Skip to main content
. 2016 Oct 24;11(10):e0164573. doi: 10.1371/journal.pone.0164573

Table 3. Reporting methods and statistical analysis: Prevalence of use.

Details VAS-P(%) RMDQ(%) Totals(%)
Usage 205 (100) 152 (100) 357 (100)
Reporting method
Mean or mean difference 179 (87) 126 (83) 300 (85)
P-values 162 (79) 109 (72) 271 (76)
Standard deviation 134 (65) 98 (64) 232 (65)
Median 22 (11) 14 (9) 36 (10)
Range or IQR 20 (10) 13 (9) 33 (9)
Standard error 18 (9) 18 (12) 36 (10)
Confidence intervals 62 (30) 76 (50) 138 (39)
Number/proportion improved* 20 (10) 8 (5) 28 (8)
Number needed to treat 2 (1) 5 (3) 7 (2)
Odds ratio (improvement) 3 (1) 9 (6) 12 (3)
Relative risk (improvement) 1 (0) 6 (4) 7 (2)
Percentage change score 25 (12) 15 (10) 40 (11)
Standardised mean difference 9 (4) 11 (7) 20 (6)
Graphs and tables
Table 174 (85) 119 (78) 293 (82)
Line chart 60 (29) 39 (26) 99 (28)
Bar chart 29 (14) 9 (6) 38 (11)
Other 8 (4) 9 (6) 17 (5)
Statistical analysis
ANCOVA regression / mixed model 53 (25) 64 (42) 117 (33)
ANOVA/MANOVA regression / mixed model 74 (36) 39 (26) 113 (32)
t test 93 (45) 43 (28) 136 (38)
Non-parametric test 68 (33) 37 (24) 105 (29)
Other 18 (9) 9 (6) 27 (8)
Within-group analysis only 11 (5) 3 (2) 14 (4)
Test not described 10 (5) 10 (7) 20 (6)

VAS = Visual analogue scale

RMDQ = Roland-Morris Disability Questionnaire

IQR = Interquartile range

* Where ‘improvement’ is defined by the change in score of a specified magnitude

e.g. Wilcoxon, Friedman’s, Mann Whitney U, or Kruskal Wallis