# ======== ライブラリ ======== #
library(readxl)
## Warning: package 'readxl' was built under R version 4.3.3
library(dplyr)
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
library(tidyr)
library(ggplot2)
library(lme4)
## Warning: package 'lme4' was built under R version 4.3.3
## Loading required package: Matrix
##
## Attaching package: 'Matrix'
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack
library(lmerTest)
##
## Attaching package: 'lmerTest'
## The following object is masked from 'package:lme4':
##
## lmer
## The following object is masked from 'package:stats':
##
## step
library(emmeans)
## Warning: package 'emmeans' was built under R version 4.3.3
## Welcome to emmeans.
## Caution: You lose important information if you filter this package's results.
## See '? untidy'
library(emmeans)
library(effectsize)
## Warning: package 'effectsize' was built under R version 4.3.3
# ======== データ読み込み ======== #
df <- read_excel("~/Library/CloudStorage/Dropbox/土居_関口/2.Analysis_Doi/writing_results_scored_4criteria.xlsx")
# 条件を統一・順序付け
df <- df %>%
mutate(
Condition = factor(case_when(
条件 == "ai-wcf" ~ "AI",
条件 == "model text" ~ "Model",
条件 == "control" ~ "Control",
TRUE ~ 条件
), levels = c("AI", "Model", "Control")),
ID = as.factor(学籍番号)
)
# ======== LMEを回す関数を定義 ======== #
run_lme <- function(data, criterion_name, pre_col, post_col) {
# データを縦長化
df_long <- data %>%
select(ID, Condition, !!sym(pre_col), !!sym(post_col)) %>%
pivot_longer(
cols = c(!!sym(pre_col), !!sym(post_col)),
names_to = "Time",
values_to = "Score"
) %>%
mutate(
Time = ifelse(Time == pre_col, "Pre", "Post"),
Time = factor(Time, levels = c("Pre", "Post"))
)
# 混合効果モデル
model <- lmer(Score ~ Condition * Time + (1 | ID), data = df_long)
cat("\n\n===== LME結果:", criterion_name, "=====\n")
print(anova(model))
print(summary(model))
# EMM
emm <- emmeans(model, ~ Condition * Time)
# プロット
emm_df <- as.data.frame(emm)
p <- ggplot(emm_df, aes(x = Time, y = emmean, group = Condition, color = Condition)) +
geom_line(linewidth = 1.2) +
geom_point(size = 3) +
geom_errorbar(aes(ymin = emmean - SE, ymax = emmean + SE), width = 0.1, linewidth = 0.8) +
scale_color_manual(values = c("AI" = "#1f78b4", "Model" = "#33a02c", "Control" = "#e31a1c")) +
labs(
title = paste("Interaction Plot:", criterion_name, "(Condition × Time)"),
x = "Time (Pre / Post)",
y = "Estimated Mean Score",
color = "Condition"
) +
theme_minimal(base_size = 14) +
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
axis.title = element_text(face = "bold"),
axis.text = element_text(size = 12),
legend.position = "top",
legend.title = element_text(face = "bold")
) +
coord_cartesian(ylim = c(0, 9))
print(p)
}
# ======== 各観点＋合計スコアでLME実行 ======== #
# Task Achievement
run_lme(df, "Task Achievement", "Pre_TA", "Post_TA")
##
##
## ===== LME結果: Task Achievement =====
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.4006 0.2003 2 91 0.3774 0.6867
## Time 15.1152 15.1152 1 91 28.4823 6.884e-07 ***
## Condition:Time 1.2658 0.6329 2 91 1.1926 0.3081
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Score ~ Condition * Time + (1 | ID)
## Data: df_long
##
## REML criterion at convergence: 539
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.14714 -0.42083 0.07508 0.57099 2.44408
