library(lme4)## 載入需要的套件：Matrixlibrary(lmerTest)##
## 載入套件：'lmerTest'## 下列物件被遮斷自 'package:lme4':
##
## lmer## 下列物件被遮斷自 'package:stats':
##
## steplibrary(readxl)
Exp1_SCJT_ANOVA_MEM_final <- read_excel("MEM/Exp1 SCJT_ANOVA&MEM_final.xls")
View(Exp1_SCJT_ANOVA_MEM_final)library(readxl)
Exp1_Naming_ANOVA_MEM_final <- read_excel("MEM/Exp1 Naming_ANOVA&MEM_final.xls")
View(Exp1_Naming_ANOVA_MEM_final)##Exp1(friends) SCJT_ACC
#隨機截距&隨機斜率
m1_lme = lmer(ACC~Task*Condition+(1+Condition|Subject)+(1|ID),data=Exp1_SCJT_ANOVA_MEM_final)
summary(m1_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ACC ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## Data: Exp1_SCJT_ANOVA_MEM_final
##
## REML criterion at convergence: -1637.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.5618 -0.0460 0.0562 0.3896 1.3783
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Subject (Intercept) 0.002009 0.04482
## Conditionunlearned 0.002200 0.04690 -0.81
## ID (Intercept) 0.002624 0.05122
## Residual 0.032814 0.18115
## Number of obs: 3081, groups: Subject, 28; ID, 23
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 9.415e-01 1.755e-02 8.562e+01 53.660 <2e-16 ***
## Task 3.247e-03 3.264e-03 3.004e+03 0.995 0.320
## Conditionunlearned 2.145e-02 1.770e-02 1.684e+02 1.212 0.227
## Task:Conditionunlearned -1.483e-15 4.616e-03 3.004e+03 0.000 1.000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.558
## Condtnnlrnd -0.574 0.553
## Tsk:Cndtnnl 0.395 -0.707 -0.783#只有隨機截距
m2_lme = lmer(ACC~Task*Condition+(1|Subject)+(1|ID),data=Exp1_SCJT_ANOVA_MEM_final)
summary(m2_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ACC ~ Task * Condition + (1 | Subject) + (1 | ID)
## Data: Exp1_SCJT_ANOVA_MEM_final
##
## REML criterion at convergence: -1614.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.4898 -0.0390 0.0966 0.3214 1.2890
##
## Random effects:
## Groups Name Variance Std.Dev.
## Subject (Intercept) 0.0008901 0.02984
## ID (Intercept) 0.0026687 0.05166
## Residual 0.0333182 0.18253
## Number of obs: 3081, groups: Subject, 28; ID, 23
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 9.413e-01 1.648e-02 8.352e+01 57.117 <2e-16 ***
## Task 3.247e-03 3.289e-03 3.029e+03 0.987 0.324
## Conditionunlearned 2.149e-02 1.544e-02 3.031e+03 1.392 0.164
## Task:Conditionunlearned -1.786e-15 4.651e-03 3.029e+03 0.000 1.000
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.599
## Condtnnlrnd -0.468 0.639
## Tsk:Cndtnnl 0.423 -0.707 -0.904anova(m1_lme, m2_lme)## refitting model(s) with ML (instead of REML)## Data: Exp1_SCJT_ANOVA_MEM_final
## Models:
## m2_lme: ACC ~ Task * Condition + (1 | Subject) + (1 | ID)
## m1_lme: ACC ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m2_lme 7 -1634.6 -1592.4 824.32 -1648.6
## m1_lme 9 -1652.6 -1598.3 835.28 -1670.6 21.931 2 1.729e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#選擇m1，隨機截距&隨機斜率##Exp1(friends) SCJT_RT
#隨機截距&隨機斜率
m3_lme = lmer(RT~Task*Condition+(1+Condition|Subject)+(1|ID),data=Exp1_SCJT_ANOVA_MEM_final)
summary(m3_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: RT ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## Data: Exp1_SCJT_ANOVA_MEM_final
##
## REML criterion at convergence: 39347.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9716 -0.5049 -0.1537 0.2395 16.4580
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Subject (Intercept) 7707.8 87.79
## Conditionunlearned 286.7 16.93 -0.82
## ID (Intercept) 1996.3 44.68
## Residual 33012.6 181.69
## Number of obs: 2963, groups: Subject, 28; ID, 23
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 745.371 22.138 62.137 33.670 < 2e-16 ***
## Task -24.234 3.355 2884.480 -7.224 6.43e-13 ***
## Conditionunlearned -46.936 16.004 490.407 -2.933 0.00352 **
## Task:Conditionunlearned 1.945 4.716 2883.924 0.412 0.68011
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.456
## Condtnnlrnd -0.474 0.630
## Tsk:Cndtnnl 0.324 -0.711 -0.886#只有隨機截距
m4_lme = lmer(RT~Task*Condition+(1|Subject)+(1|ID),data=Exp1_SCJT_ANOVA_MEM_final)
summary(m4_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: RT ~ Task * Condition + (1 | Subject) + (1 | ID)
## Data: Exp1_SCJT_ANOVA_MEM_final
##
## REML criterion at convergence: 39350.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.0186 -0.5065 -0.1546 0.2437 16.3645
