The Dimensionality of Lexical Features in General, Academic, and Disciplinary Vocabulary

Figures & data

Table 1. Description of word feature measures.

Measure	Name	Description	Corpus	Description	Citation
cd	contextual diversity	the number of documents in which a word appears	Touchstone Applied Science Associates (TASA)	Approx.120,000 paragraphs taken from 38,000 academic texts	Adelman et al. (2006)
cocazipf	COCA frequency	Zipfian-transformed word frequencies	Corpus of Contemporary American English (COCA)	Approx. 560 million words from T.V., radio, newspapers, fiction, academic papers, and popular magazines	Davies (2008)
d	dispersion	number of subject areas in which a word appears; scores range from 0 (only one area) to 1 (all areas)	The Educator’s Word Frequency Guide	Approx. 17 million words from kindergarten- to college-level texts	Zeno et al. (1995)
freqband	frequency band	frequency band grouping from the Oxford English Dictionary; bands run from 1 (infrequent) to 8 (frequent) a	Google Ngrams, version 2	Approx. 450 million words scanned from Google Books	Lin et al. (2012)
length	length	number of letters; collected from ELP	NA		Balota et al. (2007)
log_freq_hal	HAL frequency	log-transformed word frequencies; collected from the English Lexicon Project (ELP; Balota et al., 2007)	Hyperspace Analogue to Language	Approx. 131 million words from 3,000 Usenet newsgroups	Lund and Burgess (1996)
log_freq_kf	Brown frequency	log-transformed word frequencies; collected from the English Lexicon Project (ELP; Balota et al., 2007)	Brown Corpus of Standard American English	Approx. 1 million words from American English texts	Kučera and Francis (1967)
nmorph	morphemes	number of morphemes; collected from ELP	NA		Balota et al. (2007)
nphon	phonemes	number of phonemes; collected from ELP	NA		Balota et al. (2007)
nsyll	syllables	number of syllables; collected from ELP	NA		Balota et al. (2007)
og_n	phonographic neighbors	raw number of phonographic neighbors (i.e., the number of words that are one letter and one phoneme away from the word, e.g., “stove” and “stone”), excluding homophones	English Lexicon Project	Standardized behavioral and descriptive data set for 40,481 words (and 40,481 nonwords)	Balota et al. (2007)
old20	orthographic Levenshtein distance	mean Levenshtein distance of the 20 closest orthographic neighbors	English Lexicon Project	Standardized behavioral and descriptive data set for 40,481 words (and 40,481 nonwords)	Yarkoni et al. (2008)
ortho_n	orthographic neighbors	raw number of orthographic neighbors (i.e., the number of words that are one letter away from the word, e.g., “lost” and “lose”), excluding homophones	English Lexicon Project	Standardized behavioral and descriptive data set for 40,481 words (and 40,481 nonwords)	Balota et al. (2007)
phono_n	phonologic neighbors	raw number of phonologic neighbors (i.e., the number of words that are one phoneme away from the word, e.g., “hear” and “hare”), excluding homophones	English Lexicon Project	Standardized behavioral and descriptive data set for 40,481 words (and 40,481 nonwords)	Balota et al. (2007)
pld20	phonologic Levenshtein distance	mean Levenshtein distance of the 20 closest phonologic neighbors	English Lexicon Project	Standardized behavioral and descriptive data set for 40,481 words (and 40,481 nonwords)	Yarkoni et al. (2008)
semd	semantic diversity	mean cosine of the latent semantic analysis vectors for all pairwise combinations of contexts containing the word	British National Corpus	Approx. 100 million words from T.V., radio, newspapers, fiction, academic papers, and popular magazines	Hoffman et al. (2013)
subzipf	Subtlex frequency	Zipfian-transformed word frequencies	SubtlexUS corpus (Subtitle Lexicon- U.S. version)	Approx. 51 million words from American subtitles	Brysbaert and New (2009)
wordage	word age	number of years since a word was first used (as of 2000), based on oldest occurrence across parts of speech	Google Ngrams, version 2	Approx. 450 million words scanned from Google Books	Lin et al. (2012)
wordnet_lnapossam	Wordnet senses and meanings	log-transformed number of senses and meanings a word has across all parts of speech	WordNet lexical database	Database of 155,327 words organized in 175,979 synonym sets and hypernym chains	Fellbaum (2005)
wordsmyth_lnapossam	Wordsmyth senses and meanings	log-transformed number of senses and meanings a word has across all parts of speech	Wordsmyth integrated dictionary and thesaurus	Advanced dictionary and integrated thesaurus for 60,000 headwords	Parks et al. (1998)
z_sem_prec	semantic precision	z-transformed depth scores averaged across parts of speech; scores range from 0 (shallow/vague) to 10 (deep/precise)	WordNet lexical database	Database of 155,327 words organized in 175,979 synonym sets and hypernym chains	Fellbaum (2005)
zenozipf	Zeno frequency	Zipfian-transformed word frequencies	The Educator’s Word Frequency Guide	Approx. 17 million words from kindergarten- to college-level texts	Zeno et al. (1995)

