-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathDutch-modality-exclusivity-norms-RPubs.Rmd
4598 lines (3419 loc) · 234 KB
/
Dutch-modality-exclusivity-norms-RPubs.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
---
title: ' Dutch modality exclusivity norms'
author: '<a href="#section-info" style="color:#E4E5E8 !important; text-decoration:none !important; font-size:10px"> Bernabeu (2018) </a>'
output:
flexdashboard::flex_dashboard:
theme: 'spacelab'
orientation: columns
vertical_layout: scroll
favicon: https://i.ibb.co/bB9fCfr/norms-favicon.png
---
```{r packages, include = FALSE}
library(arules)
library(car)
library(contrast)
library(corpcor)
library(doBy)
library(plyr) # Load before `dplyr` (https://github.com/tidyverse/dplyr/issues/347)
library(dplyr)
library(flexdashboard)
library(formattable)
library(gdata)
library(ggplot2)
library(ggrepel)
library(GPArotation)
library(grid)
library(gridExtra)
library(gtools)
library(Hmisc)
library(htmltools)
library(irr)
library(kableExtra)
library(knitr)
library(lattice)
library(leaflet)
library(ltm)
library(MASS)
library(mnormt)
library(pander)
library(pastecs)
library(plotly)
library(png)
library(psych)
library(QuantPsyc)
library(RColorBrewer)
library(RCurl)
library(reactable)
library(reshape)
library(rmarkdown)
library(Rmisc)
library(rsconnect)
library(scales)
library(shiny)
library(shinyWidgets)
library(stringr)
library(tibble)
library(tidyr)
panderOptions('keep.trailing.zeros', TRUE)
knitr::opts_chunk$set(cache = FALSE, fig.align='center')
```
<!-- Begin definition of layout parameters -->
<head>
<style type="text/css">
<!-- Tabs format -->
/* inactive tab */
ul.nav.navbar-nav.navbar-left a {
color: white !important;
background-color: none !important;
border-bottom: none !important;
}
/* inactive tab hovered over */
ul.nav.navbar-nav.navbar-left a:hover {
color: white !important;
background-color: #34507D !important;
border-bottom: none !important;
text-shadow: none !important;
}
/* active tab */
ul.nav.navbar-nav.navbar-left li.active a {
color: white !important;
font-weight: bold !important;
background-color: #5472A3 !important;
border-bottom: none !important;
text-shadow: none !important;
}
/* visited links in active tabs */
ul.nav.navbar-nav.navbar-left li.active a:visited {
color: white !important;
font-weight: bold !important;
background-color: #5472A3 !important;
border-bottom: none !important;
text-shadow: none !important;
}
/* visited links in inactive tabs */
ul.nav.navbar-nav.navbar-left a:visited {
color: white !important;
font-weight: bold !important;
background-color: #5472A3 !important;
border-bottom: none !important;
text-shadow: none !important;
}
<!-- Links format -->
/* unvisited link */
section-properties, section-concepts, section-cf-lc-english-norms, section-sound-symbolism, section-info a {
color: #3C6CA7 !important;
border-bottom: 0.03px solid #5277A5 !important;
font-weight: normal !important;
background-color: none !important;
text-shadow: none !important;
}
/* visited link */
section-properties, section-concepts, section-cf-lc-english-norms, section-sound-symbolism, section-info a:visited {
color: #426DA1 !important;
border-bottom: none !important;
font-weight: normal !important;
background-color: none !important;
text-shadow: none !important;
}
/* mouse over link */
section-properties, section-concepts, section-cf-lc-english-norms, section-sound-symbolism, section-info a:hover {
color: #2462B0 !important;
border-bottom: none !important;
font-weight: normal !important;
background-color: none !important;
text-shadow: 1px 1px darkgrey !important;
}
/* selected link */
section-properties, section-concepts, section-cf-lc-english-norms, section-sound-symbolism, section-info a:active {
color: #1964BF !important;
border-bottom: none !important;
font-weight: bold !important;
background-color: none !important;
text-shadow: 1px 1px darkgrey !important;
}
<!-- Define CSS style for customising output to specific screen sizes -->
.desktop-only {display: inline;}
/* Smartphone Portrait and Landscape */
@media only screen
and (max-width : 765px){
.desktop-only {display: none;}
}
.mobile-only {display: inline;}
/* Smartphone Portrait and Landscape */
@media only screen
and (min-width : 766px){
.mobile-only {display: none;}
}
<!-- Define CSS style for fonts -->
body{ /* Normal */
font-size: 16px;
}
td { /* Table */
font-size: 14px;
}
h1.title {
font-size: 38px;
font-weight: bold;
color: #28002E
}
h1 { /* Header 1 */
font-size: 28px;
font-weight: bold;
color: #28002E
}
h2 { /* Header 2 */
font-size: 22px;
font-weight: bold;
color: #28002E
}
h3 { /* Header 3 */
font-size: 18px;
font-weight: bold;
}
h4 { /* Header 4 */
font-size: 16px;
}
h5 { /* Header 5 */
font-size: 15px;
}
h6 { /* Header 6 */
font-size: 14px;
}
code.r{ /* Code block */
font-size: 12px;
}
pre { /* Code-formatted output */
font-size: 14px;
padding-top: 2px;
padding-bottom: 2px;
margin-top: -15px;
margin-bottom: 18px;
}
<!-- Define CSS style for splitting columns -->
* {
box-sizing: border-box;
}
/* Create two equal columns that float next to each other */
.column {
float: left;
padding: 10px;
}
/* Clear floats after the columns */
.row:after {
content: "";
display: table;
clear: both;
}
</style>
<!-- Load library of icons -->
<script src="https://use.fontawesome.com/releases/v5.15.3/js/all.js"></script>
<!-- Javascript function to enable a hovering tooltip -->
<script>
$(document).ready(function(){
$('[data-toggle="tooltip1"]').tooltip();
});
</script>
</head>
```{r global, include = FALSE}
# Since this script contains the dashboard, the code run is minimised,and all the rest--namely, diagnostic commands--is commented out. Fully run code on the proper analysis scripts, at https://osf.io/brkjw/.
