### Data Preparation
#### Load Relevant Libraries and Functions
library(tidyverse)
library(readr)
library(lme4)
# Import data
folder_path <- "../data/complete"
csv_files <- list.files(folder_path, pattern = "*.csv", full.names = TRUE)
df_list <- lapply(csv_files, read_csv, show_col_types = FALSE)
original_df <- bind_rows(df_list)
#exclude practice trials
original_df <- original_df |>
filter(exp_part == "actual")
# convert the data type and compute correctness of each trial
original_df <- original_df %>%
mutate(correct_response = as.character(correct_response)) %>%
mutate(response = as.character(response)) %>%
mutate(correct = correct_response == response)
# compute the accuracy of each participant
accuracy_table <- original_df %>%
group_by(ID) %>%
summarise(accuracy = mean(correct, na.rm = TRUE))Replication of Study ‘What makes words special? Workds as unmotivated cues (Experiment 1A)’ by Edmiston & Lupyan (2015, Cognition)
Introduction
The current study is a replication of Experiment 1A of Edmiston & Lupyan (2015), which examines the differences in which verbal labels and environmental sounds activate categories in the mind.
Previous research suggests that environmental sounds, such as a dog’s bark, activate specific mental representations of a concept (Lupyan & Thompson-Schill, 2012). For instance, a loud, deep bark might evoke the image of a large dog (e.g., a Rottweiler), whereas a high-pitched bark may bring to mind a small dog (e.g., a Yorkie). In contrast, verbal labels like the word “dog” elicit broader mental representations that encompass knoweldge of multiple instances within a category. Because verbal labels do not inherently activate instance-specific qualities, they exhibit a “label advantage,” enabling listeners to identify and associate verbal labels with their referents more quickly and efficiently than environmental sounds.
Experiment 1A of Edmiston and Lupyan’s 2015 is a picture verification task. Participants were presented with an auditory cue (verbal label or environmental sound) followed by an image. They were then asked to respond whether the auditory cue and image matched. Analysis of reaction times (RTs) illustrates that verbal labels are more effective than environmental sounds in facilitating category activation. Moreover, when environmental sounds are paired with incongruent images (e.g., a deep bark paired with an image of a Yorkie), the activation of conceptual knowledge is impeded, as evidenced by slower reaction times.
In the current study, we aim to replicate Experiment 1A as closely as possible to determine whether these findings replicate and generalize to a new sample of participants.
Methods
Power Analysis
Based on guidance from instructional staff, the sample size was determined with an a-priori power analysis with the package simr, and is adequate to achieve at least 80% power for detecting the effect reported in the original study at a significance criterion of alpha = .05 (any random effects not specified in the original paper were taken from a small pilot study).
We plan to obtain sample of n=50 participants (near the original sample size n=43, including seven more participants to account for online uncertainty).
Planned Sample
In the original study, the participants for Experiment 1A consisted of n=43 undergraduates. For the current study, as justified above, we plan to obtain a sample of size n=50.
Participants will be recruited and compensated through Prolific, and will be pre-screened to ensure that they are fluent English-speaking adults with no reported hearing difficulties.
Materials
From the original study:
“The auditory cues comprised basic-level category labels and environmental sounds for six categories: bird, dog, drum, guitar, motorcycle, and phone. For each category, we obtained two distinct environmental sound cues, e.g.,
We will use the same materials for the current replication study, as all materials were provided by the original authors.
Procedure
From the original study:
“On each trial participants heard a cue and saw a picture. We instructed participants to decide as quickly and accurately as possible if the picture they saw came from the same basic-level category as the word or sound they heard Participants were tested in individual rooms sitting approximately 2400 from a monitor such that images subtended 10 10°. Trials began with a 250 ms. fixation cross followed immediately by the auditory cue, delivered via headphones. The target image appeared centrally 1 s after the off- set of the auditory cue and remained visible until a response was made. Each participant completed 6 practice and 384 test trials. If the picture matched the auditory cue (50% of trials) participants were instructed to respond ‘Yes’ on a gaming controller (e.g.,
We aim to follow the original procedure of Experiment 1A as precisely as possible. However, instead of presenting the trials in-person in a lab setting, the experiment will be conducted online using jsPsych hosted on Prolific.