Figure 1. Overlap between word lists for unique words with complete data (n = 10,744 words).

Table 2. Descriptive statistics for word features by word list.

	General Service List (GSL)		Academic Word List (AWL)		Academic Vocabulary List – Domain Specific (AVL-DS)
	n = 2060		n = 1051		n = 3413
Word Feature	M	SD	M	SD	M	SD
cd	1036.52	2453.23	175.54	300.74	142.06	424.10
cocazipf	4.62	0.64	4.15	0.57	3.82	0.59
d	0.80	0.14	0.71	0.19	0.56	0.21
freqband	6.09	0.67	5.83	0.62	5.31	0.60
length	5.84	2.01	8.57	2.33	7.31	2.45
log_freq_hal	9.72	1.60	8.70	1.45	7.71	1.52
log_freq_kf	1.62	0.63	1.16	0.55	0.82	0.58
nmorph	1.32	0.58	2.19	0.83	1.73	0.78
nphon	4.75	1.81	7.40	2.15	6.19	2.27
nsyll	1.75	0.86	3.02	1.05	2.45	1.13
og_n	3.01	4.54	0.36	1.33	1.54	3.30
old	2.02	0.72	2.84	0.79	2.59	0.94
ortho_n	5.00	6.38	0.65	1.84	2.52	4.59
phono_n	11.87	14.39	1.39	4.31	5.84	10.59
pld	1.87	0.82	2.94	0.98	2.57	1.15
semd	1.80	0.27	1.79	0.25	1.44	0.30
subzipf	4.44	0.75	3.41	0.67	3.38	0.77
word_age	742.83	243.53	499.54	185.46	542.02	251.93
wordnet_lnapossam	1.73	0.82	1.26	0.68	1.09	0.79
wordsmyth_lnapossam	1.70	0.73	1.06	0.61	1.07	0.72
z_sem_prec	−0.03	0.73	−0.14	0.87	0.24	0.86
zenozipf	4.67	0.64	3.91	0.60	3.71	0.68

This table includes only the final 22 features used in the models.

Table 3. Factor analysis fit by word list reference.

	General Service List (GSL)	Academic Word List (AWL)	Academic Vocabulary List – Domain Specific (AVL-DS)
Bartlett’s test
df	231	231	231
X^2	50136.55	20743.13	72063.03
p	<.0001	<.0001	<.0001
Plot suggestions
Parallel Analysis (fa.parallel)	5	5	5
Parallel Analysis (nFactors)	4	5	4
Optimal Coordinates (nFactors)	4	5	4
Eigenvalues > Mean	5	5	4
Eigenvalues >1	5	5	4
Acceleration Factor	1	1	1
5-Factor Model Fit
% Variance Explained	75%	69%	68%
Overall Sampling Adequacy (MSA)	0.92	0.90	.92
Comparative Fit Index (CFI)	0.962	0.948	.956
Tucker-Lewis Index (TLI)	0.934	0.907	.923
Root Mean Square of Residuals (RMSR)	0.02	0.02	.03
Root Mean Square Error of Approximation (RMSEA)	0.083	0.088	.084
Lower Bound	0.080	0.084	.081
Upper Bound	0.087	0.093	.086
4-Factor Model Fit
% Variance Explained	71%	65%	67%
Overall Sampling Adequacy (MSA)	0.92	0.90	.92
Comparative Fit Index (CFI)	0.943	0.923	.933
Tucker-Lewis Index (TLI)	0.912	0.88	.896
Root Mean Square of Residuals (RMSR)	0.03	0.04	.03
Root Mean Square Error of Approximation (RMSEA)	0.096	0.010	.097
Lower Bound	0.093	0.096	.095
Upper Bound	0.099	0.105	.100

