This website contains R code used for the analyses in Park et al. (in press). The scripts were also posted through an online data repository at Center for Open Science and GitHub.
Park, J., Gremp, M., & Shin, M. (in press). Effects of interactive and hands-on learning experience of assistive technology on college students’ perspectives and knowledge. Education and Information Technologies.
suppressPackageStartupMessages({
library(stringr)
library(scales)
library(reshape2)
library(gtsummary)
library(flextable)
library(purrr)
library(ggeffects)
library(ggplot2)
library(dplyr)
library(ordinal)
library(texreg)
library(MASS)
library(nlme)
library(tidyr)
})
load("data/AT-college.RData")df$group <- as.factor(df$group)
demo_tbl <- df %>%
mutate(group = case_when(
group == "1" ~ "Treatment",
group == "0" ~ "Control"
)) %>%
filter(time == "0") %>%
dplyr::select(group, gender, year, college) %>%
tbl_summary(by = group,
percent = "column")
demo_tbl_ft <- as_flex_table(demo_tbl)demo_tbl_ft Characteristic | Control | Treatment |
|---|---|---|
gender | ||
female | 37 (74%) | 49 (77%) |
male | 9 (18%) | 12 (19%) |
not_reported | 0 (0%) | 1 (1.6%) |
third | 4 (8.0%) | 2 (3.1%) |
year | ||
freshman | 38 (76%) | 40 (63%) |
junior | 1 (2.0%) | 6 (9.4%) |
sophomore | 11 (22%) | 18 (28%) |
college | ||
business | 1 (2.0%) | 0 (0%) |
education_AHS | 39 (78%) | 39 (61%) |
health_science | 3 (6.0%) | 5 (7.8%) |
justice_safety_military | 0 (0%) | 1 (1.6%) |
letter_art_Social_Science | 6 (12%) | 16 (25%) |
STEM | 1 (2.0%) | 3 (4.7%) |
1n (%) | ||
reverse_thresholds <- function(data) {
sum <- summary(data)
cf <- as.data.frame(sum$coefficients)
cf[1:4, 1] <- cf[1:4, 1] * -1
cf[1:4, 3] <- cf[1:4, 3] * -1
rownames(cf)[1:4] <- c("intercept (Y>1)", "intercept (Y>2)", "intercept (Y>3)", "intercept (Y>4)")
return(cf)
}
reverse_thresholds_alt <- function(data) {
sum <- summary(data)
cf <- as.data.frame(sum$coefficients)
cf[1:3, 1] <- cf[1:3, 1] * -1
cf[1:3, 3] <- cf[1:3, 2] * -1
rownames(cf)[1:3] <- c("intercept (Y>2)", "intercept (Y>3)", "intercept (Y>4)")
return(cf)
}SP1
df$SP1 <- as.factor(df$SP1)
main.SP1 <- clmm(SP1 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP1)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 5.21 0.87 5.98 0.00 183.76
intercept (Y>2) 4.49 0.71 6.35 0.00 88.87
intercept (Y>3) 2.10 0.44 4.80 0.00 8.13
intercept (Y>4) 0.22 0.36 0.61 0.54 1.25
group1 -0.11 0.47 -0.23 0.82 0.90
time 0.49 0.43 1.13 0.26 1.64
group1:time 1.30 0.62 2.09 0.04 3.67SP2
df$SP2 <- as.factor(df$SP2)
main.SP2 <- clmm(SP2 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds_alt(main.SP2)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>2) 4.73 0.75 -0.75 0.00 113.77
intercept (Y>3) 1.85 0.41 -0.41 0.00 6.38
intercept (Y>4) -0.22 0.36 -0.36 0.55 0.81
group1 0.06 0.46 0.13 0.90 1.06
time 0.61 0.43 1.43 0.15 1.85
group1:time 1.10 0.60 1.84 0.07 3.01SP3
df$SP3 <- as.factor(df$SP3)
main.SP3 <- clmm(SP3 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP3)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 4.79 0.69 6.99 0.00 119.74
intercept (Y>2) 4.12 0.59 6.97 0.00 61.58
intercept (Y>3) 1.66 0.40 4.18 0.00 5.27
intercept (Y>4) 0.29 0.36 0.82 0.41 1.34
group1 -0.60 0.47 -1.28 0.20 0.55
time -0.27 0.41 -0.65 0.52 0.77
group1:time 1.41 0.58 2.44 0.01 4.09SP4
df$SP4 <- as.factor(df$SP4)