For this reason, the overall task will be slightly different. Participants will be asked to respond to trials using keyboard keys (“F” indicating “Yes,”J” indicating “No”) in lieu of a gaming controller. Although we are unable to enforce the experimental environment, we will introduce an audio check prior to trials and encourage participants to wear headphones to ensure that they can adequately hear all auditory cues.
Analysis Plan
From the original study:
“All participants performed very accurately on all items (M = 97%). Response times (RTs) shorter than 250 ms. or longer than 1500 ms. were removed (292 trials removed, 1.77% of total). We fit RTs for correct responses on matching trials (‘Yes’ responses) with linear mixed regression using maximum likelihood estimation (Bates, Maechler, Bolker, & Walker, 2013), including random intercepts and random slopes for within-subject factors and random intercepts for repeated items (unique trial types) following the recommendations of Barr, Levy, Scheepers, and Tily (2013). Reported below are the parameter estimates (b) and confidence intervals for each contrast of interest. Significance tests were calculated using chi-square tests that compared nested models—models with and without the factor of interest—on improvement in log-likelihood.” (pp. 94-95)
For our replication, we will follow the original analysis as closely as possible. We will exclude participants who fall below a 90% accuracy threshold, as well as trials where RTs are shorter than 250ms and longer than 1500ms. We will then replicate the linear mixed effect model to examine the same main effects and random structure. Finally, we will perform chi-square tests that compared the models with and without the main effect using log-likelihood.
We anticipate that the results will successfully replicate. We should find that verbal labels elicit the lowest overall reaction times, and that congruent sounds elicit lower reaction times than incongruent sounds.
Clarify key analysis of interest here: Linear mixed regression model, chi-square tests of nested models with and without the factor of interest.
Differences from Original Study
This replication will differ slightly from the original study in terms of setting and procedure. As previously mentioned, the experiments will be conducted online using jsPsych as opposed to in-person in a lab. Loading delays, internet speed, and other technical difficulties encountered on participants’ personal devices may affect overall reaction times. Participants will also be using keyboard keys to respond to trials, as opposed to responding on a gaming controller. Despite these differences, we do not anticipate that these differences will significantly affect results or ability to obtain the same effect as described in the original study.
Design Review
In our analysis, one factor is directly manipulated: the auditory cue that was presented to participants. There are three conditions of that factor: verbal label, congruent environmental sound, and incongruent environmental sound.
There is one measure taken: reaction time (RT).
This study uses a within-participants design.
Measures were repeated; there were 384 experimental trials in the original experiment.
It would not make sense for the experiment to be changed to between-participants because the measures would not be as accurate. There will be individual differences of RTs regardless of experimental conditions.
No known steps were taken to reduce demand characteristics.
The original authors should have added attention checks between trials to ensure participants are attentive throughout trials. They could have also kept “No” responses. Incongruent sounds may have prompted participants to respond “No”, but these are still valid responses.
The original study involved undergraduate participants, which raises generalizability issues because they are inherently a WEIRD population.
Methods Addendum (Post Data Collection)
Actual Sample
The actual sample size obtained is n=50. While the original study consisted of n=43 participants, we included seven more participants to account for online uncertainty. Participants were pre-screened to ensure that they are adult English speakers without hearing difficulties.
In our replication experiment, only 84% of participants performed at an accuracy of above 90% (compared to 97% in the original study). Participants who fell below this accuracy threshold were excluded. Trials where reaction times fell below 250ms and exceeded 1500ms were also excluded.
42 of 50 participants performed with at least 90% accuracy, and were thus included in our confirmatory analysis. In total, 404 trials were removed (0.05% of total).
Differences from pre-data collection methods plan
NA
Results
Data preparation
As proposed in the above analysis plan, we will exclude participants who perform below 90% accuracy on the overall task, as well as trials where recorded reaction times fall below 250ms or over 1500ms.
First, we preprocess the data. This involves:
Importing the data,
Excluding practice trials,
Labeling trials as correct or incorrect,
Calculating overall accuracy for each participant
Next, we check the distribution of accuracy.
library(ggplot2)
# set the accuracy filtering threshold
accuracy_thresh <- 0.9
# plot the distribution of accuract
ggplot(accuracy_table, aes(x = accuracy)) +
geom_histogram(
breaks=seq(0, 1, length.out=21),
fill = "blue", color = "black", alpha = 0.7) +
geom_vline(xintercept = accuracy_thresh, color = "red", linetype = "dashed", size = 1) + # mark the threshold for QA
labs(
title = "Distribution of Accuracy",
x = "Accuracy",
y = ""
) +
theme_minimal()
# Number of participants that perform above 90% accuracy threshold
print(sum(accuracy_table$accuracy > accuracy_thresh))[1] 42# Percent of total sample
print((sum(accuracy_table$accuracy > accuracy_thresh))/50)[1] 0.8442 of 50 participants completed the experimental task with at least 90% accuracy. This is 84% of our total sample. This is notably lower than the accuracy obtained in the original study (97% of total sample).
