nobegosu
2022-11-15
library(tidyverse)## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──## ✔ ggplot2 3.3.6 ✔ purrr 0.3.4
## ✔ tibble 3.1.8 ✔ dplyr 1.0.9
## ✔ tidyr 1.2.0 ✔ stringr 1.4.0
## ✔ readr 2.1.2 ✔ forcats 0.5.1## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()library(ggpubr)
library(rstatix)##
## Attaching package: 'rstatix'## The following object is masked from 'package:stats':
##
## filterlibrary(readxl)
dat <- read_excel("/Users/riku/Documents/koto_nobe_12nin.xlsx")
dat$id <- as.factor(dat$id)
dat$trueid <- as.factor(dat$trueid)
dat$task <- ordered(dat$task, levels=c("L","M","H"))
dat$proficiency <- as.factor(dat$proficiency)
dat$proficiency <- ordered(dat$proficiency, levels=c("lower","middle","upper"))
dat %>%
group_by(task, proficiency) %>%
get_summary_stats( type = "mean_sd")## # A tibble: 36 × 6
## task proficiency variable n mean sd
## <ord> <ord> <chr> <dbl> <dbl> <dbl>
## 1 L lower koto 11 47.6 9.43
## 2 L lower nobe 11 89.2 25.1
## 3 L lower wariai 11 0.554 0.09
## 4 L lower wariai2 11 3.59 0.278
## 5 L middle koto 12 56.9 18.2
## 6 L middle nobe 12 115 49.5
## 7 L middle wariai 12 0.529 0.097
## 8 L middle wariai2 12 3.77 0.548
## 9 L upper koto 12 60.3 13.7
## 10 L upper nobe 12 127. 41.0
## # … with 26 more rows
## # ℹ Use `print(n = ...)` to see more rows1 Boxplot
bxp <- ggboxplot(
dat, x = "task", y = "wariai", add = "jitter",
color = "proficiency", palette = "uchicago"
)
bxp + ggtitle("Wariai")
bxp <- ggboxplot(
dat, x = "task", y = "wariai2", add = "jitter",
color = "proficiency", palette = "uchicago"
)
bxp + ggtitle("Wariai2")
bxp <- ggboxplot(
dat, x = "task", y = "nobe", add = "jitter",
color = "proficiency", palette = "uchicago"
)
bxp + ggtitle("Nobegosu")
bxp <- ggboxplot(
dat, x = "task", y = "koto", add = "jitter",
color = "proficiency", palette = "uchicago"
)
bxp + ggtitle("Kotonarugosu")
2 ANOVA Wariai
2.1 Check assumptions
dat %>%
group_by(task, proficiency) %>%
identify_outliers(wariai)## # A tibble: 3 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 L lower 32 33 36 28 0.778 3.30 TRUE FALSE
## 2 L upper 34 35 70 45 0.643 3.80 TRUE FALSE
## 3 M upper 11 11 98 59 0.602 4.21 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat %>%
group_by(task, proficiency) %>%
shapiro_test(wariai)## # A tibble: 9 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 L lower wariai 0.865 0.0664
## 2 L middle wariai 0.975 0.959
## 3 L upper wariai 0.977 0.968
## 4 M lower wariai 0.847 0.0392
## 5 M middle wariai 0.976 0.960
## 6 M upper wariai 0.826 0.0190
## 7 H lower wariai 0.880 0.104
## 8 H middle wariai 0.954 0.692
## 9 H upper wariai 0.949 0.623ggqqplot(dat, "wariai", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat %>%
group_by(task) %>%
levene_test(wariai ~ proficiency)## # A tibble: 3 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 L 2 32 0.286 0.753
## 2 M 2 32 0.653 0.527
## 3 H 2 32 0.813 0.4522.2 Computation
res.aov <- anova_test(
data = dat, dv = wariai, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 2.469 0.101 0.082
## 2 task 2 64 1.602 0.209 0.021
## 3 proficiency:task 4 64 0.123 0.974 0.003interaction.plot(x.factor = dat$task, trace.factor = dat$proficiency,
response = dat$wariai, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
3 ANOVA koto
3.1 Check assumptions
dat %>%
group_by(task, proficiency) %>%
identify_outliers(koto)## # A tibble: 1 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 H lower 20 20 149 65 0.436 3.77 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat %>%
group_by(task, proficiency) %>%
shapiro_test(koto)## # A tibble: 9 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 L lower koto 0.955 0.711
## 2 L middle koto 0.878 0.0814
## 3 L upper koto 0.980 0.982
## 4 M lower koto 0.922 0.335
## 5 M middle koto 0.943 0.537
## 6 M upper koto 0.933 0.410
## 7 H lower koto 0.872 0.0832
## 8 H middle koto 0.904 0.177
## 9 H upper koto 0.888 0.111ggqqplot(dat, "koto", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat %>%
group_by(task) %>%
levene_test(koto ~ proficiency)## # A tibble: 3 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 L 2 32 1.71 0.197