main.SP4 <- clmm(SP4 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP4)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 5.26 0.73 7.21 0.00 192.46
intercept (Y>2) 3.85 0.54 7.12 0.00 47.22
intercept (Y>3) 1.09 0.39 2.81 0.01 2.98
intercept (Y>4) -0.97 0.38 -2.52 0.01 0.38
group1 -0.18 0.49 -0.37 0.71 0.84
time -0.22 0.39 -0.56 0.57 0.80
group1:time 0.67 0.54 1.24 0.21 1.94SP5
df$SP5 <- as.factor(df$SP5)
main.SP5 <- clmm(SP5 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP5)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 4.46 0.60 7.49 0.00 86.58
intercept (Y>2) 1.38 0.37 3.73 0.00 3.96
intercept (Y>3) -1.05 0.36 -2.92 0.00 0.35
intercept (Y>4) -3.09 0.45 -6.82 0.00 0.05
group1 -0.38 0.45 -0.85 0.39 0.68
time -0.28 0.38 -0.74 0.46 0.75
group1:time 1.04 0.52 1.99 0.05 2.82SP6
df$SP6 <- as.factor(df$SP6)
main.SP6 <- clmm(SP6 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP6)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 5.66 0.75 7.52 0.00 286.62
intercept (Y>2) 3.86 0.57 6.76 0.00 47.60
intercept (Y>3) 1.15 0.44 2.61 0.01 3.16
intercept (Y>4) -0.76 0.43 -1.75 0.08 0.47
group1 -0.44 0.56 -0.80 0.43 0.64
time -0.05 0.40 -0.12 0.91 0.95
group1:time 0.55 0.55 1.00 0.32 1.73SP7
df$SP7 <- as.factor(df$SP7)
main.SP7 <- clmm(SP7 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP7)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 4.92 0.73 6.73 0.00 137.60
intercept (Y>2) 4.26 0.65 6.58 0.00 70.72
intercept (Y>3) 2.33 0.46 5.06 0.00 10.23
intercept (Y>4) 0.28 0.37 0.74 0.46 1.32
group1 -0.76 0.49 -1.55 0.12 0.47
time 0.25 0.42 0.61 0.54 1.29
group1:time 1.81 0.62 2.91 0.00 6.10SP8
df$SP8 <- as.factor(df$SP8)
main.SP8 <- clmm(SP8 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.SP8)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 3.10 0.51 6.10 0.00 22.22
intercept (Y>2) 0.01 0.39 0.04 0.97 1.01
intercept (Y>3) -4.02 0.59 -6.83 0.00 0.02
intercept (Y>4) -6.10 0.94 -6.49 0.00 0.00
group1 -0.25 0.51 -0.48 0.63 0.78
time 0.68 0.41 1.64 0.10 1.97
group1:time -1.17 0.56 -2.07 0.04 0.31screenreg(list(main.SP1, main.SP2, main.SP3, main.SP4, main.SP5, main.SP6, main.SP7, main.SP8),
custom.model.names = c("SP 1", "SP 2", "SP 3", "SP 4", "SP 5", "SP 6", "SP 7", "SP 8"))
======================================================================================================================================
SP 1 SP 2 SP 3 SP 4 SP 5 SP 6 SP 7 SP 8
--------------------------------------------------------------------------------------------------------------------------------------
group1 -0.11 0.06 -0.60 -0.18 -0.38 -0.44 -0.76 -0.25
(0.47) (0.46) (0.47) (0.49) (0.45) (0.56) (0.49) (0.51)
time 0.49 0.61 -0.27 -0.22 -0.28 -0.05 0.25 0.68
(0.43) (0.43) (0.41) (0.39) (0.38) (0.40) (0.42) (0.41)
group1:time 1.30 * 1.10 1.41 * 0.67 1.04 * 0.55 1.81 ** -1.17 *
(0.62) (0.60) (0.58) (0.54) (0.52) (0.55) (0.62) (0.56)
1|2 -5.21 *** -4.79 *** -5.26 *** -4.46 *** -5.66 *** -4.92 *** -3.10 ***
(0.87) (0.69) (0.73) (0.60) (0.75) (0.73) (0.51)
2|3 -4.49 *** -4.73 *** -4.12 *** -3.85 *** -1.38 *** -3.86 *** -4.26 *** -0.01
(0.71) (0.75) (0.59) (0.54) (0.37) (0.57) (0.65) (0.39)
3|4 -2.10 *** -1.85 *** -1.66 *** -1.09 ** 1.05 ** -1.15 ** -2.33 *** 4.02 ***
(0.44) (0.41) (0.40) (0.39) (0.36) (0.44) (0.46) (0.59)
4|5 -0.22 0.22 -0.29 0.97 * 3.09 *** 0.76 -0.28 6.10 ***