Higher values (at least above .90) indicate adequate model fit for MSA, CFI, and TLI. Lower values (at least below .10 indicate adequate model fit for RMSR and RMSEA.

Table 4. Factor loadings by Wordlist.

	Frequency			Complexity			Proximity			Polysemy (Polysemy/Diversity)^a			Diversity
	GSL	AWL	AVL-DS	GSL	AWL	AVL-DS	GSL	AWL	AVL-DS	GSL	AWL	AVL-DS	GSL	AWL
cocazipf	1.00	0.99	0.98	0.02	0.02	0.02	−0.01	0.01	0.01	0.01	−0.04	−0.03	−0.02	−0.02
log_freq_kf	0.96	0.86	0.83	0.08	0.07	0.09	0.00	0.00	−0.03	−0.01	0.03	0.06	0.03	0.06
zenozipf	0.95	0.89	0.84	−0.07	−0.01	−0.07	−0.01	0.01	0.02	0.02	0.02	−0.01	−0.06	0.06
log_freq_hal	0.90	0.89	0.84	0.02	−0.06	−0.03	0.02	0.00	0.02	0.06	0.02	0.03	0.04	−0.07
subzipf	0.84	0.73	0.77	−0.17	−0.11	−0.21	0.02	0.03	0.04	0.01	0.09	−0.02	−0.05	0.07
freqband	0.75	0.72	0.76	0.08	−0.07	0.10	−0.01	−0.08	−0.06	0.02	−0.01	−0.01	0.12	−0.07
cd	0.61	0.61	0.58	−0.06	0.04	0.05	0.09	0.12	0.07	−0.17	−0.02	0.01	0.15	0.04
wordage	0.30	0.29	0.29	0.06	0.06	0.02	0.00	−0.02	−0.02	0.11	0.16	0.29	0.16	0.32
nphon	−0.02	−0.02	−0.01	0.98	0.97	0.98	0.01	0.01	0.00	0.06	0.05	0.03	−0.01	0.01
nsyll	0.04	0.05	−0.01	0.90	0.84	0.90	0.02	−0.02	0.01	−0.06	−0.06	−0.03	0.01	−0.11
length	−0.05	−0.03	−0.01	0.89	0.97	0.94	−0.09	0.01	−0.05	0.03	0.07	0.07	−0.02	0.01
pld	0.03	0.02	0.00	0.87	0.86	0.88	−0.06	−0.02	0.00	−0.07	−0.08	−0.11	−0.02	0.00
nmorph	−0.07	−0.09	−0.02	0.78	0.69	0.77	0.16	0.04	0.08	−0.04	0.00	0.04	0.03	0.01
old	0.00	0.01	0.00	0.73	0.80	0.78	−0.22	−0.08	−0.11	−0.09	−0.12	−0.12	0.00	0.05
og_n	−0.01	0.01	−0.02	0.05	0.06	0.06	0.96	0.99	0.98	0.03	0.01	0.02	−0.02	0.00
ortho_n	0.03	0.00	0.02	−0.03	−0.02	−0.01	0.95	0.95	0.98	0.01	−0.01	−0.02	0.00	−0.01
phono_n	0.01	0.02	0.02	−0.27	−0.11	−0.17	0.64	0.78	0.73	0.01	−0.01	0.02	0.02	0.01
wordnet_lnapossam	0.00	−0.03	0.01	0.04	0.00	0.00	0.02	0.02	0.01	0.96	0.87	0.90	0.02	0.05
wordsmyth_lnapossam	0.05	0.10	0.03	−0.11	−0.08	−0.11	0.00	−0.02	0.04	0.79	0.68	0.70	0.00	−0.07
semd	0.00	0.01	0.24	−0.01	−0.01	0.06	−0.01	0.00	−0.01	0.01	0.05	0.39	0.98	0.74
d	0.27	0.33	0.52	0.03	−0.04	0.00	−0.01	0.00	−0.01	0.10	0.02	0.19	0.46	0.55
z_sem_prec	0.07	0.21	0.04	0.07	0.09	−0.14	0.04	0.02	0.00	0.03	0.24	−0.26	−0.42	−0.44
Explained Variance	26%	24%	24%	22%	21%	22%	12%	12%	12%	8%	7%	9%	7%	6%
Eigenvalue	5.65	5.24	5.21	4.91	4.65	4.93	2.55	2.58	2.62	1.77	1.49	1.93	1.59	1.31
Standardized Alpha	0.94	0.93	0.93	0.96	0.95	0.96	0.93	0.94	0.95	0.90	0.79	0.75	0.68	0.62