## 2 M 2 32 1.60 0.217
## 3 H 2 32 0.438 0.6493.2 Computation
res.aov <- anova_test(
data = dat, dv = koto, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2.00 32.00 4.510 0.019 * 0.175
## 2 task 1.63 52.28 4.039 0.031 * 0.030
## 3 proficiency:task 3.27 52.28 1.531 0.215 0.023interaction.plot(x.factor = dat$task, trace.factor = dat$proficiency,
response = dat$koto, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
4 Comparison between L and M
dat_LM <- dat %>% filter(task == "L"| task == "M")
dat_LM$task <- ordered(dat_LM$task, levels=c("L","M"))4.1 Wariai LM
4.1.1 Check assumptions
dat_LM %>%
group_by(task, proficiency) %>%
identify_outliers(wariai)## # A tibble: 3 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 L lower 32 33 36 28 0.778 3.30 TRUE FALSE
## 2 L upper 34 35 70 45 0.643 3.80 TRUE FALSE
## 3 M upper 11 11 98 59 0.602 4.21 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat_LM %>%
group_by(task, proficiency) %>%
shapiro_test(wariai)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 L lower wariai 0.865 0.0664
## 2 L middle wariai 0.975 0.959
## 3 L upper wariai 0.977 0.968
## 4 M lower wariai 0.847 0.0392
## 5 M middle wariai 0.976 0.960
## 6 M upper wariai 0.826 0.0190ggqqplot(dat_LM, "wariai", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_LM %>%
group_by(task) %>%
levene_test(wariai ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 L 2 32 0.286 0.753
## 2 M 2 32 0.653 0.5274.1.2 Computation
res.aov <- anova_test(
data = dat_LM, dv = wariai, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 2.734 0.080 0.094
## 2 task 1 32 0.961 0.334 0.012
## 3 proficiency:task 2 32 0.084 0.920 0.002interaction.plot(x.factor = dat_LM$task, trace.factor = dat_LM$proficiency,
response = dat_LM$wariai, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
習熟度の主効果は有意傾向です。
4.2 Wariai2 LM
4.2.1 Check assumptions
dat_LM %>%
group_by(task, proficiency) %>%
identify_outliers(wariai2)## # A tibble: 2 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 L middle 7 7 29 21 0.724 2.76 TRUE FALSE
## 2 L middle 12 12 42 26 0.619 2.84 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat_LM %>%
group_by(task, proficiency) %>%
shapiro_test(wariai2)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 L lower wariai2 0.955 0.705
## 2 L middle wariai2 0.908 0.201
## 3 L upper wariai2 0.945 0.571
## 4 M lower wariai2 0.926 0.373
## 5 M middle wariai2 0.962 0.807
## 6 M upper wariai2 0.838 0.0265ggqqplot(dat_LM, "wariai2", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_LM %>%
group_by(task) %>%
levene_test(wariai2 ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 L 2 32 1.11 0.341
## 2 M 2 32 0.148 0.8634.2.2 Computation
res.aov <- anova_test(
data = dat_LM, dv = wariai2, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 2.127 0.136 0.087
## 2 task 1 32 6.472 0.016 * 0.054
## 3 proficiency:task 2 32 1.252 0.300 0.021interaction.plot(x.factor = dat_LM$task, trace.factor = dat_LM$proficiency,
response = dat_LM$wariai2, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
認知負荷の主効果は有意です。
4.3 koto LM
4.3.1 Check assumptions
dat_LM %>%
group_by(task, proficiency) %>%
identify_outliers(koto)## [1] task proficiency id trueid nobe koto
## [7] wariai wariai2 is.outlier is.extreme
## <0 rows> (or 0-length row.names)dat_LM %>%
group_by(task, proficiency) %>%
shapiro_test(koto)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 L lower koto 0.955 0.711
## 2 L middle koto 0.878 0.0814
## 3 L upper koto 0.980 0.982
## 4 M lower koto 0.922 0.335
## 5 M middle koto 0.943 0.537
## 6 M upper koto 0.933 0.410ggqqplot(dat_LM, "koto", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_LM %>%
group_by(task) %>%
levene_test(koto ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 L 2 32 1.71 0.197
## 2 M 2 32 1.60 0.2174.3.2 Computation
res.aov <- anova_test(
data = dat_LM, dv = koto, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 5.199 0.011 * 0.193
## 2 task 1 32 3.372 0.076 0.027
## 3 proficiency:task 2 32 1.787 0.184 0.028interaction.plot(x.factor = dat_LM$task, trace.factor = dat_LM$proficiency,
response = dat_LM$koto, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
習熟度の主効果は有意です。
