.4 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.20000 -0.76773 -0.03987 0.72934 2.57619 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 4.300e-02 2.074e-01 # Item (Intercept) 8.368e-02 2.893e-01 # Experimenter (Intercept) 3.639e-06 1.908e-03 # Set (Intercept) 2.860e-09 5.348e-05 # Order (Intercept) 7.374e-07 8.587e-04 # Residual 4.390e-01 6.626e-01 # Number of obs: 664, groups: ID, 42; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.155313 0.797925 33.789510 -1.448 0.15686 # EFB 0.073929 0.033371 31.460898 2.215 0.03410 * # Age2 0.017460 0.020789 32.806643 0.840 0.40707 # STM 0.025195 0.017597 31.583876 1.432 0.16203 # WM 0.065611 0.037762 31.674263 1.738 0.09201 . # ICverb -0.003361 0.009641 32.543165 -0.349 0.72960 # Vocab 0.015999 0.005074 31.435387 3.153 0.00354 ** # Gram 0.010611 0.011430 31.371671 0.928 0.36028 # Complements 0.116710 0.041331 634.431033 2.824 0.00489 ** # MV1 0.260896 0.080569 31.665780 3.238 0.00282 ** # Modals -0.262049 0.085080 31.352978 -3.080 0.00428 ** # No effect of discarding Mental Verbs: Third person anova(luc1compf.lmer, luc1compg.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compg.lmer 17 1432.8 1509.3 -699.42 1398.8 # luc1compf.lmer 18 1434.8 1515.8 -699.40 1398.8 0.0315 1 0.8591 # Verbal Inhibitory Control discarded: luc1comph.lmer <-lmer(Complements2 ~ EFB+Age2+STM+WM+Vocab+Gram+Complements+MV1+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1480.7 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.24960 -0.75952 -0.04789 0.71477 2.58014 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 3.965e-02 1.991e-01 # Item (Intercept) 8.524e-02 2.920e-01 # Experimenter (Intercept) 1.165e-09 3.413e-05 # Set (Intercept) 3.894e-08 1.973e-04 # Order (Intercept) 0.000e+00 0.000e+00 # Residual 4.344e-01 6.591e-01 # Number of obs: 680, groups: ID, 43; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -1.128675 0.772613 35.776378 -1.461 0.15278 # EFB 0.073364 0.031904 33.384854 2.300 0.02786 * # Age2 0.017521 0.020238 34.763737 0.866 0.39257 # STM 0.023457 0.016251 33.842183 1.443 0.15808 # WM 0.062255 0.035420 33.401874 1.758 0.08797 . # Vocab 0.015455 0.004480 33.621155 3.450 0.00153 ** # Gram 0.009849 0.010815 33.359772 0.911 0.36899 # Complements 0.112442 0.040745 645.567747 2.760 0.00595 ** # MV1 0.258033 0.077665 33.435872 3.322 0.00217 ** # Modals -0.255933 0.079404 33.244830 -3.223 0.00284 ** # Model comparison anova(luc1compg.lmer, luc1comph.lmer) # Error in anova.merMod(luc1g.glmer, luc1h.glmer) : # models were not all fitted to the same size of dataset # [Error in comparison due to one child having NA on ICverb] # Age at Time 2 discarded: luc1compi.lmer <-lmer(Complements2 ~ EFB+STM+WM+Vocab+Gram+Complements+MV1+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1475.5 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.29189 -0.75825 -0.05609 0.71270 2.55687 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 3.872e-02 1.968e-01 # Item (Intercept) 8.458e-02 2.908e-01 # Experimenter (Intercept) 1.575e-10 1.255e-05 # Set (Intercept) 0.000e+00 0.000e+00 # Order (Intercept) 8.643e-10 2.940e-05 # Residual 4.346e-01 6.593e-01 # Number of obs: 680, groups: ID, 43; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -0.486027 0.213949 41.519567 -2.272 0.028351 * # EFB 0.078479 0.031129 34.274585 2.521 0.016513 * # STM 0.024301 0.016110 34.529639 1.508 0.140534 # WM 0.069015 0.034302 34.449167 2.012 0.052097 . # Vocab 0.016902 0.004125 35.039373 4.097 0.000235 *** # Gram 0.010060 0.010737 34.087554 0.937 0.355377 # Complements 0.115959 0.040569 640.169684 2.858 0.004398 ** # MV1 0.269046 0.076077 34.382384 3.537 0.001183 ** # Modals -0.262697 0.078472 33.941978 -3.348 0.002004 ** # No effect of discarding Age at Time 2: anova(luc1comph.lmer, luc1compi.