GSL – General Service List, AWL – Academic Word List, AVL-DS – Academic Vocabulary List (Domain-Specific). Loadings greater than .30 are bolded for ease.

^aFor the 4-factor solutions, results for the Polysemy-Diversity factor are reported under Polysemy.

Figure 2. Factor correlations for the GSL-reference model estimated on each word sample. Note: Scatterplots below the diagonal contain random 200-word samples while density plots and correlations on and above the diagonal are based on entire word samples. Word samples include AVL_DS (red) Academic Vocabulary List Domain-Specific; AWL (green) Academic Word List; GSL (blue) General Service List; Words (purple) in any word list.

Figure 3. Factor correlations for the AWL-reference model estimated on each word sample. Note: Scatterplots below the diagonal contain random 200-word samples while density plots and correlations on and above the diagonal are based on entire word samples. Word samples include AVL_DS (red) Academic Vocabulary List Domain-Specific; AWL (green) Academic Word List; GSL (blue) General Service List; Words (purple) in any word list.

Figure 4. Factor correlations for the AVLDS-reference model estimated on each word sample. Note: Scatterplots below the diagonal contain random 200-word samples while density plots and correlations on and above the diagonal are based on entire word samples. Word samples include AVL_DS (red) Academic Vocabulary List Domain-Specific; AWL (green) Academic Word List; GSL (blue) General Service List; Words (purple) in any word list.

Table 5. Example words and factor estimates based on each model.