##
## Random effects:
## Groups Name Variance Std.Dev.
## ID (Intercept) 0.6964 0.8345
## Residual 0.5307 0.7285
## Number of obs: 188, groups: ID, 94
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.1000 0.2022 137.6606 15.328 < 2e-16 ***
## ConditionModel 0.2125 0.2815 137.6606 0.755 0.451634
## ConditionControl 0.1500 0.2815 137.6606 0.533 0.595011
## TimePost 0.7333 0.1881 91.0000 3.899 0.000185 ***
## ConditionModel:TimePost -0.1083 0.2618 91.0000 -0.414 0.680011
## ConditionControl:TimePost -0.3896 0.2618 91.0000 -1.488 0.140206
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CndtnM CndtnC TimPst CnM:TP
## ConditinMdl -0.718
## CondtnCntrl -0.718 0.516
## TimePost -0.465 0.334 0.334
## CndtnMdl:TP 0.334 -0.465 -0.240 -0.718
## CndtnCnt:TP 0.334 -0.240 -0.465 -0.718 0.516
# Coherence & Cohesion
run_lme(df, "Coherence & Cohesion", "Pre_CC", "Post_CC")
##
##
## ===== LME結果: Coherence & Cohesion =====
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.6492 0.3246 2 91 0.7369 0.4814
## Time 12.4004 12.4004 1 91 28.1522 7.839e-07 ***
## Condition:Time 1.6613 0.8307 2 91 1.8859 0.1576
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Score ~ Condition * Time + (1 | ID)
## Data: df_long
##
## REML criterion at convergence: 519.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.15376 -0.47666 0.03972 0.63654 1.76965
##
## Random effects:
## Groups Name Variance Std.Dev.
## ID (Intercept) 0.7186 0.8477
## Residual 0.4405 0.6637
## Number of obs: 188, groups: ID, 94
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.133333 0.196561 131.467173 15.941 < 2e-16 ***
## ConditionModel 0.085417 0.273602 131.467173 0.312 0.75539
## ConditionControl -0.008333 0.273602 131.467173 -0.030 0.97575
## TimePost 0.666667 0.171362 91.000000 3.890 0.00019 ***
## ConditionModel:TimePost -0.041667 0.238526 91.000000 -0.175 0.86172
## ConditionControl:TimePost -0.416667 0.238526 91.000000 -1.747 0.08404 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CndtnM CndtnC TimPst CnM:TP
## ConditinMdl -0.718
## CondtnCntrl -0.718 0.516
## TimePost -0.436 0.313 0.313
## CndtnMdl:TP 0.313 -0.436 -0.225 -0.718
## CndtnCnt:TP 0.313 -0.225 -0.436 -0.718 0.516
# Lexical Resource
run_lme(df, "Lexical Resource", "Pre_LR", "Post_LR")
##
##
## ===== LME結果: Lexical Resource =====
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.9444 0.4722 2 91 0.9056 0.4079
## Time 22.0153 22.0153 1 91 42.2193 4.264e-09 ***
## Condition:Time 0.7607 0.3803 2 91 0.7294 0.4850
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Score ~ Condition * Time + (1 | ID)
## Data: df_long
##
## REML criterion at convergence: 546.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.39589 -0.48692 -0.00265 0.59679 1.94321
##
## Random effects:
## Groups Name Variance Std.Dev.
## ID (Intercept) 0.8027 0.8960
## Residual 0.5215 0.7221
## Number of obs: 188, groups: ID, 94
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.23333 0.21010 133.09007 15.390 < 2e-16 ***
## ConditionModel 0.26667 0.29244 133.09007 0.912 0.363
## ConditionControl -0.07708 0.29244 133.09007 -0.264 0.793
## TimePost 0.86667 0.18645 91.00000 4.648 1.13e-05 ***
## ConditionModel:TimePost -0.27292 0.25953 91.00000 -1.052 0.296
## ConditionControl:TimePost -0.27292 0.25953 91.00000 -1.052 0.296
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CndtnM CndtnC TimPst CnM:TP
## ConditinMdl -0.718
## CondtnCntrl -0.718 0.516
## TimePost -0.444 0.319 0.319
## CndtnMdl:TP 0.319 -0.444 -0.229 -0.718
## CndtnCnt:TP 0.319 -0.229 -0.444 -0.718 0.516
# Grammatical Range & Accuracy
run_lme(df, "Grammatical Range & Accuracy", "Pre_GRA", "Post_GRA")
##
##
## ===== LME結果: Grammatical Range & Accuracy =====
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.5846 0.2923 2 91 0.6447 0.5272
## Time 12.9082 12.9082 1 91 28.4726 6.91e-07 ***
## Condition:Time 0.4735 0.2368 2 91 0.5222 0.5950
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Score ~ Condition * Time + (1 | ID)
## Data: df_long
##
## REML criterion at convergence: 519.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.06640 -0.44075 -0.06999 0.57978 1.76979