# Perform analyses for PCA
# RELATION AMONG MODALITIES
# Below is a Principal Components Analysis (PCA) with plots. Firstly it is performed
# on the Dutch norms, and then on Lynott and Connell's (2009, 2013) English norms
# (leaving out gustatory and olfactory scores and words).
all = read.csv('https://raw.githubusercontent.com/pablobernabeu/Modality-exclusivity-norms-747-Dutch-English-replication/master/all.csv', fileEncoding = 'Latin1')
#nrow(all)
# Set correct numeric class for standard deviation variables
all$SD_Auditory = as.numeric(all$SD_Auditory)
all$SD_Haptic = as.numeric(all$SD_Haptic)
all$SD_Visual = as.numeric(all$SD_Visual)
# PROPERTIES
# Principal component analysis on the three modalities
# Check conditions for a PCA
# matrix
prop = all[all$cat == 'Property' & !is.na(all$word), c('Auditory', 'Haptic', 'Visual')]
#nrow(prop)
prop_matrix = cor(prop, use = 'complete.obs')
#prop_matrix
#round(prop_matrix, 2)
# POOR: correlations not apt for a PCA, with too many below .3
# now on the raw vars:
#nrow(prop)
#cortest.bartlett(prop)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
#KMO(prop_matrix)
# Result: .56 = mediocre. PCA not strongly recommended. But we still do it
# because the purpose is graphical only.
# check determinant
#det(prop_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_prop = psych::principal(prop, nfactors = 3, rotate = "none")
#pc1_prop
# RESULT: Only PC1, with eigenvalue > 1, should be extracted,
# acc to Kaiser's criterion (Jolliffe's threshold of 0.7 way too lax;
# Field, Miles, & Field, 2012)
# Unrotated: scree plot
#plot(pc1_prop$values, type = "b")
# Result: one or two RCs should be extracted, converging with eigenvalues
# Now with varimax rotation, Kaiser-normalized (by default).
# Always preferable because it captures explained variance best.
# Compare eigenvalues w/ 1 & 2 factors
pc2_prop = psych::principal(prop, nfactors = 2, rotate = "varimax", scores = TRUE)
#pc2_prop
#pc2_prop$loadings
# good to extract 2 factors, as they both explain quite the same variance,
# and both surpass 1 eigenvalue
#pc2_prop$residual
#pc2_prop$fit
#pc2_prop$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals OK: fewer than 50% have absolute
# values > 0.05 (exactly 50% do).Model fit good, > .90.
# Communalities good, all > .7 (av = .83).
# subset and add PCs
props = all[all$cat == 'Property' & !is.na(all$word),]
#nrow(props)
props = cbind(props, pc2_prop$scores)
#nrow(props)
# Set to character format
props$word = as.character(props$word)
# Replace NAs in corpora with 0 to allow selection
props[is.na(props$phonemes_DUTCHPOND), 'phonemes_DUTCHPOND'] = 0
props[is.na(props$freq_lg10CD_SUBTLEXNL), 'freq_lg10CD_SUBTLEXNL'] = 0
props[is.na(props$freq_lg10WF_SUBTLEXNL), 'freq_lg10WF_SUBTLEXNL'] = 0
props[is.na(props$inflected_adj_freq_lg10CD_SUBTLEXNL), 'inflected_adj_freq_lg10CD_SUBTLEXNL'] = 0
props[is.na(props$freq_CELEX_lem), 'freq_CELEX_lem'] = 0
props[is.na(props$orth_neighbours_DUTCHPOND), 'orth_neighbours_DUTCHPOND'] = 0
props[is.na(props$phon_neighbours_DUTCHPOND), 'phon_neighbours_DUTCHPOND'] = 0
props[is.na(props$AoA_Brysbaertetal2014), 'AoA_Brysbaertetal2014'] = 0
props[is.na(props$concrete_Brysbaertetal2014), 'concrete_Brysbaertetal2014'] = 0
# Turn modality exclusivity into percentage format
props$Exclusivity = props$Exclusivity * 100
# CONCEPTS
# Principal component analysis on the three modalities
# Check conditions for a PCA
# matrix
conc = all[all$cat == 'Concept' & !is.na(all$word), c('Auditory', 'Haptic', 'Visual')]
#nrow(conc)
conc_matrix = cor(conc, use = 'complete.obs')
#conc_matrix
#round(conc_matrix, 2)
# POOR: correlations not apt for a PCA, with too many below .3
# now on the raw data:
#nrow(conc)
#cortest.bartlett(conc)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
#KMO(conc_matrix)
# Result: .49 = poor. PCA not strongly recommended. But we still do it
# because the purpose is graphical really.
# check determinant
#det(conc_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_conc = psych::principal(conc, nfactors = 3, rotate = "none")
#pc1_conc
# RESULT good: PC1 and PC2, with eigenvalue > 1, should be extracted,
# acc to Kaiser's criterion (Jolliffe's threshold of 0.7 way too lax;
# Field, Miles, & Field, 2012)
# Unrotated: scree plot
#plot(pc1_conc$values, type = "b")