		GSL (General Service List) Reference Model					AWL (Academic Word List) Reference Model					AVLDS (Domain-Specific Word) Reference Model
word	Word List	Frequency	Complexity	Proximity	Polysemy	Diversity	Freq	Comp	Prox	Poly	Div	Freq	Comp	Prox	Poly-Div
accept	GSL	0.41	0.49	−0.64	0.66	1.04	1.47	−0.89	−0.07	1.52	1.12	2.04	−0.22	−0.38	1.62
honest	GSL	−0.05	0.09	−0.75	0.15	0.22	0.93	−1.16	−0.27	0.92	0.81	1.48	−0.53	−0.53	1.11
jaw	GSL	−0.88	−1.08	1.91	0.07	−1.21	−0.10	−2.31	8.36	0.86	−0.78	0.46	−1.56	3.15	0.89
library	GSL	0.34	1.18	−0.78	−0.43	−1.66	1.42	−0.27	−0.31	0.25	−0.36	1.86	0.31	−0.56	0.37
meet	GSL	0.91	−0.91	0.41	0.98	1.34	2.09	−2.13	3.89	1.93	1.45	2.62	−1.36	1.08	1.97
nice	GSL	0.68	−0.93	0.68	0.17	−0.60	1.82	−2.08	4.83	1.03	−0.30	2.27	−1.36	1.44	1.08
race	GSL	0.65	−1.05	1.68	0.92	−1.13	1.69	−2.22	8.42	1.86	−0.44	2.18	−1.44	2.80	1.89
size	GSL	0.91	−0.96	0.09	0.63	1.01	2.18	−2.15	3.10	1.51	1.17	2.62	−1.38	0.61	1.68
usual	GSL	0.32	0.39	−0.77	−1.12	1.22	1.33	−0.94	−0.28	−0.57	1.61	1.81	−0.29	−0.55	0.00
wood	GSL	0.30	−0.89	0.93	0.52	−0.44	1.43	−2.08	4.96	1.43	−0.02	1.86	−1.34	1.84	1.40
abandon	AWL	−0.82	0.99	−0.77	0.16	1.06	−0.01	−0.42	−0.31	0.97	0.50	0.62	0.18	−0.54	1.09
compile	AWL	−1.97	0.55	−0.70	−0.74	0.03	−1.17	−0.77	−0.28	−0.16	−0.62	−0.51	−0.17	−0.52	0.11
evidence	AWL	0.93	1.22	−0.77	0.04	0.70	2.02	−0.22	−0.30	0.89	0.84	2.55	0.34	−0.57	1.07
illustrate	AWL	−0.82	1.85	−0.77	−0.82	0.49	0.01	0.39	−0.32	−0.14	0.74	0.58	0.84	−0.57	0.24
maximize	AWL	−1.53	1.46	−0.71	−1.46	−0.73	−0.82	0.03	−0.32	−1.00	−0.87	−0.13	0.58	−0.55	−0.57
nonetheless	AWL	−0.74	2.65	−0.82	−2.24	0.96	−0.03	1.02	−0.35	−1.86	1.17	0.60	1.35	−0.60	−1.06
process	AWL	1.18	0.56	−0.55	0.97	0.89	2.41	−0.78	0.13	1.91	1.12	2.92	−0.15	−0.23	1.97
ratio	AWL	−0.39	0.20	−0.43	−1.26	−0.52	0.47	−1.11	0.54	−0.71	−0.33	0.98	−0.47	−0.16	−0.32
significant	AWL	0.60	3.10	−0.77	−0.71	1.34	1.50	1.36	−0.34	−0.09	1.37	2.14	1.84	−0.60	0.45
underlying	AWL	−0.62	1.88	−0.71	−0.26	0.59	0.16	0.38	−0.32	0.47	0.46	0.81	0.87	−0.55	0.85
absenteeism	AVLDS	−2.76	3.53	−0.76	−2.59	−1.44	−2.23	1.68	−0.38	−2.25	−1.58	−1.51	1.99	−0.58	−1.68
autopsy	AVLDS	−1.85	0.89	−0.76	−1.44	−2.14	−1.19	−0.47	−0.30	−0.95	−2.18	−0.49	0.07	−0.55	−0.69
carbon	AVLDS	−0.19	0.13	−0.42	−0.49	−2.29	0.84	−1.06	0.36	0.23	−1.50	1.26	−0.41	−0.06	0.31
habitable	AVLDS	−2.47	1.74	−0.68	−2.18	−1.47	−1.91	0.32	−0.31	−1.76	−1.03	−1.32	0.78	−0.53	−1.16
membrane	AVLDS	−1.25	1.21	−0.80	−1.37	−2.50	−0.33	−0.13	−0.31	−0.79	−1.63	0.12	0.36	−0.57	−0.65
linguistic	AVLDS	−1.27	2.50	−0.74	−1.55	−2.86	−0.57	0.93	−0.35	−1.09	−1.49	0.08	1.40	−0.58	−0.65
positivism	AVLDS	−2.86	2.90	−0.56	−1.81	−3.68	−2.34	1.21	−0.16	−1.36	−3.16	−1.69	1.67	−0.39	−1.03
sacrificial	AVLDS	−2.37	2.88	−0.75	−2.36	−2.01	−1.79	1.22	−0.36	−1.98	−1.86	−1.12	1.58	−0.57	−1.41
tuberculosis	AVLDS	−1.50	4.60	−0.89	−2.51	−1.72	−0.76	2.58	−0.43	−2.06	−0.87	−0.26	2.85	−0.63	−1.73
wavelength	AVLDS	−1.75	2.04	−0.86	−1.61	−2.43	−0.97	0.53	−0.35	−1.04	−2.70	−0.44	0.89	−0.59	−0.93

