E1_TerminologyRecognition
Heping Xie
2024-07-10
Preparations in R
Install the packages if they are not already installed.
rm(list = ls())
options(scipen = 999)
if (!("pacman" %in% installed.packages())) install.packages("pacman")Load the R packages.
pacman::p_load(tidyverse, bruceR)Load and name the dataset.
dat <- bruceR::import("E1_TerminologyRecognition.csv")Check the data frame and for missings.
dim(dat)## [1] 3200 7colnames(dat)## [1] "Participant" "Item" "Tracing_type" "Cueing_type"
## [5] "Score" "Prior_knowledge" "Spatial_ability"str(dat)## 'data.frame': 3200 obs. of 7 variables:
## $ Participant : chr "A1003" "A1003" "A1003" "A1003" ...
## $ Item : chr "recognition1" "recognition2" "recognition3" "recognition4" ...
## $ Tracing_type : chr "yes" "yes" "yes" "yes" ...
## $ Cueing_type : chr "yes" "yes" "yes" "yes" ...
## $ Score : int 0 1 1 1 1 1 1 1 0 1 ...
## $ Prior_knowledge: int 16 16 16 16 16 16 16 16 16 16 ...
## $ Spatial_ability: int 6 6 6 6 6 6 6 6 6 6 ...summary(dat)## Participant Item Tracing_type Cueing_type
## Length:3200 Length:3200 Length:3200 Length:3200
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
## Score Prior_knowledge Spatial_ability
## Min. :0.000 Min. : 4.00 Min. : 0.000
## 1st Qu.:0.000 1st Qu.: 8.00 1st Qu.: 5.000
## Median :1.000 Median :12.00 Median : 8.000
## Mean :0.635 Mean :10.73 Mean : 8.613
## 3rd Qu.:1.000 3rd Qu.:14.00 3rd Qu.:12.000
## Max. :1.000 Max. :16.00 Max. :20.000head(dat)## Participant Item Tracing_type Cueing_type Score Prior_knowledge
## 1 A1003 recognition1 yes yes 0 16
## 2 A1003 recognition2 yes yes 1 16
## 3 A1003 recognition3 yes yes 1 16
## 4 A1003 recognition4 yes yes 1 16
## 5 A1003 recognition5 yes yes 1 16
## 6 A1003 recognition6 yes yes 1 16
## Spatial_ability
## 1 6
## 2 6
## 3 6
## 4 6
## 5 6
## 6 6Dataset for ANCOVA
dat_ANCOVA <- dat %>%
dplyr::filter(Cueing_type == "no") %>%
dplyr::group_by(Participant, Tracing_type) %>%
dplyr::summarise(Sum_score = sum(Score, na.rm = TRUE),
Prior_knowledge = mean(Prior_knowledge, na.rm = TRUE),
Spatial_ability = mean(Spatial_ability, na.rm = TRUE)) %>%
dplyr::ungroup() %>%
dplyr::mutate(
Spatial_ability = case_when(
Spatial_ability < mean(Spatial_ability, na.rm = TRUE) ~ "low",
Spatial_ability >= mean(Spatial_ability, na.rm = TRUE) ~ "high"
)
)## `summarise()` has grouped output by 'Participant'. You can override using the
## `.groups` argument.dat_ANCOVA %>%
dplyr::group_by(Tracing_type, Spatial_ability) %>%
dplyr::summarise(n = n()) %>%
bruceR::print_table(digits = 0)## `summarise()` has grouped output by 'Tracing_type'. You can override using the
## `.groups` argument.## ──────────────────────────────────
## Tracing_type Spatial_ability n
## ──────────────────────────────────
## 1 no high 18
## 2 no low 22
## 3 yes high 17
## 4 yes low 23
## ──────────────────────────────────ANCOVA
m <- bruceR::MANOVA(dat_ANCOVA, dv = "Sum_score", between = c("Tracing_type", "Spatial_ability"), covariate = "Prior_knowledge")## Warning: Numerical variables NOT centered on 0 (matters if variable in interaction):
## Prior_knowledge##
## ====== ANOVA (Between-Subjects Design) ======
##
## Descriptives:
## ───────────────────────────────────────────────────
## "Tracing_type" "Spatial_ability" Mean S.D. n
## ───────────────────────────────────────────────────
## no high 12.500 (3.899) 18
## no low 12.409 (3.112) 22
## yes high 15.353 (2.957) 17
## yes low 12.826 (3.284) 23
## ───────────────────────────────────────────────────
## Total sample size: N = 80
##
## ANOVA Table:
## Dependent variable(s): Sum_score
## Between-subjects factor(s): Tracing_type, Spatial_ability
## Within-subjects factor(s): –
## Covariate(s): Prior_knowledge
## ──────────────────────────────────────────────────────────────────────────────────────────────
## MS MSE df1 df2 F p η²p [90% CI of η²p] η²G
## ──────────────────────────────────────────────────────────────────────────────────────────────
## Tracing_type 50.350 11.161 1 75 4.511 .037 * .057 [.002, .160] .057
## Spatial_ability 35.623 11.161 1 75 3.192 .078 . .041 [.000, .136] .041
## Prior_knowledge 1.918 11.161 1 75 0.172 .680 .002 [.000, .050] .002
## Tracing_type * Spatial_ability 29.868 11.161 1 75 2.676 .106 .034 [.000, .126] .034
## ──────────────────────────────────────────────────────────────────────────────────────────────
## MSE = mean square error (the residual variance of the linear model)
## η²p = partial eta-squared = SS / (SS + SSE) = F * df1 / (F * df1 + df2)
## ω²p = partial omega-squared = (F - 1) * df1 / (F * df1 + df2 + 1)
## η²G = generalized eta-squared (see Olejnik & Algina, 2003)
## Cohen’s f² = η²p / (1 - η²p)
##
## Levene’s Test for Homogeneity of Variance:
## ───────────────────────────────────────────
## Levene’s F df1 df2 p
## ───────────────────────────────────────────
## DV: Sum_score 0.382 3 76 .766
## ───────────────────────────────────────────emmeans::emmip(m, Tracing_type ~ Spatial_ability, CIs = TRUE)
emmeans::emmip(m, Spatial_ability ~ Tracing_type, CIs = TRUE)