# Result: with no point of inflexion along the y axis, two PCs would obtain.
# Now with varimax rotation, Kaiser-normalized (by default):
# Always preferable because it captures explained variance best.
# Compare eigenvalues w/ 1 & 2 Principal Components
pc2_conc = psych::principal(conc, nfactors = 2, rotate = "varimax", scores = TRUE)
#pc2_conc
#pc2_conc$loadings
# good to extract 2 Principal Components, as they both explain quite the same variance,
# and both surpass 1 eigenvalue
#pc2_conc$residual
#pc2_conc$fit
#pc2_conc$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals bad: over 50% have absolute
# values > 0.05. Model fit good, > .90. Communalities good, all > .7 (av = .82).
# subset and add PCs
concs = all[all$cat == 'Concept' & !is.na(all$word),]
#nrow(concs)
concs = cbind(concs, pc2_conc$scores)
#nrow(concs)
# Set to character format
concs$word = as.character(concs$word)
# Replace NAs in corpora with 0 to allow selection
concs[is.na(concs$phonemes_DUTCHPOND), 'phonemes_DUTCHPOND'] = 0
concs[is.na(concs$freq_lg10WF_SUBTLEXNL), 'freq_lg10WF_SUBTLEXNL'] = 0
concs[is.na(concs$freq_lg10CD_SUBTLEXNL), 'freq_lg10CD_SUBTLEXNL'] = 0
concs[is.na(concs$freq_CELEX_lem), 'freq_CELEX_lem'] = 0
concs[is.na(concs$orth_neighbours_DUTCHPOND), 'orth_neighbours_DUTCHPOND'] = 0
concs[is.na(concs$phon_neighbours_DUTCHPOND), 'phon_neighbours_DUTCHPOND'] = 0
concs[is.na(concs$AoA_Brysbaertetal2014), 'AoA_Brysbaertetal2014'] = 0
concs[is.na(concs$concrete_Brysbaertetal2014), 'concrete_Brysbaertetal2014'] = 0
# Turn modality exclusivity into percentage format
concs$Exclusivity = concs$Exclusivity * 100
# Colors for plots
colours = c('Auditory' = 'firebrick1', 'Haptic' = 'dodgerblue', 'Visual' = 'forestgreen')
```
<i class="fas fa-code" aria-hidden="true"></i> Info {style="background-color: #FCFCFC; data-width: 100%; width: 900px; margin: 0 auto;"}
=======================================================================
Column {style="height:1300px; background-color:#FCFCFC;"}
--------------------------------------------------------
<div style = "padding-left: 60px; padding-right: 60px; padding-bottom:5px; text-align: justify; background-color: #FCFCFC;">
<div style="text-align: center; font-size: 18px; padding-top: 31px; margin-top: 2px; margin-bottom: 0px;"> This is a reduced version of a dashboard. </div>
<!-- Links -->
<div style = "text-align: center; padding-top: 12px;">
<a href="https://pablobernabeu.shinyapps.io/Dutch-modality-exclusivity-norms/" target="_top" style='color:#827182; border-bottom:none !important; font-size: 17px; font-weight: bold; font-family: "Courier New", Courier, monospace;'> <i class="fas fa-drafting-compass" aria-hidden="true" style='font-size:13px; color:#5A647B;'></i> See complete dashboard </a>
<a href="https://osf.io/58gzs/" target="_top" style="color:#577787; border-bottom:none !important;font-size: 18px; font-weight: bold; font-family: 'Courier New', Courier, monospace;"> <i class='fas fa-database' aria-hidden="true" style='font-size:16px; color:#577787'></i> Data </a>
<a href="https://github.com/pablobernabeu/Modality-exclusivity-norms-747-Dutch-English-replication/blob/master/Dutch-modality-exclusivity-norms-RPubs.Rmd" target="_top" style="color:#577787; border-bottom:none !important;font-size: 18px; font-weight: bold; font-family: 'Courier New', Courier, monospace;"> <i class='fab fa-r-project' aria-hidden="true" style='font-size:16px; color:#577787'></i><i class='fas fa-code' aria-hidden="true" style='font-size:9px; color:#526772'></i> Code </a>
</div>
<div style='text-align:right !important;'>
<a href="http://creativecommons.org/licenses/by/4.0/" rel="Attribution licence" target="_top" style="color:#CEC7BF; font-size:13px;"> Licence <img style="border-width:0;" src="https://i.creativecommons.org/l/by/4.0/80x15.png" alt="Creative Commons License"></img> </a>
</div>
<!-- Enable this once the Binder feature is enabled (https://github.com/jupyter/repo2docker/issues/799) -->
<!-- <div style="text-align: center; color: #A9AEC1; font-size: 12px; padding-top: 21px; margin-top: 2px; margin-bottom: 2px;"> In case of downtime, please [visit here](https://mybinder.org/v2/gh/pablobernabeu/Modality-exclusivity-norms-747-Dutch-English-replication/master?urlpath=shiny/Shiny-app/){target="_top"}. </div> -->
</div>
<div style = "padding-left: 60px; padding-right: 60px; text-align: justify; background-color:#FCFCFC; font-size: 16px; line-height: 1.6;">
<h2> Information </h2>
This dashboard presents the data and analyses of a set of modality exclusivity norms for Dutch. Various tabs and interactive plots are available. In addition, the data set is available in [CSV format](https://osf.io/ge7pn/){target="_top"} and [Excel format including column definitions](https://osf.io/58gzs/){target="_top"}.