Next, we check the distribution of response times (RTs) across experimental trials.
library(ggplot2)
# get the list of IDs of participants whose accuracy is above the threshold
id_pass_qa <- accuracy_table %>%
filter(accuracy > accuracy_thresh) %>%
pull(ID)
# filter the original dataframe to keep only data from the qualified participants
original_df_with_good_subj <- original_df %>%
filter(ID %in% id_pass_qa)
# plot the distribution of response time
ggplot(original_df_with_good_subj, aes(x = rt)) +
geom_histogram(
breaks=seq(0, 2000, length.out=40),
fill = "blue", color = "black", alpha = 0.7) +
geom_vline(xintercept = 250, color = "red", linetype = "dashed", size = 1) +
geom_vline(xintercept = 1500, color = "red", linetype = "dashed", size = 1) +
labs(
title = "Distribution of Reaction Time",
x = "Reaction Time",
y = ""
) +
theme_minimal()
# compute the quantile
ecdf_func <- ecdf(original_df_with_good_subj$rt)
rt_quantiles <- ecdf_func(c(250, 1500))
# compute the mask for whether each trial has passed the check
rt_pass_mask <- (original_df_with_good_subj$rt > 250) & (original_df_with_good_subj$rt < 1500)# Total number of trials that are within 250ms and 1500ms
print(sum(rt_pass_mask))[1] 15083# Percent of total trials
print(round(mean(rt_pass_mask), 2))[1] 0.9415083 trials have RTs that fall between 250ms and 1500ms. This is 94% of total trials collected.
Finally, we prepare the data for confirmatory analysis according to our analysis plan:
Exclude participants who do not pass the accuracy threshold of 90%,
Filter the trials to only include correct “Yes” trials (where participants correctly responded to matching trials),
Re-label experimental conditions (i.e.
label,congruent,incongruent),Exclude trials where RTs are below 250ms and above 1500ms.
# keep only people who passed the QA check
combined_df <- original_df_with_good_subj
# exclude only correct 'YES' trials
combined_df <- combined_df %>%
filter(response == "f") %>%
filter(correct_response == "f")
prepare_fitting_ready_df <- function(df) {
# rename "sound_subtype" to "cue"
df <- df %>%
rename(cue = sound_subtype)
# create congruency column
df <- df %>%
mutate(congruency = case_when(
correct == FALSE ~ NA_character_,
cue == "label" ~ "label",
(img_subtype %in% c("song", "york", "bongo", "acoustic", "harley", "rotary") & sound_version == "A") ~ "incongruent",
TRUE ~ "congruent"
))
# drop the correctness column
df <- df %>%
select(-correct)
# filter reaction time
df <- df %>%
filter(rt >250, rt <1500)
# Prepare data for analysis - keep only columns relevant to the model.
df <- df %>%
select(rt, ID, sound_category,cue, congruency)
return(df)
}
combined_df <- prepare_fitting_ready_df(combined_df)Confirmatory analysis
Following the confirmatory analysis of Edmiston & Lupyan (2015), we will perform a linear mixed effects regression model fit on response times on correct matching trials (trials where participants correctly responded “Yes” to matching trials). The model included random intercepts and random slopes for within-subjects faactors (subject) and random intercepts for repeated items (sound_category). The main effect investigated in this model is congruency, which denotes whether a particular trial was presented with a label, congruent, or incongruent sound. We then perform chi-square test to examine whether congruency significantly affects the log-likelihood. The resulting chi-square test values and p-values are reported.
The lmer model assesses how auditory cue type (label, congruent, or incongruent) affects reaction time in milliseconds. We implement a pairwise comparison to assess the relationship between label and congruent conditions. Comparison between nested models (with and without the factor of interest) is also implemented as described in the original study to examine whether manipulation of auditory cues truly motivates the observed effects.