##
## Random effects:
## Groups Name Variance Std.Dev.
## ID (Intercept) 0.6813 0.8254
## Residual 0.4534 0.6733
## Number of obs: 188, groups: ID, 94
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 3.06667 0.19448 133.76897 15.768 < 2e-16 ***
## ConditionModel 0.27708 0.27071 133.76897 1.024 0.307894
## ConditionControl 0.02708 0.27071 133.76897 0.100 0.920457
## TimePost 0.66667 0.17385 91.00000 3.835 0.000231 ***
## ConditionModel:TimePost -0.19792 0.24199 91.00000 -0.818 0.415563
## ConditionControl:TimePost -0.22917 0.24199 91.00000 -0.947 0.346138
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CndtnM CndtnC TimPst CnM:TP
## ConditinMdl -0.718
## CondtnCntrl -0.718 0.516
## TimePost -0.447 0.321 0.321
## CndtnMdl:TP 0.321 -0.447 -0.231 -0.718
## CndtnCnt:TP 0.321 -0.231 -0.447 -0.718 0.516
# Total
run_lme(df, "Total Score", "Pre_Total", "Post_Total")
##
##
## ===== LME結果: Total Score =====
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 8.625 4.312 2 91 0.6507 0.5241
## Time 246.304 246.304 1 91 37.1629 2.585e-08 ***
## Condition:Time 13.283 6.642 2 91 1.0021 0.3711
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Score ~ Condition * Time + (1 | ID)
## Data: df_long
##
## REML criterion at convergence: 1017
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.26134 -0.42777 0.00483 0.53980 1.99887
##
## Random effects:
## Groups Name Variance Std.Dev.
## ID (Intercept) 11.399 3.376
## Residual 6.628 2.574
## Number of obs: 188, groups: ID, 94
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.53333 0.77516 130.01427 16.169 < 2e-16 ***
## ConditionModel 0.84167 1.07898 130.01427 0.780 0.437
## ConditionControl 0.09167 1.07898 130.01427 0.085 0.932
## TimePost 2.93333 0.66472 91.00000 4.413 2.8e-05 ***
## ConditionModel:TimePost -0.62083 0.92524 91.00000 -0.671 0.504
## ConditionControl:TimePost -1.30833 0.92524 91.00000 -1.414 0.161
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) CndtnM CndtnC TimPst CnM:TP
## ConditinMdl -0.718
## CondtnCntrl -0.718 0.516
## TimePost -0.429 0.308 0.308
## CndtnMdl:TP 0.308 -0.429 -0.221 -0.718
## CndtnCnt:TP 0.308 -0.221 -0.429 -0.718 0.516
# ==== 単純効果比較関数 ==== #
compare_prepost_by_condition <- function(data, criterion_name, pre_col, post_col) {
df_long <- data %>%
select(ID, Condition, !!sym(pre_col), !!sym(post_col)) %>%
pivot_longer(
cols = c(!!sym(pre_col), !!sym(post_col)),
names_to = "Time",
values_to = "Score"
) %>%
mutate(
Time = ifelse(Time == pre_col, "Pre", "Post"),
Time = factor(Time, levels = c("Pre", "Post"))
)
# LME
model <- lmer(Score ~ Condition * Time + (1 | ID), data = df_long)
emm <- emmeans(model, ~ Condition * Time)
cat("\n\n===== 単純効果比較: ", criterion_name, "=====\n")
# 各群内のPre vs Post
result <- pairs(emm, by = "Condition", adjust = "tukey")
print(result)
# 平均とSEを一緒に出す
cat("\n推定平均（参考）:\n")
print(summary(emm))
}
# ==== 各観点で実行 ==== #
compare_prepost_by_condition(df, "Task Achievement", "Pre_TA", "Post_TA")
##
##
## ===== 単純効果比較: Task Achievement =====
## Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.733 0.188 91 -3.899 0.0002
##
## Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.625 0.182 91 -3.432 0.0009
##
## Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.344 0.182 91 -1.887 0.0623
##
## Degrees-of-freedom method: kenward-roger
##
## 推定平均（参考）:
## Condition Time emmean SE df lower.CL upper.CL
## AI Pre 3.10 0.202 138 2.70 3.50
## Model Pre 3.31 0.196 138 2.93 3.70
## Control Pre 3.25 0.196 138 2.86 3.64
## AI Post 3.83 0.202 138 3.43 4.23
## Model Post 3.94 0.196 138 3.55 4.32
## Control Post 3.59 0.196 138 3.21 3.98
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
compare_prepost_by_condition(df, "Coherence & Cohesion", "Pre_CC", "Post_CC")
##
##
## ===== 単純効果比較: Coherence & Cohesion =====
## Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.667 0.171 91 -3.890 0.0002
##
## Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.625 0.166 91 -3.767 0.0003
##
## Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.250 0.166 91 -1.507 0.1353