The norms, which are ratings of linguistic stimuli, served a twofold purpose: first, the creation of linguistic stimuli ([Bernabeu, 2018](https://psyarxiv.com/s2c5h){target="_top"}; see also Speed & Majid, 2017), and second, a conceptual replication of Lynott and Connell's (2009, 2013) analyses. In the collection of the ratings, forty-two respondents completed surveys for the properties or the concepts separately. Each word was rated by eight participants on average, with a minimum of five (e.g., for *bevriezend*) and a maximum of ten ratings per word (e.g., for *donzig*). The [instructions to participants](https://osf.io/ungey/){target="_top"} were similar to those used by Lynott and Connell (2009, 2013), except that we elicited three modalities (auditory, haptic, visual) instead of five.
> <div style = "text-align: justify; background-color: #FCFCFC; font-size: 15px;"> 'This is a stimulus validation for a future experiment. The task is to rate how much you experience everyday' [properties/concepts] 'using three different perceptual senses: feeling by touch, hearing and seeing. Please rate every word on each of the three senses, from 0 (not experienced at all with that sense) to 5 (experienced greatly with that sense). If you do not know the meaning of a word, leave it blank.' </div>
<h3> Definitions (as in Lynott & Connell, 2009, 2013) </h3>
* Dominant modality: highest-rated modality;
* Modality exclusivity: range of the three modality ratings divided by the sum;
* Perceptual strength: highest rating across modalities.
<br>
```{r out.width=500}
allNL = all[!is.na(all$main),]
allNL$Range = floor(allNL$Exclusivity * 4)
allNL$Range = mapvalues(allNL$Range, from = c(0, 1, 2, 3, 4),
to = c("0-20%", "20-40%", "40-60%", "60-80%", "80-100%"))
allNL$cat = dplyr::recode(allNL$cat, Concept = "Concepts", Property = "Properties")
# Set order to display properties first, instead of alphabetical 'Concepts-Properties' order
allNL$cat = as.factor(allNL$cat)
allNL$cat = factor(allNL$cat, levels = rev(levels(allNL$cat)))
ggplot(allNL) +
geom_bar(mapping = aes(x = main, fill = Range), position = position_stack(reverse = TRUE)) +
scale_fill_grey(start=.9, end=0, labels = c("0-20%", "20-40%", "40-60%", "60-80%", "80-100%"),
guide = guide_legend(reverse = TRUE, override.aes = list(size = 9))) +
scale_x_discrete(expand = c(.24,0)) + scale_y_continuous(expand = expand_scale(mult = c(0, .05))) +
facet_grid(. ~ cat) + labs(fill = "Modality\nExclusivity", x = 'Dominant Modality', y = 'Number of Words') +
theme_bw() + theme(legend.position = c(.17, .59), legend.title = element_text(size = 14, face = 'bold'),
legend.text = element_text(size = 12), legend.background = element_rect(fill=alpha('white', 0)),
axis.title = element_text(size = 17, face = "bold"), axis.text = element_text(size = 15),
axis.title.x = element_text(margin = margin(.09, 0, 0, 0, "cm")),
axis.text.x = element_text(margin = margin(.09, 0, 0, 0, "cm")),
panel.border = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"),
strip.text = element_text(size = 17, face = 'bold', vjust = .7,
margin = margin(.3, 0, .3, 0, "cm")),
axis.ticks.x = element_blank())
```
<h5 style='text-align:justify; padding-left:5px; padding-right:5px; padding-bottom:10px; line-height: 1.4;'> Figure 1. Number of words in the norms per word category, dominant modality, and modality exclusivity. <br> [<i class='fas fa-external-link-alt' aria-hidden=TRUE style='font-size:7'></i> Github](https://raw.githubusercontent.com/pablobernabeu/Modality-exclusivity-norms-747-Dutch-English-replication/master/stacked_exc.png) </h5>
These norms were validated in an experiment demonstrating that shifts across trials with different dominant modalities incurred semantic processing costs ([Bernabeu, Willems, & Louwerse, 2017](https://doi.org/10.31234/osf.io/a5pcz){target="_top"}). All data for that study are [available](https://osf.io/97unm/wiki/home/){target="_top"}, including a [dashboard](https://pablobernabeu.shinyapps.io/ERP-waveform-visualization_CMS-experiment/){target="_top"} (in case of downtime of the dashboard site, please see [this alternative](https://mybinder.org/v2/gh/pablobernabeu/Modality-switch-effects-emerge-early-and-increase-throughout-conceptual-processing/master?urlpath=shiny/Shiny-app/){target="_top"}).
The [**properties**](#properties){style='background-color:#FDFFFF'} and the [**concepts**](#concepts){style='background-color:#FDFFFF'} may also be consulted together on a [**table**](#table){style='background-color:#FDFFFF'}. Distinct relationships emerged among the modalities, with the visual and haptic modalities being more related to each other than to the auditory modality. This ties in with findings that, in conceptual processing, modalities can be collated based on language statistics (Louwerse & Connell, 2011). Furthermore, properties were found to be more strongly perceptual than concepts ([**cf. English norms by Lynott & Connell, 2009, 2013**](#cf-lc-english-norms){style='background-color:#FDFFFF'}).