(0.36) (0.36) (0.36) (0.38) (0.45) (0.43) (0.37) (0.94)
--------------------------------------------------------------------------------------------------------------------------------------
Log Likelihood -205.36 -214.11 -252.19 -282.25 -301.08 -287.26 -228.83 -250.23
AIC 426.71 442.22 520.38 580.51 618.16 590.52 473.66 516.45
BIC 454.15 466.23 547.82 607.94 645.59 617.95 501.10 543.89
Num. obs. 228 228 228 228 228 228 228 228
Groups (student) 114 114 114 114 114 114 114 114
Variance: student: (Intercept) 1.65 1.82 1.82 2.77 2.05 4.36 2.30 3.17
======================================================================================================================================
*** p < 0.001; ** p < 0.01; * p < 0.05PK1
df$PK1 <- as.factor(df$PK1)
main.PK1 <- clmm(PK1 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK1)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 1.36 0.34 4.06 0.00 3.91
intercept (Y>2) -1.30 0.34 -3.86 0.00 0.27
intercept (Y>3) -3.86 0.49 -7.83 0.00 0.02
intercept (Y>4) -6.19 0.68 -9.06 0.00 0.00
group1 0.20 0.40 0.51 0.61 1.23
time 0.88 0.40 2.20 0.03 2.41
group1:time 3.02 0.59 5.08 0.00 20.43PK2
df$PK2 <- as.factor(df$PK2)
main.PK2 <- clmm(PK2 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK2)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 0.29 0.31 0.96 0.34 1.34
intercept (Y>2) -1.63 0.35 -4.68 0.00 0.20
intercept (Y>3) -3.60 0.45 -7.99 0.00 0.03
intercept (Y>4) -5.92 0.63 -9.45 0.00 0.00
group1 0.64 0.40 1.61 0.11 1.90
time 1.37 0.40 3.42 0.00 3.92
group1:time 2.05 0.55 3.74 0.00 7.73PK3
df$PK3 <- as.factor(df$PK3)
main.PK3 <- clmm(PK3 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK3)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 0.60 0.33 1.81 0.07 1.83
intercept (Y>2) -1.55 0.36 -4.28 0.00 0.21
intercept (Y>3) -3.75 0.49 -7.73 0.00 0.02
intercept (Y>4) -6.21 0.68 -9.15 0.00 0.00
group1 0.20 0.42 0.48 0.63 1.23
time 1.23 0.40 3.08 0.00 3.43
group1:time 2.65 0.57 4.64 0.00 14.11PK4
df$PK4 <- as.factor(df$PK4)
main.PK4 <- clmm(PK4 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK4)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 0.74 0.34 2.18 0.03 2.09
intercept (Y>2) -1.15 0.35 -3.31 0.00 0.32
intercept (Y>3) -3.13 0.44 -7.18 0.00 0.04
intercept (Y>4) -5.76 0.63 -9.10 0.00 0.00
group1 0.60 0.43 1.40 0.16 1.82
time 1.06 0.39 2.75 0.01 2.90
group1:time 1.93 0.54 3.56 0.00 6.89PK5
df$PK5 <- as.factor(df$PK5)
main.PK5 <- clmm(PK5 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK5)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 1.66 0.40 4.14 0.00 5.25
intercept (Y>2) -0.66 0.37 -1.81 0.07 0.51
intercept (Y>3) -2.99 0.46 -6.49 0.00 0.05
intercept (Y>4) -6.07 0.70 -8.68 0.00 0.00
group1 0.32 0.47 0.68 0.50 1.37
time 0.84 0.39 2.16 0.03 2.32
group1:time 2.95 0.59 4.98 0.00 19.18PK6
df$PK6 <- as.factor(df$PK6)
main.PK6 <- clmm(PK6 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK6)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 0.26 0.30 0.87 0.38 1.30
intercept (Y>2) -1.73 0.34 -5.02 0.00 0.18
intercept (Y>3) -3.90 0.47 -8.25 0.00 0.02
intercept (Y>4) -6.41 0.67 -9.54 0.00 0.00
group1 0.64 0.40 1.61 0.11 1.89
time 1.41 0.40 3.57 0.00 4.11
group1:time 2.29 0.56 4.11 0.00 9.84PK7
df$PK7 <- as.factor(df$PK7)
main.PK7 <- clmm(PK7 ~ group + time + group:time +
(1 | student), data = df, Hess=TRUE, nAGQ=7)cf <- reverse_thresholds(main.PK7)