##
## Degrees-of-freedom method: kenward-roger
##
## 推定平均（参考）:
## Condition Time emmean SE df lower.CL upper.CL
## AI Pre 3.13 0.197 131 2.74 3.52
## Model Pre 3.22 0.190 131 2.84 3.60
## Control Pre 3.12 0.190 131 2.75 3.50
## AI Post 3.80 0.197 131 3.41 4.19
## Model Post 3.84 0.190 131 3.47 4.22
## Control Post 3.38 0.190 131 3.00 3.75
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
compare_prepost_by_condition(df, "Lexical Resource", "Pre_LR", "Post_LR")
##
##
## ===== 単純効果比較: Lexical Resource =====
## Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.867 0.186 91 -4.648 <.0001
##
## Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.594 0.181 91 -3.289 0.0014
##
## Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.594 0.181 91 -3.289 0.0014
##
## Degrees-of-freedom method: kenward-roger
##
## 推定平均（参考）:
## Condition Time emmean SE df lower.CL upper.CL
## AI Pre 3.23 0.210 133 2.82 3.65
## Model Pre 3.50 0.203 133 3.10 3.90
## Control Pre 3.16 0.203 133 2.75 3.56
## AI Post 4.10 0.210 133 3.68 4.52
## Model Post 4.09 0.203 133 3.69 4.50
## Control Post 3.75 0.203 133 3.35 4.15
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
compare_prepost_by_condition(df, "Grammatical Range & Accuracy", "Pre_GRA", "Post_GRA")
##
##
## ===== 単純効果比較: Grammatical Range & Accuracy =====
## Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.667 0.174 91 -3.835 0.0002
##
## Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.469 0.168 91 -2.785 0.0065
##
## Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.438 0.168 91 -2.599 0.0109
##
## Degrees-of-freedom method: kenward-roger
##
## 推定平均（参考）:
## Condition Time emmean SE df lower.CL upper.CL
## AI Pre 3.07 0.194 134 2.68 3.45
## Model Pre 3.34 0.188 134 2.97 3.72
## Control Pre 3.09 0.188 134 2.72 3.47
## AI Post 3.73 0.194 134 3.35 4.12
## Model Post 3.81 0.188 134 3.44 4.18
## Control Post 3.53 0.188 134 3.16 3.90
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
compare_prepost_by_condition(df, "Total Score", "Pre_Total", "Post_Total")
##
##
## ===== 単純効果比較: Total Score =====
## Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -2.93 0.665 91 -4.413 <.0001
##
## Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -2.31 0.644 91 -3.593 0.0005
##
## Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -1.62 0.644 91 -2.525 0.0133
##
## Degrees-of-freedom method: kenward-roger
##
## 推定平均（参考）:
## Condition Time emmean SE df lower.CL upper.CL
## AI Pre 12.5 0.775 130 11.0 14.1
## Model Pre 13.4 0.751 130 11.9 14.9
## Control Pre 12.6 0.751 130 11.1 14.1
## AI Post 15.5 0.775 130 13.9 17.0
## Model Post 15.7 0.751 130 14.2 17.2
## Control Post 14.2 0.751 130 12.8 15.7
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
# --- データを縦長に変換（Total Score用） ---
df_long <- df %>%
select(ID, Condition, Pre_Total, Post_Total) %>%
pivot_longer(
cols = c(Pre_Total, Post_Total),
names_to = "Time",
values_to = "Score"
) %>%
mutate(
Time = factor(ifelse(Time == "Pre_Total", "Pre", "Post"),
levels = c("Pre", "Post")),
Condition = factor(Condition, levels = c("AI", "Model", "Control"))
)
# --- LMEモデル構築 ---
model_total <- lmer(Score ~ Condition * Time + (1 | ID), data = df_long)
# --- 効果量算出 ---
eta_sq <- eta_squared(model_total, partial = TRUE)
eta_sq
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ----------------------------------------------
## Condition | 0.01 | [0.00, 1.00]
## Time | 0.29 | [0.17, 1.00]
## Condition:Time | 0.02 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
# ============================================
# 3要因混合効果モデル
# Criterion × Condition × Time + (1 | ID)
# ============================================
# --------------------------------------------
# 縦長に変換（Pre/Post × 4観点）
# --------------------------------------------
df_long <- df %>%
select(ID, Condition,
starts_with("Pre_TA"), starts_with("Pre_CC"), starts_with("Pre_LR"), starts_with("Pre_GRA"),
starts_with("Post_TA"), starts_with("Post_CC"), starts_with("Post_LR"), starts_with("Post_GRA")) %>%
pivot_longer(
cols = -c(ID, Condition),
names_to = c("Time", "Criterion"),