<br>
```{r out.width=775, out.length=775}
# ENGLISH PROPERTIES
# check conditions for a PCA
# matrix
eng_prop = all[all$cat == 'Property', c('English_Auditory_Lynott_Connell_2009_2013', 'English_Haptic_Lynott_Connell_2009_2013', 'English_Visual_Lynott_Connell_2009_2013')]
# nrow(eng_prop)
eng_prop_matrix = cor(eng_prop, use = 'complete.obs')
# eng_prop_matrix
# round(eng_prop_matrix, 2)
# OK: correlations good for a PCA, with enough < .3
# now on the raw vars:
# nrow(eng_prop)
# cortest.bartlett(eng_prop)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
# KMO(eng_prop_matrix)
# Result: .56 = mediocre. PCA not strongly recommended. But we still do it
# because the purpose is graphical only.
# check determinant
# det(eng_prop_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_eng_prop = psych::principal(eng_prop, nfactors = 3, rotate = "none")
# pc1_eng_prop
# RESULT: Extract either one PC, acc to Kaiser's criterion, or two RCs, acc to
# Joliffe's (Field, Miles, & Field, 2012)
# Unrotated: scree plot
# plot(pc1_eng_prop$values, type = "b")
# Result: again one or two RCs should be extracted
# Now with varimax rotation, Kaiser-normalized (by default)
pc2_eng_prop = psych::principal(eng_prop, nfactors = 2, rotate = "varimax", scores = TRUE)
# pc2_eng_prop
# pc2_eng_prop$loadings
# two components are good, as they both have eigenvalues over 1
# pc2_eng_prop$residual
# pc2_eng_prop$fit
# pc2_eng_prop$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals bad: more than 50% have absolute
# values > 0.05. Model fit good, > .90. Communalities good, all > .7.
# subset and add PCs
eng_props = all[all$cat == 'Property', ]
# nrow(eng_props)
eng_props = cbind(eng_props, pc2_eng_prop$scores)
# nrow(eng_props)
#
# head(eng_props)
# Finally, plot
# Set sample words to show on plot (first word in each modality)
auditory_w = as.character(sort(eng_props[eng_props$English_Main_Lynott_Connell_2009_2013=='Auditory', 'English_Word_Lynott_Connell_2009_2013'])[1])
haptic_w = as.character(sort(eng_props[eng_props$English_Main_Lynott_Connell_2009_2013=='Haptic', 'English_Word_Lynott_Connell_2009_2013'])[1])
visual_w = as.character(sort(eng_props[eng_props$English_Main_Lynott_Connell_2009_2013=='Visual', 'English_Word_Lynott_Connell_2009_2013'])[1])
w_set = c(auditory_w, haptic_w, visual_w)
eng_props$English_Main_Lynott_Connell_2009_2013 = dplyr::recode(eng_props$English_Main_Lynott_Connell_2009_2013, Auditory = "a", Haptic = "h", Visual = "v")
Engprops = ggplot(eng_props,
aes(RC1, RC2, label = as.character(English_Main_Lynott_Connell_2009_2013))) +
stat_density2d(color = "gray87") +
geom_text(size = ifelse(eng_props$English_Word_Lynott_Connell_2009_2013 %in% w_set, 4, 2.3),
fontface = ifelse(eng_props$English_Word_Lynott_Connell_2009_2013 %in% w_set, 'bold', 'plain')) +
geom_point(data=eng_props[eng_props$English_Word_Lynott_Connell_2009_2013 %in% w_set,],
pch=21, fill=NA, size=4, stroke=2, alpha=.6) +
ggtitle('English properties (Lynott & Connell, 2009)') +
labs(x = "", y = "Varimax-rotated Principal Component 2") + theme_bw() +
theme( plot.background = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), panel.border = element_blank(),
axis.line = element_line(color = 'black'),
axis.title.x = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.title.y = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.text.x = element_text(size=5), axis.text.y = element_text(size=5),
plot.title = element_text(hjust = 0.5, size = 10, margin=margin(4,4,4,4)),
plot.margin=unit(c(0,0,0,0), "lines")) +
geom_label_repel(data = eng_props[eng_props$English_Word_Lynott_Connell_2009_2013 %in% w_set,],
aes(label = English_Word_Lynott_Connell_2009_2013), size = 3,
alpha = 0.77, color = 'black', box.padding = 1.5 )
# ENGLISH CONCEPTS
# check conditions for a PCA
# matrix
eng_conc = all[all$cat == 'Concept', c('English_Auditory_Lynott_Connell_2009_2013', 'English_Haptic_Lynott_Connell_2009_2013', 'English_Visual_Lynott_Connell_2009_2013')]
# nrow(eng_conc)
eng_conc_matrix = cor(eng_conc, use = 'complete.obs')
# eng_conc_matrix
# round(eng_conc_matrix, 2)
# POOR: correlations not apt for a PCA, with too many below .3
# now on the raw data:
# nrow(eng_conc)
# cortest.bartlett(eng_conc)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
# KMO(eng_conc_matrix)
# Result: .48 = poor. PCA not strongly recommended. But we still do it
# because the purpose is graphical really.
# check determinant
# det(eng_conc_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_eng_conc = psych::principal(eng_conc, nfactors = 3, rotate = "none")
# pc1_eng_conc
# RESULT: Extract either one PC, acc to Kaiser's criterion, or two RCs, acc to
# Joliffe's (Field, Miles, & Field, 2012)
# Unrotated: scree plot
# plot(pc1_eng_conc$values, type = "b")
# Result: two PCs obtain.
# Now with varimax rotation, Kaiser-normalized (by default):
# always preferable because it captures explained variance best.
pc2_eng_conc = psych::principal(eng_conc, nfactors = 2, rotate = "varimax", scores = TRUE)