##
## Random effects:
## Groups Name Variance Std.Dev.
## Subject (Intercept) 6493 80.58
## ID (Intercept) 1974 44.42
## Residual 33082 181.88
## Number of obs: 2963, groups: Subject, 28; ID, 23
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 745.164 21.116 73.998 35.289 < 2e-16 ***
## Task -24.235 3.358 2910.578 -7.217 6.75e-13 ***
## Conditionunlearned -46.874 15.697 2910.837 -2.986 0.00285 **
## Task:Conditionunlearned 1.928 4.721 2910.542 0.408 0.68302
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.478
## Condtnnlrnd -0.376 0.643
## Tsk:Cndtnnl 0.340 -0.711 -0.904anova(m3_lme,m4_lme)## refitting model(s) with ML (instead of REML)## Data: Exp1_SCJT_ANOVA_MEM_final
## Models:
## m4_lme: RT ~ Task * Condition + (1 | Subject) + (1 | ID)
## m3_lme: RT ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m4_lme 7 39387 39429 -19686 39373
## m3_lme 9 39387 39441 -19685 39369 3.2869 2 0.1933#選擇m4，只有隨機截距##Exp1(friends) Naming_ACC
#隨機截距&隨機斜率
m5_lme = lmer(ACC~Task*Condition+(1+Condition|Subject)+(1|ID),data=Exp1_Naming_ANOVA_MEM_final)
summary(m5_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ACC ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## Data: Exp1_Naming_ANOVA_MEM_final
##
## REML criterion at convergence: -2408.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -6.0478 -0.0092 0.0212 0.1637 2.2262
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Subject (Intercept) 0.0006313 0.02513
## Conditionunlearned 0.0016202 0.04025 -0.66
## ID (Intercept) 0.0053850 0.07338
## Residual 0.0272253 0.16500
## Number of obs: 3360, groups: Subject, 28; ID, 24
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 9.417e-01 1.833e-02 4.586e+01 51.368 < 2e-16 ***
## Task 9.524e-03 2.847e-03 3.280e+03 3.346 0.00083 ***
## Conditionunlearned 2.173e-02 1.537e-02 1.693e+02 1.414 0.15923
## Task:Conditionunlearned -1.042e-02 4.026e-03 3.280e+03 -2.588 0.00971 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.466
## Condtnnlrnd -0.400 0.556
## Tsk:Cndtnnl 0.329 -0.707 -0.786#只有隨機截距
m6_lme = lmer(ACC~Task*Condition+(1|Subject)+(1|ID),data=Exp1_Naming_ANOVA_MEM_final)
summary(m6_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ACC ~ Task * Condition + (1 | Subject) + (1 | ID)
## Data: Exp1_Naming_ANOVA_MEM_final
##
## REML criterion at convergence: -2389.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.9405 -0.0523 0.0623 0.1705 2.0129