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compi.lmer 15 1456.0 1523.8 -713.00 1426.0 # luc1comph.lmer 16 1457.1 1529.5 -712.57 1425.1 0.861 1 0.3535 # Receptive Grammar discarded: luc1compj.lmer <-lmer(Complements2 ~ EFB+STM+WM+Vocab+Complements+MV1+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp) # REML criterion at convergence: 1469.1 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.31050 -0.76005 -0.05145 0.72085 2.56312 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 3.846e-02 1.961e-01 # Item (Intercept) 8.433e-02 2.904e-01 # Experimenter (Intercept) 4.682e-09 6.842e-05 # Set (Intercept) 1.778e-08 1.333e-04 # Order (Intercept) 1.173e-08 1.083e-04 # Residual 4.347e-01 6.593e-01 # Number of obs: 680, groups: ID, 43; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -0.444073 0.208844 42.856291 -2.126 0.039275 * # EFB 0.074205 0.030736 35.200422 2.414 0.021101 * # STM 0.028576 0.015422 35.746557 1.853 0.072156 . # WM 0.061725 0.033348 35.493068 1.851 0.072510 . # Vocab 0.018113 0.003909 36.349994 4.634 4.49e-05 *** # Complements 0.117705 0.040525 641.021652 2.904 0.003805 ** # MV1 0.273511 0.075783 35.334150 3.609 0.000943 *** # Modals -0.265688 0.078258 34.896094 -3.395 0.001725 ** # No effect of discarding Receptive Grammar: anova(luc1compi.lmer, luc1compj.lmer) # Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq) # luc1compj.lmer 14 1455 1518.3 -713.52 1427 # luc1compi.lmer 15 1456 1523.8 -713.00 1426 1.0483 1 0.3059 # check for outliers (large standardised residuals: >2.5) # One outlying observation (out 680), no changes in significance luc1compj.lmerB <-lmer(Complements2 ~ EFB+STM+WM+Vocab+Complements+MV1+Modals+(1|ID)+(1|Item)+(1|Experimenter)+(1|Set)+(1|Order), data = lucid1comp[abs(scale(resid(luc1compj.lmer)))< 2.5,]) summary(luc1compj.lmerB) # REML criterion at convergence: 1467.1 # # Scaled residuals: # Min 1Q Median 3Q Max # -2.31088 -0.75729 -0.05336 0.71009 2.56152 # # Random effects: # Groups Name Variance Std.Dev. # ID (Intercept) 3.822e-02 0.1955037 # Item (Intercept) 8.452e-02 0.2907279 # Experimenter (Intercept) 8.270e-08 0.0002876 # Set (Intercept) 1.003e-07 0.0003168 # Order (Intercept) 0.000e+00 0.0000000 # Residual 4.348e-01 0.6593782 # Number of obs: 679, groups: ID, 43; Item, 16; Experimenter, 2; Set, 2; Order, 2 # # Fixed effects: # Estimate Std. Error df t value Pr(>|t|) # (Intercept) -0.448053 0.208589 42.962839 -2.148 0.037394 * # EFB 0.075101 0.030695 35.284445 2.447 0.019540 * # STM 0.028210 0.015399 35.807487 1.832 0.075282 . # WM 0.062477 0.033297 35.551957 1.876 0.068838 . # Vocab 0.018137 0.003902 36.375604 4.648 4.29e-05 *** # Complements 0.119252 0.040554 639.133867 2.941 0.003394 ** # MV1 0.275339 0.075672 35.397839 3.639 0.000866 *** # Modals -0.268072 0.078157 34.980574 -3.430 0.001565 ** # Marginal effects plot library(ggeffects) plot(ggpredict(luc1compj.lmer, terms = c("EFB", "MV1"))) # Post hoc power # Make sure to prune the data set for rows containing NA's before conducting the simulations, otherwise every simulation will fail, and you'll see observed powers of zero all way. library(simr) powerSim(luc1compj.lmer, fixed("EFB", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("STM", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("WM", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("Vocab", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("Complements", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("MV1", "t"), nsim=1000) powerSim(luc1compj.lmer, fixed("Modals", "t"), nsim=1000)