# pc2_eng_conc
# pc2_eng_conc$loadings
#
#
# pc2_eng_conc$residual
# pc2_eng_conc$fit
# pc2_eng_conc$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals bad: over 50% have absolute
# values > 0.05. Model fit good, > .90. Communalities good, all > .7.
# subset and add PCs
eng_concs = all[all$cat == 'Concept', ]
# nrow(eng_concs)
eng_concs = cbind(eng_concs, pc2_eng_conc$scores)
# summary(eng_concs$RC1, eng_concs$RC2)
eng_concs = eng_concs[eng_concs$normed == 'Dut_Eng' | eng_concs$normed == 'English',]
# nrow(eng_concs)
# summary(eng_concs$RC1, eng_concs$RC2)
# Finally, plot
# Set sample words to show on plot (first word in each modality)
auditory_w = as.character(sort(eng_concs[eng_concs$English_Main_Lynott_Connell_2009_2013=='Auditory', 'English_Word_Lynott_Connell_2009_2013'])[1])
haptic_w = as.character(sort(eng_concs[eng_concs$English_Main_Lynott_Connell_2009_2013=='Haptic', 'English_Word_Lynott_Connell_2009_2013'])[1])
visual_w = as.character(sort(eng_concs[eng_concs$English_Main_Lynott_Connell_2009_2013=='Visual', 'English_Word_Lynott_Connell_2009_2013'])[1])
w_set = c(auditory_w, haptic_w, visual_w)
eng_concs$English_Main_Lynott_Connell_2009_2013 = dplyr::recode(eng_concs$English_Main_Lynott_Connell_2009_2013, Auditory = "a", Haptic = "h", Visual = "v")
Engconcs = ggplot(eng_concs,
aes(RC1, RC2, label = as.character(English_Main_Lynott_Connell_2009_2013))) +
stat_density2d(color = "gray87") +
geom_text(size = ifelse(eng_concs$English_Word_Lynott_Connell_2009_2013 %in% w_set, 4, 2.3),
fontface = ifelse(eng_concs$English_Word_Lynott_Connell_2009_2013 %in% w_set, 'bold', 'plain')) +
geom_point(data=eng_concs[eng_concs$English_Word_Lynott_Connell_2009_2013 %in% w_set,],
pch=21, fill=NA, size=4, stroke=2, alpha=.6) +
ggtitle('English concepts (Lynott & Connell, 2013)') +
labs(x = "Varimax-rotated Principal Component 1", y = "Varimax-rotated Principal Component 2") +
theme_bw() +
theme( plot.background = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), panel.border = element_blank(),
axis.line = element_line(color = 'black'),
axis.title.x = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.title.y = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.text.x = element_text(size=5), axis.text.y = element_text(size=5),
plot.title = element_text(hjust = 0.5, size = 10, margin=margin(4,4,4,4)),
plot.margin=unit(c(0,0,0,0), "lines")) +
geom_label_repel(data = eng_concs[eng_concs$English_Word_Lynott_Connell_2009_2013 %in% w_set,],
aes(label = English_Word_Lynott_Connell_2009_2013), size = 3,
alpha = 0.77, color = 'black', box.padding = 1.5 )
# DUTCH
# Properties
# check conditions for a PCA
# matrix
Property = all[all$cat == 'Property' & !is.na(all$word), c('Auditory', 'Haptic', 'Visual')]
# nrow(Property)
prop_matrix = cor(Property, use = 'complete.obs')
# prop_matrix
# round(prop_matrix, 2)
# POOR: correlations not apt for a PCA, with too many below .3
# now on the raw vars:
# nrow(Property)
# cortest.bartlett(Property)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
# KMO(prop_matrix)
# Result: .56 = mediocre. PCA not strongly recommended. But we still do it
# because the purpose is graphical only.
# check determinant
# det(prop_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_prop = psych::principal(Property, nfactors = 3, rotate = "none")
# pc1_prop
# RESULT: Only PC1, with eigenvalue > 1, should be extracted,
# acc to Kaiser's criterion (Jolliffe's threshold of 0.7 way too lax;
# Field, Miles, & Field, 2012)
# Unrotated: scree plot
# plot(pc1_prop$values, type = "b")
# Result: one or two RCs should be extracted, converging with eigenvalues
# Now with varimax rotation, Kaiser-normalized (by default).
# Always preferable because it captures explained variance best.
# Compare eigenvalues w/ 1 & 2 factors
pc2_prop = psych::principal(Property, nfactors = 2, rotate = "varimax", scores = TRUE)