The figure below shows the results of our replication (right), depicting average reaction time across the three conditions. It is presented side-by-side with the results of the original study (left, Experiment 1A) for comparison.
#load libraries:
library(ggplot2)
library(gridExtra)
Attaching package: 'gridExtra'The following object is masked from 'package:dplyr':
combinelibrary(jpeg)
library(grid)
# rename and reorder some columns for plotting
plot_df <- combined_df |>
mutate(congruency=factor(
congruency, levels = c("label", "congruent", "incongruent"))) |>
mutate(congruency = fct_recode(congruency,
"Label"="label",
"Congruent Sound"="congruent",
"Incongruent Sound"="incongruent")
)
# plot the result of our replication
our_plot <- ggplot(data = plot_df,
mapping = aes(x = congruency,
y = rt,
color = congruency,
fill = congruency)) +
geom_bar(stat = "summary", fun = "mean",
width = 1,
color = "black") +
stat_summary(fun.data = "mean_cl_boot",
geom = "linerange",
color = "black") +
geom_errorbar(stat = "summary",
fun.data = "mean_cl_boot",
width = 0.5,
size=0.5,
color = "black") +
scale_fill_manual(values = c("firebrick", "darkblue", "slategray2")) +
scale_color_manual(values = c("firebrick", "darkblue", "slategray2")) +
labs(
x = "Experiment 1A",
y = "Verification Speed (ms)",
color = "Cue Type",
fill = "Cue Type") +
coord_cartesian(xlim=c(0, 4), ylim = c(400, 725), expand=FALSE) +
scale_y_continuous(breaks = seq(400, 700, by = 50), expand=c(0, 25)) +
theme_classic() +
theme(
axis.text.x = element_blank(),
axis.text.y = element_text(size = 10, color = "black"),
axis.title.x = element_text(size = 10, color = "black"),
axis.title.y = element_text(size = 10, color = "black"),
legend.title = element_text(size = 10, color = "black", face='bold'),
legend.text = element_text(size = 10, color = "black"),
plot.margin = unit(c(0.0, 0.1, 0.0, 0.1), "npc")
) +
theme(
aspect.ratio = 1.8)
# load the plot from the original study
edmiston_img <- rasterGrob(readJPEG("edmiston_exp1a.jpg"), interpolate = TRUE)
# display side by side
plot_title <- textGrob("Our Result", gp = gpar(fontface = "bold", fontsize = 14))
image_title <- textGrob("Original Study Result", gp = gpar(fontface = "bold", fontsize = 14))
grid.arrange(
arrangeGrob(image_title, edmiston_img, ncol = 1, heights = c(0.5, 5)),
arrangeGrob(plot_title, ggplotGrob(our_plot), ncol = 1, heights = c(0.5, 5)),
ncol = 2, widths=c(1, 1.7))
We then fit the linear mixed effects regression model on reaction times and calculate the associated 95% confidence intervals.
#load libraries
library(lmerTest)
library(emmeans)
library(lme4)
#find best optimizer
#allFit(model_full)
#set up model and view results
model_full <- lmer(rt ~ congruency + (1 + congruency|ID) + (1|sound_category), data = combined_df, control = lmerControl(optimizer = "bobyqa"))
summary(model_full)Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: rt ~ congruency + (1 + congruency | ID) + (1 | sound_category)
Data: combined_df
Control: lmerControl(optimizer = "bobyqa")
REML criterion at convergence: 99407.8
Scaled residuals:
Min 1Q Median 3Q Max
-2.3748 -0.6827 -0.1974 0.4509 4.2050
Random effects:
Groups Name Variance Std.Dev. Corr
ID (Intercept) 18945.75 137.644
congruencyincongruent 822.76 28.684 -0.10
congruencylabel 1768.00 42.048 -0.46 0.30
sound_category (Intercept) 16.17 4.021
Residual 44351.37 210.598
Number of obs: 7328, groups: ID, 42; sound_category, 6
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 690.928 21.686 41.063 31.861 < 2e-16 ***
congruencyincongruent 15.047 9.365 41.783 1.607 0.116
congruencylabel -52.267 8.390 39.661 -6.229 2.33e-07 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Correlation of Fixed Effects:
(Intr) cngrncyn
cngrncyncng -0.126
congrncylbl -0.439 0.317 #get 95% CI
confint.merMod(model_full,method="Wald") 2.5 % 97.5 %
.sig01 NA NA
.sig02 NA NA
.sig03 NA NA
.sig04 NA NA
.sig05 NA NA
.sig06 NA NA
.sig07 NA NA
.sigma NA NA
(Intercept) 648.424610 733.43226
congruencyincongruent -3.307661 33.40178
congruencylabel -68.712145 -35.82268Reported reaction times for the congruent condition is 690.93ms on average, (95% confidence interval, [648.42, 733.43]). Average reaction time for the label condition, is shorter on average compared to the congruent condition by -52.27ms. This is consistent with our hypothesis. Average reaction time for the incongruent condition is indeed greater than the congruent condition, however this difference is not statistically significant.