Simple-effect analyses
bruceR::EMMEANS(m, effect = "Tracing_type", by = "Spatial_ability")## ------ EMMEANS (effect = "Tracing_type") ------
##
## Joint Tests of "Tracing_type":
## ──────────────────────────────────────────────────────────────────────────────
## Effect "Spatial_ability" df1 df2 F p η²p [90% CI of η²p]
## ──────────────────────────────────────────────────────────────────────────────
## Tracing_type high 1 75 6.314 .014 * .078 [.009, .189]
## Tracing_type low 1 75 0.138 .711 .002 [.000, .047]
## Prior_knowledge high 1 75 0.172 .680 .002 [.000, .050]
## Prior_knowledge low 1 75 0.172 .680 .002 [.000, .050]
## ──────────────────────────────────────────────────────────────────────────────
## Note. Simple effects of repeated measures with 3 or more levels
## are different from the results obtained with SPSS MANOVA syntax.
##
## Estimated Marginal Means of "Tracing_type":
## ─────────────────────────────────────────────────────────────────
## "Tracing_type" "Spatial_ability" Mean [95% CI of Mean] S.E.
## ─────────────────────────────────────────────────────────────────
## no high 12.548 [10.963, 14.134] (0.796)
## yes high 15.388 [13.765, 17.012] (0.815)
## no low 12.399 [10.979, 13.819] (0.713)
## yes low 12.772 [11.360, 14.184] (0.709)
## ─────────────────────────────────────────────────────────────────
##
## Pairwise Comparisons of "Tracing_type":
## ───────────────────────────────────────────────────────────────────────────────────────
## Contrast "Spatial_ability" Estimate S.E. df t p Cohen’s d [95% CI of d]
## ───────────────────────────────────────────────────────────────────────────────────────
## yes - no high 2.840 (1.130) 75 2.513 .014 * 0.850 [ 0.176, 1.524]
## yes - no low 0.373 (1.002) 75 0.372 .711 0.112 [-0.486, 0.709]
## ───────────────────────────────────────────────────────────────────────────────────────
## Pooled SD for computing Cohen’s d: 3.341
## No need to adjust p values.
##
## Disclaimer:
## By default, pooled SD is Root Mean Square Error (RMSE).
## There is much disagreement on how to compute Cohen’s d.
## You are completely responsible for setting `sd.pooled`.
## You might also use `effectsize::t_to_d()` to compute d.bruceR::EMMEANS(m, effect = "Spatial_ability", by = "Tracing_type")## ------ EMMEANS (effect = "Spatial_ability") ------
##
## Joint Tests of "Spatial_ability":
## ───────────────────────────────────────────────────────────────────────────
## Effect "Tracing_type" df1 df2 F p η²p [90% CI of η²p]
## ───────────────────────────────────────────────────────────────────────────
## Spatial_ability no 1 75 0.019 .890 .000 [.000, .021]
## Spatial_ability yes 1 75 5.760 .019 * .071 [.007, .181]
## Prior_knowledge no 1 75 0.172 .680 .002 [.000, .050]
## Prior_knowledge yes 1 75 0.172 .680 .002 [.000, .050]
## ───────────────────────────────────────────────────────────────────────────
## Note. Simple effects of repeated measures with 3 or more levels
## are different from the results obtained with SPSS MANOVA syntax.
##
## Estimated Marginal Means of "Spatial_ability":
## ─────────────────────────────────────────────────────────────────
## "Spatial_ability" "Tracing_type" Mean [95% CI of Mean] S.E.
## ─────────────────────────────────────────────────────────────────
## high no 12.548 [10.963, 14.134] (0.796)
## low no 12.399 [10.979, 13.819] (0.713)
## high yes 15.388 [13.765, 17.012] (0.815)
## low yes 12.772 [11.360, 14.184] (0.709)
## ─────────────────────────────────────────────────────────────────
##
## Pairwise Comparisons of "Spatial_ability":
## ───────────────────────────────────────────────────────────────────────────────────────
## Contrast "Tracing_type" Estimate S.E. df t p Cohen’s d [95% CI of d]
## ───────────────────────────────────────────────────────────────────────────────────────
## low - high no -0.149 (1.071) 75 -0.139 .890 -0.045 [-0.683, 0.594]
## low - high yes -2.617 (1.090) 75 -2.400 .019 * -0.783 [-1.433, -0.133]
## ───────────────────────────────────────────────────────────────────────────────────────
## Pooled SD for computing Cohen’s d: 3.341
## No need to adjust p values.
##
## Disclaimer:
## By default, pooled SD is Root Mean Square Error (RMSE).
## There is much disagreement on how to compute Cohen’s d.
## You are completely responsible for setting `sd.pooled`.
## You might also use `effectsize::t_to_d()` to compute d.