# pc2_prop
# pc2_prop$loadings
# good to extract 2 factors, as they both explain quite the same variance,
# and both surpass 1 eigenvalue
# pc2_prop$residual
# pc2_prop$fit
# pc2_prop$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals OK: fewer than 50% have absolute
# values > 0.05 (exactly 50% do).Model fit good, > .90.
# Communalities good, all > .7 (av = .83).
# subset and add PCs
properties = all[all$cat == 'Property' & !is.na(all$word), ]
# nrow(properties)
properties = cbind(properties, pc2_prop$scores)
# nrow(properties)
# Finally, plot: letters+density (cf. Lynott & Connell, 2009, 2013)
# Set sample words to show on plot (first word in each modality)
auditory_w = as.character(sort(properties[properties$main=='Auditory', 'word'])[1])
haptic_w = as.character(sort(properties[properties$main=='Haptic', 'word'])[1])
visual_w = as.character(sort(properties[properties$main=='Visual', 'word'])[1])
w_set = c(auditory_w, haptic_w, visual_w)
properties$main = dplyr::recode(properties$main, Auditory = "a", Haptic = "h", Visual = "v")
NLprops = ggplot(properties,
aes(RC1, RC2, label = as.character(main))) +
stat_density2d(color = "gray87") +
geom_text(size = ifelse(props$word %in% w_set, 4, 2.3),
fontface = ifelse(props$word %in% w_set, 'bold', 'plain')) +
geom_point(data=props[props$word %in% w_set,],
pch=21, fill=NA, size=4, stroke=2, alpha=.6) +
ggtitle('Dutch properties') +
labs(x = "", y = "") +
theme_bw() +
theme( plot.background = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), panel.border = element_blank(),
axis.line = element_line(color = 'black'),
axis.title.x = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.title.y = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.text.x = element_text(size=5), axis.text.y = element_text(size=5),
plot.title = element_text(hjust = 0.5, size = 10, margin=margin(4,4,4,4)),
plot.margin=unit(c(0,0,0,0), "lines")) +
geom_label_repel(data = props[props$word %in% w_set,],
aes(label = word), size = 3,
alpha = 0.77, color = 'black', box.padding = 1.5 )
# CONCEPTS
# check conditions for a PCA
# matrix
Concept = all[all$cat == 'Concept' & !is.na(all$word), c('Auditory', 'Haptic', 'Visual')]
# nrow(Concept)
conc_matrix = cor(Concept, use = 'complete.obs')
# conc_matrix
# round(conc_matrix, 2)
# POOR: correlations not apt for a PCA, with too many below .3
# now on the raw data:
# nrow(Concept)
# cortest.bartlett(Concept)
# GOOD: Bartlett's test significant
# KMO: Kaiser-Meyer-Olkin Measure of Sampling Adequacy
# KMO(conc_matrix)
# Result: .49 = poor. PCA not strongly recommended. But we still do it
# because the purpose is graphical really.
# check determinant
# det(conc_matrix)
# GOOD: > 0.00001
# start off with unrotated PCA
pc1_conc = psych::principal(Concept, nfactors = 3, rotate = "none")
# pc1_conc
# RESULT good: PC1 and PC2, with eigenvalue > 1, should be extracted,
# acc to Kaiser's criterion (Jolliffe's threshold of 0.7 way too lax;
# Field, Miles, & Field, 2012)
# Unrotated: scree plot
# plot(pc1_conc$values, type = "b")
# Result: with no point of inflexion along the y axis, two PCs would obtain.
# Now with varimax rotation, Kaiser-normalized (by default):
# Always preferable because it captures explained variance best.
# Compare eigenvalues w/ 1 & 2 Principal Components
pc2_conc = psych::principal(Concept, nfactors = 2, rotate = "varimax", scores = TRUE)
# pc2_conc
# pc2_conc$loadings
# good to extract 2 Principal Components, as they both explain quite the same variance,
# and both surpass 1 eigenvalue
# pc2_conc$residual
# pc2_conc$fit
# pc2_conc$communality
# Results based on a Kaiser-normalizalized orthogonal (varimax) rotation
# (by default in psych::stats). Residuals bad: over 50% have absolute
# values > 0.05. Model fit good, > .90. Communalities good, all > .7 (av = .82).
# subset and add PCs
concepts = all[all$cat == 'Concept' & !is.na(all$word), ]
# nrow(concepts)
concepts = cbind(concepts, pc2_conc$scores)
# nrow(concepts)
# Finally, plot
# Set sample words to show on plot (first word in each modality)
auditory_w = as.character(sort(concepts[concepts$main=='Auditory', 'word'])[1])
haptic_w = as.character(sort(concepts[concepts$main=='Haptic', 'word'])[1])
visual_w = as.character(sort(concepts[concepts$main=='Visual', 'word'])[1])
w_set = c(auditory_w, haptic_w, visual_w)
concepts$main = dplyr::recode(concepts$main, Auditory = "a", Haptic = "h", Visual = "v")
NLconcs = ggplot(concepts,
aes(RC1, RC2, label = as.character(main))) +
stat_density2d(color = "gray87") +
geom_text(size = ifelse(concs$word %in% w_set, 4, 2.3),
fontface = ifelse(concs$word %in% w_set, 'bold', 'plain')) +
geom_point(data=concs[concs$word %in% w_set,],
pch=21, fill=NA, size=4, stroke=2, alpha=.6) +
ggtitle('Dutch concepts') +
labs(x = "Varimax-rotated Principal Component 1", y = "") +
theme_bw() +
theme( plot.background = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), panel.border = element_blank(),
axis.line = element_line(color = 'black'),
axis.title.x = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.title.y = element_text(colour = 'black', size = 7, margin=margin(2,2,2,2)),
axis.text.x = element_text(size=5), axis.text.y = element_text(size=5),
plot.title = element_text(hjust = 0.5, size = 10, margin=margin(4,4,4,4)),
plot.margin=unit(c(0,0,0,0), "lines")) +
geom_label_repel(data = concs[concs$word %in% w_set,],
aes(label = word), size = 3,
alpha = 0.77, color = 'black', box.padding = 1.5 )
# Four plots on a grid
multiplot(Engprops, Engconcs, NLprops, NLconcs, cols = 2)
```
<h5 style='text-align:justify; padding-left:5px; padding-right:5px; padding-bottom:5px; line-height:1.4;'> Figure 2. Dutch norms compared to English norms (reanalysis of Lynott & Connell, 2009, 2013, narrowed to three modalities) based on a principal component analysis of the auditory, haptic, and visual ratings for each word. Letters indicate the dominant modality of each word (*a* = auditory, *h* = haptic, *v* = visual), and contours further display the degree of consistency of the modalities. [<i class='fas fa-external-link-alt' aria-hidden=TRUE style='font-size:7'></i> Github](https://raw.githubusercontent.com/pablobernabeu/Modality-exclusivity-norms-747-Dutch-English-replication/master/allfour_lowres.png)
<!-- Begin two-column format -->
<div class="row">
<div class="column" style="width: 65%;">
<div>
The norms also served to investigate [**sound symbolism**](#sound-symbolism){style='background-color:#FDFFFF'}, which is the relation between the form of words and their meaning. The form of words rests on their sound more than on their visual or tactile properties (at least in spoken language). Therefore, auditory ratings should more reliably predict the lexical properties of words (length, frequency, distinctiveness) than haptic or visual ratings would. Lynott and Connell's (2013) findings were replicated, as auditory ratings were either the best predictor of lexical properties, or yielded an effect that was opposite in polarity to the effects of haptic and visual ratings. The present analyses and further ones will be reported in a forthcoming paper.