The following post-hoc analysis report allows us to examine all relationships: congruent - incongruent, congruent - label, and incongruent - label. The 95% confidence interval for each comparison is reported.
#post-hoc test:
fit_result <- model_full %>%
emmeans(
pairwise ~ congruency, # contrast across 'congruency' conditions
adjust = "bonferroni" # p-value adjusted for multiple comparisons
) %>%
pluck("contrasts")
fit_result contrast estimate SE df z.ratio p.value
congruent - incongruent -15.0 9.36 Inf -1.607 0.3243
congruent - label 52.3 8.39 Inf 6.229 <.0001
incongruent - label 67.3 10.40 Inf 6.470 <.0001
Degrees-of-freedom method: asymptotic
P value adjustment: bonferroni method for 3 tests ci <- confint(fit_result, level=0.95)
ci contrast estimate SE df asymp.LCL asymp.UCL
congruent - incongruent -15.0 9.36 Inf -37.5 7.37
congruent - label 52.3 8.39 Inf 32.2 72.35
incongruent - label 67.3 10.40 Inf 42.4 92.22
Degrees-of-freedom method: asymptotic
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 3 estimates Finally, we conduct chi-square tests to examine significance of observed effects.
#create model without factor of interest
model_reduced <- lmer(rt ~ (1 + congruency|ID) + (1|sound_category), data = combined_df)
#run an anova to get chi-squared values
anova(model_full, model_reduced)Data: combined_df
Models:
model_reduced: rt ~ (1 + congruency | ID) + (1 | sound_category)
model_full: rt ~ congruency + (1 + congruency | ID) + (1 | sound_category)
npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
model_reduced 9 99481 99543 -49731 99463
model_full 11 99450 99526 -49714 99428 34.967 2 2.553e-08 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The results of the confirmatory analysis suggest that, as expected, verbal labels elicit faster reaction times than congruent environmental sounds (b = -52.3, CI [-68.71, -35.82]) and incongruent environmental sounds ((b = 67.3, CI [42.4, 92.22]). However, congruent trials were not significantly faster than incongruent trials (b = 15.04, CI [-3.31 33.4]). Finally, the chi-square test demonstrates that model with the factor of interest (congruency) outperforms the model without it. The full model with the congruency factor explains the data more successfully than the model without congruency.
Overall, the findings from the original Edmiston & Lupyan (2015) only partially replicated. Verbal labels did indeed elicit faster reaction times than environmental sounds, however we did not a statistically significant effect between congruent and incongruent environmental sounds.
Exploratory analyses
The box plots below shows the distribution of RTs for each individual trial for each condition. This gives insight possible biases in our reported measures that may have affected the results of our replication.
plot_df <- combined_df |>
mutate(congruency=factor(
congruency, levels = c("label", "congruent", "incongruent"))) |>
mutate(congruency = fct_recode(congruency,
"Label"="label",
"Congruent Sound"="congruent",
"Incongruent Sound"="incongruent")
)
ggplot(data = plot_df,
mapping = aes(x = congruency,
y = rt,
fill = congruency)) +
geom_boxplot(width = 0.6, outlier.shape = NA, alpha=0.5) +
geom_jitter(shape = 21, color = "black", width = 0.2, alpha = 0.2, size = 1.5) +
geom_hline(yintercept = 1200, color = "red", linetype = "dashed", size = 1) +
scale_fill_manual(values = c("firebrick", "darkblue", "slategray")) +
labs(
y = "Reaction Time (Ms)",
color = "Cue Type",
x = "Cue Type",
fill = "Cue Type "
) +
theme_classic() +
theme(
axis.text.x = element_blank(),
axis.text.y = element_text(size = 14, color = "black"),
axis.title.x = element_text(size = 16, color = "black", face = "bold"),
axis.title.y = element_text(size = 16, color = "black", face = "bold"),
legend.title = element_text(size = 14, color = "black"),
legend.text = element_text(size = 14, color = "black")
)
The distribution reveals that there is no clearly visible distinction between reaction times for congruent trials and incongruent trials. In fact, there are many congruent trials with longer RTs (greater than 1200ms). These data points pull the average RT for the congruent condition towards a higher value, which explains why we are unable to observe a statistically significant different between congruent and incongruent conditions.
