Table 1 Fluency measures and POS tags in the two corpora, with statistical comparison.

From: Characterizing the patient experience of physical restraint in psychiatric settings via a linguistic, sentiment, and metaphor analysis

Analysis

Variable

PR narratives

Reference corpus

W

p-value

M ± SD

Range

N

%

M ± SD

Range

N

%

Fluency

WPN

84.82 ± 99.61

1–391

–

–

91.76 ± 101.86

2–787

–

–

8,936

0.008

SPN

8.43 ± 7.76

1–32

–

–

12.84 ± 11.83

1–92

–

–

9,682

< 0.001

WPS

13.55 ± 8.93

1–53

–

–

16.17 ± 7.47

2–46

–

–

9,146.5

0.003

TTR

0.71 ± 0.12

0.35–1

–

–

0.90 ± 0.10

0.60–1

–

–

1,790.5

< 0.001

Word frequency (log)

4.47 ± 0.59

1.71–5.43

–

–

4.59 ± 0.26

2.96–5.04

–

–

7,140

0.770

POS tagging

ADJ

6.53 ± 7.62

0–50

435

4.63

8.07 ± 5.54

0–33.33

1,002

6.40

9,312.0

< 0.001

ADP

8.92 ± 5.16

0–28.57

910

9.68

12.26 ± 4.26

0–22.45

1,959

12.51

10,383.5

< 0.001

ADV

9.25 ± 6.30

0–33.33

930

9.90

6.68 ± 4.24

0–25

1,044

6.67

5,167.5

< 0.001

AUX

9.56 ± 7.45

0–50

871

9.27

5.90 ± 3.54

0–22.22

904

5.77

4,593.0

< 0.001

CCONJ

3.44 ± 3.12

0–16.67

439

4.67

4.21 ± 2.60

0–13.33

641

4.09

8,418.0

< 0.001

DET

7.74 ± 5.38

0–28.57

841

8.95

9.64 ± 3.72

0–18.75

1,565

10.00

9,307.0

< 0.001

NOUN

15.29 ± 12.66

0 −100

1,266

13.47

18.11 ± 5.64

0–33.33

2,646

16.90

10,418.0

< 0.001

PRON

7.48 ± 5.37

0–25

928

9.88

4.94 ± 3.24

0–18.33

851

5.44

5,069.0

< 0.001

PROPN

2.95 ± 12.86

0–100

104

1.11

5.20 ± 7.85

0–52.56

1,164

7.44

11,718.5

< 0.001

PUNCT

10.74 ± 7.46

0–40

863

9.18

13.41 ± 4.31

3.03–33.33

1,938

12.38

9,906.0

< 0.001

SCONJ

2.66 ± 3.01

0–14.29

301

3.20

0.99 ± 1.17

0–4.30

188

1.20

4,931.5

< 0.001

VERB

14.80 ± 6.95

0–50

1,470

15.64

9.32 ± 3.92

0–18.60

1,528

9.76

2,932.5

< 0.001

  1. The table reports, for each variable, the mean number of occurrences per narrative (M), the standard deviation (SD), the min and maximum number of occurrences per narrative (Range), the total number of occurrences in the corpus (N), and proportion of the POS in the entire corpus (%). Percentages of interjections, numerals, and particles are not reported, and account for 0.42% and 1.44% in the PR narratives and in the reference corpus, respectively. Statistics and p-values are derived from Wilcoxon rank sum tests between independent groups (PR narratives vs. reference corpus), adjusted for multiple comparisons with False Discovery Rate. WPN  word per narrative, SPN  sentence per narrative, WPS  word per sentence, TTR  Type/Token Ratio, ADJ  adjectives, ADP  adpositions, ADV  adverbs, AUX  auxiliaries, CCONJ  coordinating conjunctions, DET  determiners, M  mean: average number of occurrences per each narrative, NOUN   nouns, PR  physical restraint, PRON  pronouns, PROPN   proper nouns, PUNCT  punctuations, SCONJ  subordinating conjunctions, VERB  verbs.
Back to article page