All data and code are [available for re-use](https://osf.io/brkjw/wiki/home/){target="_top"} under a [CC BY licence](https://creativecommons.org/licenses/by/4.0/){target="_top"}, by citing the source:
</div>
</div>
<div class="column" style="width: 35%; padding-left: 8px;">
<div style = "padding-top: 5px; padding-bottom: 8px; float: right; font-size: 13px; display: inline-block; background-color: #FAFAFA;">
<div style = "text-align: center !important; padding-bottom: 0 !important;"><img src="https://upload.wikimedia.org/wikipedia/commons/e/e7/Booba-Kiki.svg" alt="Bouba-Kiki" width=80px height=55px></img></div>
<div style="text-align: justify !important; padding-top: 0 !important; padding-right: 5px; padding-left: 5px;"> Sound symbolism is a psycholinguistic effect whereby the pronunciation and the meaning of words bear a non-arbitrary relationship. For instance, when people are asked to match the pseudowords *bouba* and *kiki* to the above objects, the vast majority name the angular object *kiki* and the smooth one *bouba* (Köhler, 1929; Sourav et al., 2019). </div>
<div style="font-size:65%; padding-top:4px; padding-right: 5px; padding-left: 5px; text-align:left;"> Image: <a href="https://commons.wikimedia.org/wiki/File:Booba-Kiki.svg" target="_top" title="via Wikimedia Commons">Monochrome version 1 June 2007 by BendžVectorized with Inkscape --Qef (talk) 21:21, 23 June 2008 (UTC)</a> [<a href="http://creativecommons.org/licenses/by-sa/3.0/" target="_top">CC BY-SA</a>].</div>
</div>
</div>
<!-- End two-column format -->
</div>
> <div style = 'text-align: justify; text-indent: -1.5em; margin-left: 1.5em; font-size: 16px; background-color: #FCFCFC; font-size: 15px;'> Bernabeu, P. (2018). Dutch modality exclusivity norms for 336 properties and 411 concepts [Data dashboard]. Retrieved from [https://pablobernabeu.shinyapps.io/Dutch-Modality-Exclusivity-Norms/](https://pablobernabeu.shinyapps.io/Dutch-Modality-Exclusivity-Norms/){target="_top"}.</div>
<!-- Create heading anchor for the corpora section below -->
<a name="external-corpora"></a>
<h3> External corpora </h3>
* Concreteness and age of acquisition: norms by Brysbaert, Warriner, and Kuperman (2014);
* Phonological and orthographic neighbours: DutchPOND (Marian et al., 2012);
* Word frequency and contextual diversity: SUBTLEX-NL (Keuleers, Brysbaert, & New, 2010);
* Lemma frequency: CELEX (Baayen, Piepenbrock, & van Rijn, 1993).
<h3 style='padding-top:5px;'> Acknowledgements </h3>
This research was greatly supported by the supervision from Max Louwerse and Roel Willems; the financial help from Tilburg University; Wendy Leijten's help with the translations; and the forty-two students from Tilburg University and Radboud University who completed the surveys.
<!-- Create heading anchor for the References section below -->
<a name="references"></a>
<h3 style='padding-top:5px;'> References </h3>
<div style = "text-align: justify; text-indent:-2em; margin-left:2em; background-color:#FCFCFC; font-size:15px;">
Baayen, R. H., Piepenbrock, R., & van Rijn, H. (1993). *The CELEX Lexical Database* [CD-ROM]. Philadelphia: Linguistic Data Consortium, University of Pennsylvania
Bernabeu, P. (2018). *Dutch modality exclusivity norms for 336 properties and 411 concepts* [Unpublished manuscript]. School of Humanities, Tilburg University, the Netherlands. [https://psyarxiv.com/s2c5h](https://psyarxiv.com/s2c5h){target="_top"}
Bernabeu, P., Willems, R. M., & Louwerse, M. M. (2017). Modality switch effects emerge early and increase throughout conceptual processing: Evidence from ERPs. In G. Gunzelmann, A. Howes, T. Tenbrink, & E. J. Davelaar (Eds.), *Proceedings of the 39th Annual Conference of the Cognitive Science Society* (pp. 1629-1634). Austin, TX: Cognitive Science Society. [https://doi.org/10.31234/osf.io/a5pcz](https://doi.org/10.31234/osf.io/a5pcz){target="_top"}
Brysbaert, M., Warriner, A.B., & Kuperman, V. (2014). Concreteness ratings for 40 thousand generally known English word lemmas. *Behavior Research Methods, 46*, 3, 904-911. <br>
[https://doi.org/10.3758/s13428-013-0403-5](https://doi.org/10.3758/s13428-013-0403-5){target="_top"}
Field, A. P., Miles, J., & Field, Z. (2012). *Discovering Statistics Using R*. London, UK: Sage
Keuleers, E., Brysbaert, M. & New, B. (2010). SUBTLEX-NL: A new frequency measure for Dutch words based on film subtitles. *Behavior Research Methods, 42*, 3, 643-650. [https://doi.org/10.3758/BRM.42.3.643](https://doi.org/10.3758/BRM.42.3.643){target="_top"}
Köhler, W. (1929). *Gestalt Psychology*. New York: Liveright