names_pattern = "(Pre|Post)_(.*)",
values_to = "Score"
) %>%
mutate(
Time = factor(Time, levels = c("Pre", "Post")),
Criterion = factor(Criterion, levels = c("TA", "CC", "LR", "GRA"))
)
# --------------------------------------------
# 3要因混合効果モデルの構築
# --------------------------------------------
model_3way <- lmer(Score ~ Criterion * Condition * Time + (1 | ID), data = df_long)
# --------------------------------------------
# 結果の出力
# --------------------------------------------
cat("\n===== Type III ANOVA: Criterion × Condition × Time =====\n")
##
## ===== Type III ANOVA: Criterion × Condition × Time =====
anova(model_3way, type = 3)
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Criterion 5.863 1.954 3 637 5.8615 0.000596 ***
## Condition 0.434 0.217 2 91 0.6507 0.524114
## Time 61.576 61.576 1 637 184.6821 < 2.2e-16 ***
## Criterion:Condition 0.840 0.140 6 637 0.4201 0.865799
## Criterion:Time 0.863 0.288 3 637 0.8628 0.460115
## Condition:Time 3.321 1.660 2 637 4.9800 0.007144 **
## Criterion:Condition:Time 0.840 0.140 6 637 0.4201 0.865799
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# --------------------------------------------
# 効果量（η²）を算出
# --------------------------------------------
eta_sq <- eta_squared(model_3way, partial = TRUE)
cat("\n===== 効果量 (partial η²) =====\n")
##
## ===== 効果量 (partial η²) =====
print(eta_sq)
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------------------
## Criterion | 0.03 | [0.01, 1.00]
## Condition | 0.01 | [0.00, 1.00]
## Time | 0.22 | [0.18, 1.00]
## Criterion:Condition | 3.94e-03 | [0.00, 1.00]
## Criterion:Time | 4.05e-03 | [0.00, 1.00]
## Condition:Time | 0.02 | [0.00, 1.00]
## Criterion:Condition:Time | 3.94e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
# --------------------------------------------
# 交互作用の推定平均（EMM）と単純効果比較
# --------------------------------------------
emm_3way <- emmeans(model_3way, ~ Criterion * Condition * Time)
cat("\n===== 単純効果比較: 各Criterion内のPre/Post差 =====\n")
##
## ===== 単純効果比較: 各Criterion内のPre/Post差 =====
pairs(emm_3way, by = c("Criterion", "Condition"))
## Criterion = TA, Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.733 0.149 637 -4.919 <.0001
##
## Criterion = CC, Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.667 0.149 637 -4.472 <.0001
##
## Criterion = LR, Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.867 0.149 637 -5.813 <.0001
##
## Criterion = GRA, Condition = AI:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.667 0.149 637 -4.472 <.0001
##
## Criterion = TA, Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.625 0.144 637 -4.330 <.0001
##
## Criterion = CC, Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.625 0.144 637 -4.330 <.0001
##
## Criterion = LR, Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.594 0.144 637 -4.113 <.0001
##
## Criterion = GRA, Condition = Model:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.469 0.144 637 -3.247 0.0012
##
## Criterion = TA, Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.344 0.144 637 -2.381 0.0175
##
## Criterion = CC, Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.250 0.144 637 -1.732 0.0838
##
## Criterion = LR, Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.594 0.144 637 -4.113 <.0001
##
## Criterion = GRA, Condition = Control:
## contrast estimate SE df t.ratio p.value
## Pre - Post -0.438 0.144 637 -3.031 0.0025
##
## Degrees-of-freedom method: kenward-roger
# --------------------------------------------
# 推定平均を確認
# --------------------------------------------
emm_df <- as.data.frame(emm_3way)
head(emm_df)
## Criterion Condition Time emmean SE df lower.CL upper.CL
## TA AI Pre 3.100000 0.2009374 155.66 2.703084 3.496916
## CC AI Pre 3.133333 0.2009374 155.66 2.736417 3.530249
## LR AI Pre 3.233333 0.2009374 155.66 2.836418 3.630249
## GRA AI Pre 3.066667 0.2009374 155.66 2.669751 3.463582
## TA Model Pre 3.312500 0.1945568 155.66 2.928188 3.696812
## CC Model Pre 3.218750 0.1945568 155.66 2.834438 3.603062
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95