Since the original study excluded incorrect trials (more specifically, trials where the sound and image matched, but participants responded “No”), we were curious to see whether the inclusion of this data will yield different results. To do this, we fit another linear mixed effect regression model and conducted post-hoc analysis with the inclusion of incorrect trials.
# include wrong trials
include_wrong_df <- original_df_with_good_subj %>%
filter(correct_response=='f')
include_wrong_df <- prepare_fitting_ready_df(include_wrong_df)
# lmer
# set reference level to congruent
include_wrong_df$congruency <- factor(
include_wrong_df$congruency, levels = c("congruent", "label", "incongruent"))
model_wrong_full <- lmer(rt ~ congruency + (1 + congruency|ID) + (1|sound_category), data = include_wrong_df)
# summary(model_wrong_full)
# EMM contrast
include_wrong_fit_result <- model_wrong_full %>%
emmeans(
pairwise ~ congruency, # contrast across 'congruency' conditions
adjust = "bonferroni", # p-value adjusted for multiple comparisons
) %>%
pluck("contrasts")
include_wrong_fit_result contrast estimate SE df z.ratio p.value
congruent - label 52.3 8.39 Inf 6.229 <.0001
congruent - incongruent -15.0 9.36 Inf -1.607 0.3243
label - incongruent -67.3 10.40 Inf -6.470 <.0001
Degrees-of-freedom method: asymptotic
P value adjustment: bonferroni method for 3 tests confint(include_wrong_fit_result, level=0.95) contrast estimate SE df asymp.LCL asymp.UCL
congruent - label 52.3 8.39 Inf 32.2 72.35
congruent - incongruent -15.0 9.36 Inf -37.5 7.37
label - incongruent -67.3 10.40 Inf -92.2 -42.41
Degrees-of-freedom method: asymptotic
Confidence level used: 0.95
Conf-level adjustment: bonferroni method for 3 estimates The results of this model reinforce our previous findings. RTs for the label conditions are significantly faster than congruent (p<.0001), however RTs for congruent trials are not significant faster than incongruent trials (p=0.3243).
Discussion
Summary of Replication Attempt
In our replication of Experiment 1A, we observed that on average:
There is statistically significant evidence that verbal labels elicit faster RTs than environmental sounds. This is consistent with results of the original study.
We do not have statistically significant evidence that congruent environment sounds elicit faster RTs than incongruent environmental sounds. This is inconsistent with the results of the original study.
Ultimately, based on the confirmatory analysis, our results only partially replicated the results of the original study.
Commentary
We posit that the inability to fully replicate the original results may stem from the differences in our experimental setting. As previously mentioned, the replication study was conducted online, which differed from the original study that was conducted in a lab. This may have posed many challenges. For instance, technical issues that could not be controlled for or reported (such as loading delays and slow internet speeds) could have affected our overall results. It was impossible to control participant attention in an online setting, especially since the experimental paradigm was relatively long in duration. Confusion or misunderstanding of the task, as well the inability to provide real-time feedback, may have also contributed to the discrepancies between our results and the original study.
References
- Lupyan, G., & Thompson-Schill, S. L. (2012). The evocative power of words: Activation of concepts by verbal and nonverbal means. Journal of Experimental Psychology-General, 141(1), 170–186. http://dx.doi.org/10.1037/a0024904.
Statement of Contributions
Jenna Brooks: Project Administration (lead), Software (JsPsych Experiment) (supporting), Writing and Editing (equal) Noah Khaloo: Formal Analysis (lead), Data Curation (lead), Statistical Analysis and Visualization (lead), Writing and Editing (equal) Sihan Yang: Conceptualization (lead), Software (JsPsych Experiment) (lead), Writing and Editing (equal) Reeka Estacio: Software (JsPsych Experiment) (supporting), Validation (lead), JsPsych Experiment (supporting), Presentation (lead), Writing and Editing (equal)