5 Comparison between M and H
dat_MH <- dat %>% filter(task == "M"| task == "H")
dat_MH$task <- ordered(dat_MH$task, levels=c("M","H"))5.1 Wariai MH
5.1.1 Check assumptions
dat_MH %>%
group_by(task, proficiency) %>%
identify_outliers(wariai)## # A tibble: 1 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 M upper 11 11 98 59 0.602 4.21 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat_MH %>%
group_by(task, proficiency) %>%
shapiro_test(wariai)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 M lower wariai 0.847 0.0392
## 2 M middle wariai 0.976 0.960
## 3 M upper wariai 0.826 0.0190
## 4 H lower wariai 0.880 0.104
## 5 H middle wariai 0.954 0.692
## 6 H upper wariai 0.949 0.623ggqqplot(dat_MH, "wariai", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_MH %>%
group_by(task) %>%
levene_test(wariai ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 M 2 32 0.653 0.527
## 2 H 2 32 0.813 0.4525.1.2 Computation
res.aov <- anova_test(
data = dat_MH, dv = wariai, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 2.015 0.150 0.079
## 2 task 1 32 3.378 0.075 0.033
## 3 proficiency:task 2 32 0.096 0.909 0.002interaction.plot(x.factor = dat_MH$task, trace.factor = dat_MH$proficiency,
response = dat_MH$wariai, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
認知負荷の主効果は有意傾向です。
5.2 Wariai2 MH
5.2.1 Check assumptions
dat_MH %>%
group_by(task, proficiency) %>%
identify_outliers(wariai2)## [1] task proficiency id trueid nobe koto
## [7] wariai wariai2 is.outlier is.extreme
## <0 rows> (or 0-length row.names)dat_MH %>%
group_by(task, proficiency) %>%
shapiro_test(wariai2)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 M lower wariai2 0.926 0.373
## 2 M middle wariai2 0.962 0.807
## 3 M upper wariai2 0.838 0.0265
## 4 H lower wariai2 0.906 0.219
## 5 H middle wariai2 0.938 0.476
## 6 H upper wariai2 0.868 0.0622ggqqplot(dat_MH, "wariai2", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_MH %>%
group_by(task) %>%
levene_test(wariai2 ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 M 2 32 0.148 0.863
## 2 H 2 32 0.0157 0.9845.2.2 Computation
res.aov <- anova_test(
data = dat_MH, dv = wariai2, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 2.182 0.129 0.100
## 2 task 1 32 1.643 0.209 0.009
## 3 proficiency:task 2 32 0.689 0.509 0.008interaction.plot(x.factor = dat_MH$task, trace.factor = dat_MH$proficiency,
response = dat_MH$wariai2, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
5.3 koto MH
5.3.1 Check assumptions
dat_MH %>%
group_by(task, proficiency) %>%
identify_outliers(koto)## # A tibble: 1 × 10
## task proficiency id trueid nobe koto wariai wariai2 is.outlier is.extr…¹
## <ord> <ord> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <lgl> <lgl>
## 1 H lower 20 20 149 65 0.436 3.77 TRUE FALSE
## # … with abbreviated variable name ¹is.extremedat_MH %>%
group_by(task, proficiency) %>%
shapiro_test(koto)## # A tibble: 6 × 5
## task proficiency variable statistic p
## <ord> <ord> <chr> <dbl> <dbl>
## 1 M lower koto 0.922 0.335
## 2 M middle koto 0.943 0.537
## 3 M upper koto 0.933 0.410
## 4 H lower koto 0.872 0.0832
## 5 H middle koto 0.904 0.177
## 6 H upper koto 0.888 0.111ggqqplot(dat_MH, "koto", ggtheme = theme_bw()) +
facet_grid(task ~ proficiency)
dat_MH %>%
group_by(task) %>%
levene_test(koto ~ proficiency)## # A tibble: 2 × 5
## task df1 df2 statistic p
## <ord> <int> <int> <dbl> <dbl>
## 1 M 2 32 1.60 0.217
## 2 H 2 32 0.438 0.6495.3.2 Computation
res.aov <- anova_test(
data = dat_MH, dv = koto, wid = id,
between = proficiency, within = task
)
get_anova_table(res.aov)## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 proficiency 2 32 5.222 0.011 * 0.216
## 2 task 1 32 0.239 0.629 0.001
## 3 proficiency:task 2 32 2.123 0.136 0.020interaction.plot(x.factor = dat_MH$task, trace.factor = dat_MH$proficiency,
response = dat_MH$koto, fun = mean,
type = "b", legend = TRUE, trace.label = "TASK")
習熟度の主効果は有意です。
6 まとめ
6.1 LとM（外在性負荷の有無）
発話の語彙の多様さの結果は、
延べ語数に対する異なり語数の割合では、習熟度の主効果は有意傾向でした。
延べ語数 x2の平方根に対する異なり語数の割合では、認知負荷の主効果は有意でした。
異なる語数では、習熟度の主効果は有意でした。
6.2 MとH（内在性負荷の有無）
延べ語数に対する異なり語数の割合では、認知負荷の主効果は有意傾向でした。
延べ語数 x2の平方根に対する異なり語数の割合では、有意な効果はなかった。
異なる語数では、習熟度の主効果は有意でした。