##
## Random effects:
## Groups Name Variance Std.Dev.
## Subject (Intercept) 0.0003704 0.01925
## ID (Intercept) 0.0055246 0.07433
## Residual 0.0275991 0.16613
## Number of obs: 3360, groups: Subject, 28; ID, 24
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 9.417e-01 1.827e-02 4.474e+01 51.543 <2e-16 ***
## Task 9.524e-03 2.866e-03 3.306e+03 3.323 0.0009 ***
## Conditionunlearned 2.173e-02 1.344e-02 3.306e+03 1.616 0.1061
## Task:Conditionunlearned -1.042e-02 4.053e-03 3.306e+03 -2.570 0.0102 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.471
## Condtnnlrnd -0.368 0.640
## Tsk:Cndtnnl 0.333 -0.707 -0.905anova(m5_lme,m6_lme)## refitting model(s) with ML (instead of REML)## Data: Exp1_Naming_ANOVA_MEM_final
## Models:
## m6_lme: ACC ~ Task * Condition + (1 | Subject) + (1 | ID)
## m5_lme: ACC ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m6_lme 7 -2410.3 -2367.4 1212.1 -2424.3
## m5_lme 9 -2424.7 -2369.6 1221.3 -2442.7 18.396 2 0.0001012 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#選擇m5，隨機截距&隨機斜率#分割資料
fixA <- split(Exp1_Naming_ANOVA_MEM_final,Exp1_Naming_ANOVA_MEM_final$Task)
fixB <- split(Exp1_Naming_ANOVA_MEM_final,Exp1_Naming_ANOVA_MEM_final$Condition)#在不同檢測階段的簡單主要效果
fixA1.aov <- aov(ACC ~ Condition, data = fixA$`1`)
fixA2.aov <- aov(ACC ~ Condition, data = fixA$`2`)
fixA3.aov <- aov(ACC ~ Condition, data = fixA$`3`)
fixA4.aov <- aov(ACC ~ Condition, data = fixA$`4`)
fixA5.aov <- aov(ACC ~ Condition, data = fixA$`5`)summary(fixA1.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 1 0.12 0.12054 2.443 0.119
## Residuals 670 33.06 0.04934summary(fixA2.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 1 0.054 0.05357 1.855 0.174
## Residuals 670 19.351 0.02888summary(fixA3.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 1 0.037 0.03720 1.227 0.268
## Residuals 670 20.307 0.03031summary(fixA4.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 1 0.037 0.0372 1.353 0.245
## Residuals 670 18.426 0.0275summary(fixA5.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Condition 1 0.121 0.12054 3.993 0.0461 *
## Residuals 670 20.223 0.03018
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#在不同同音鄰群類型的簡單主要效果
fixB1.aov <- aov(ACC ~ Task, data = fixB$`learned`)
fixB2.aov <- aov(ACC ~ Task, data = fixB$`unlearned`)summary(fixB1.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Task 1 0.30 0.30476 10.61 0.00115 **
## Residuals 1678 48.21 0.02873
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1summary(fixB2.aov)## Df Sum Sq Mean Sq F value Pr(>F)
## Task 1 0.0 0.00268 0.071 0.79
## Residuals 1678 63.4 0.03779##Exp1(friends) Naming_RT
#隨機截距&隨機斜率用
m7_lme = lmer(RT~Task*Condition+(1+Condition|Subject)+(1|ID),data=Exp1_Naming_ANOVA_MEM_final)
summary(m7_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: RT ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## Data: Exp1_Naming_ANOVA_MEM_final
##
## REML criterion at convergence: 38502.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.6553 -0.4722 -0.0773 0.3395 15.1467
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Subject (Intercept) 5115.1 71.52
## Conditionunlearned 835.9 28.91 -0.75
## ID (Intercept) 1964.0 44.32
## Residual 9275.1 96.31
## Number of obs: 3200, groups: Subject, 28; ID, 24
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 543.1939 17.2321 54.1818 31.522 < 2e-16 ***
## Task -5.8537 1.7046 3121.5655 -3.434 0.000603 ***
## Conditionunlearned 1.4730 9.7149 124.2674 0.152 0.879729
## Task:Conditionunlearned -0.8376 2.4123 3121.4607 -0.347 0.728457
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.299
## Condtnnlrnd -0.526 0.531
## Tsk:Cndtnnl 0.212 -0.707 -0.749#只有隨機截距
m8_lme = lmer(RT~Task*Condition+(1|Subject)+(1|ID),data=Exp1_Naming_ANOVA_MEM_final)
summary(m8_lme)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: RT ~ Task * Condition + (1 | Subject) + (1 | ID)
## Data: Exp1_Naming_ANOVA_MEM_final
##
## REML criterion at convergence: 38542.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5717 -0.4690 -0.0728 0.3318 15.5507
##
## Random effects:
## Groups Name Variance Std.Dev.
## Subject (Intercept) 3753 61.26
## ID (Intercept) 2157 46.44
## Residual 9460 97.26
## Number of obs: 3200, groups: Subject, 28; ID, 24
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 543.082 16.030 61.541 33.879 < 2e-16 ***
## Task -5.811 1.721 3146.121 -3.376 0.000743 ***
## Conditionunlearned 1.522 8.111 3146.110 0.188 0.851171
## Task:Conditionunlearned -0.850 2.436 3145.994 -0.349 0.727151
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Task Cndtnn
## Task -0.325
## Condtnnlrnd -0.254 0.642
## Tsk:Cndtnnl 0.230 -0.707 -0.905anova(m7_lme,m8_lme)## refitting model(s) with ML (instead of REML)## Data: Exp1_Naming_ANOVA_MEM_final
## Models:
## m8_lme: RT ~ Task * Condition + (1 | Subject) + (1 | ID)
## m7_lme: RT ~ Task * Condition + (1 + Condition | Subject) + (1 | ID)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## m8_lme 7 38574 38616 -19280 38560
## m7_lme 9 38539 38593 -19260 38521 39.085 2 3.256e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#選擇m7，隨機截距&隨機斜率