cf %>% mutate(odd_ratio = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 2)) Estimate Std. Error z value Pr(>|z|) odd_ratio
intercept (Y>1) 1.27 0.38 3.34 0.00 3.55
intercept (Y>2) -1.39 0.38 -3.67 0.00 0.25
intercept (Y>3) -3.66 0.49 -7.42 0.00 0.03
intercept (Y>4) -6.29 0.70 -8.94 0.00 0.00
group1 0.84 0.46 1.81 0.07 2.31
time 1.13 0.40 2.80 0.01 3.09
group1:time 2.29 0.57 4.03 0.00 9.90screenreg(list(main.PK1, main.PK2, main.PK3, main.PK4, main.PK5, main.PK6, main.PK7),
custom.model.names = c("PK 1", "PK 2", "PK 3", "PK 4", "PK 5", "PK 6", "PK 7"))
=========================================================================================================================
PK 1 PK 2 PK 3 PK 4 PK 5 PK 6 PK 7
-------------------------------------------------------------------------------------------------------------------------
group1 0.20 0.64 0.20 0.60 0.32 0.64 0.84
(0.40) (0.40) (0.42) (0.43) (0.47) (0.40) (0.46)
time 0.88 * 1.37 *** 1.23 ** 1.06 ** 0.84 * 1.41 *** 1.13 **
(0.40) (0.40) (0.40) (0.39) (0.39) (0.40) (0.40)
group1:time 3.02 *** 2.05 *** 2.65 *** 1.93 *** 2.95 *** 2.29 *** 2.29 ***
(0.59) (0.55) (0.57) (0.54) (0.59) (0.56) (0.57)
1|2 -1.36 *** -0.29 -0.60 -0.74 * -1.66 *** -0.26 -1.27 ***
(0.34) (0.31) (0.33) (0.34) (0.40) (0.30) (0.38)
2|3 1.30 *** 1.63 *** 1.55 *** 1.15 *** 0.66 1.73 *** 1.39 ***
(0.34) (0.35) (0.36) (0.35) (0.37) (0.34) (0.38)
3|4 3.86 *** 3.60 *** 3.75 *** 3.13 *** 2.99 *** 3.90 *** 3.66 ***
(0.49) (0.45) (0.49) (0.44) (0.46) (0.47) (0.49)
4|5 6.19 *** 5.92 *** 6.21 *** 5.76 *** 6.07 *** 6.41 *** 6.29 ***
(0.68) (0.63) (0.68) (0.63) (0.70) (0.67) (0.70)
-------------------------------------------------------------------------------------------------------------------------
Log Likelihood -261.26 -285.13 -281.08 -298.32 -288.53 -274.59 -283.70
AIC 538.52 586.25 578.15 612.64 593.05 565.18 583.41
BIC 565.96 613.69 605.59 640.07 620.49 592.61 610.84
Num. obs. 228 228 228 228 228 228 228
Groups (student) 114 114 114 114 114 114 114
Variance: student: (Intercept) 0.57 0.54 1.12 1.25 2.19 0.47 1.83
=========================================================================================================================
*** p < 0.001; ** p < 0.01; * p < 0.05main.DK <- glmmPQL(DK ~ group + time + group:time,
random = ~ 1 | student, data = df, family=poisson)sum <- summary(main.DK)
cf <- sum$tTable %>% as.data.frame()
cf$Estimate <- cf$Value
cf %>% mutate(irr = exp(Estimate)) %>%
mutate_if(is.numeric, ~ round(., 3)) %>%
dplyr::select(-Value) %>%
dplyr::select(Estimate, everything()) Estimate Std.Error DF t-value p-value irr
(Intercept) 0.565 0.170 112 3.333 0.001 1.760
group1 -0.447 0.249 112 -1.797 0.075 0.640
time 0.393 0.181 112 2.176 0.032 1.481
group1:time 1.555 0.251 112 6.185 0.000 4.735VarCorr(main.DK)student = pdLogChol(1)
Variance StdDev
(Intercept) 0.3767362 0.6137884
Residual 2.0265750 1.4235782screenreg(list(main.DK),
custom.model.names = c("DK Main Effects"))
=====================================
DK Main Effects
-------------------------------------
(Intercept) 0.57 **
(0.17)
group1 -0.45
(0.25)
time 0.39 *
(0.18)
group1:time 1.55 ***
(0.25)
-------------------------------------
AIC
BIC
Log Likelihood
Num. obs. 228
Num. groups: student 114
=====================================
*** p < 0.001; ** p < 0.01; * p < 0.05