Bolded estimates indicate when words were part of the sample used to create that model. Not bolded estimates were estimated separately, after creating models.

For brevity, Freq – Frequency, Comp – Complexity, Prox – Proximity, Poly – Polysemy, Div – Diversity, and Poly-Div – Polysemy/Diversity factors.

Figure 5. Scaled density plots by reference model and latent factor. Note: Word list samples are: GSL (blue) General Service List; AWL (yellow) Academic Word List; AVLDS (red) Academic Vocabulary List Domain-Specific; Words (purple) in any word list.

Adelman, J. S., Brown, G. D. A., & Quesada, J. F. (2006). Contextual diversity, not word frequency, determines word-naming and lexical decision times. Psychological Science, 17(9), 814–823. https://doi.org/10.1111/j.1467-9280.2006.01787.x

 PubMed Web of Science ®Google Scholar

Davies, M. (2008). The corpus of contemporary American English (COCA): One billion million words, 1990-2019. https://www.english-corpora.org/coca

 Google Scholar

Zeno, S., Ivens, S. H., Millard, R. T., & Duvvuri, R. (1995). The educator’s word frequency guide. Touchstone Applied Science Associates.

 Google Scholar

Lin, Y., Michel, J. B., Lieberman, E. A., Orwant, J., Brockman, W., & Petrov, S. (2012, July). Syntactic annotations for the google books ngram corpus. In Proceedings of the ACL 2012 system demonstrations, Jeju Island, Korea, (pp. 169–174).

 Google Scholar

Balota, D. A., Yap, M. J., Hutchison, K. A., Cortese, M. J., Kessler, B., Loftis, B., Neely, J. H., Nelson, D. L., Simpson, G. B., & Treiman, R. (2007). The English lexicon project. Behavior Research Methods, 39(3), 445–459. https://doi.org/10.3758/BF03193014

 PubMed Web of Science ®Google Scholar

Lund, K., & Burgess, C. (1996). Producing high-dimensional semantic spaces from lexical co-occurrence. Behavior Research Methods, Instruments, & Computers, 28(2), 203–208. https://doi.org/10.3758/BF03204766

 Google Scholar

Kučera, H., & Francis, W. N. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press.

 Google Scholar

Yarkoni, T., Balota, D., & Yap, M. (2008). Moving beyond Coltheart’s N: A new measure of orthographic similarity. Psychonomic Bulletin & Review, 15(5), 971–979. https://doi.org/10.3758/PBR.15.5.971

 PubMed Web of Science ®Google Scholar

Hoffman, P., Lambon Ralph, M. A., & Rogers, T. T. (2013). Semantic diversity: A measure of semantic ambiguity based on variability in the contextual usage of words. Behavior Research Methods, 45(3), 718–730. https://doi.org/10.3758/s13428-012-0278-x

 PubMed Web of Science ®Google Scholar

Brysbaert, M., & New, B. (2009). Moving beyond Kučera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41(4), 977–990. https://doi.org/10.3758/BRM.41.4.977

 PubMed Web of Science ®Google Scholar

Fellbaum, C. (2005). WordNet and wordnets. https://doi.org/10.1016/B0-08-044854-2/00946-9.

 Google Scholar

Parks, R., Ray, J., & Bland, S. (1998). Wordsmyth English dictionary-thesaurus [electronic version]. University of Chicago. www.wordsmyth.net

 Google Scholar