Reliability of the PEDro scale: comparison between trials published in predatory and non-predatory journals

Table 1 Reliability of the PEDro scale assessed in the whole sample of trials

PEDro score	ICC (95% C.I.)				SEM
	0.639 (0.566–0.701)				0.795
	% agreement	k	interpretation	P + +	Sa		Sd
PEDro scale item
2 Random allocation	95.0	.01	none to slight	94.6	0.99	0.03
3 Concealed allocation	91.6	.71	substantial	13.1	-0.71	0.00
4 Groups similar at baseline	69.5	.36	fair	45.0	0.29	0.02
5 Subject blinding	93.0	.65	substantial	7.7	-0.83	0.03
6 Therapist blinding	97.3	.59	moderate	2.0	-0.96	0.00
7 Assessor blinding	88.6	.65	substantial	15.1	-0.66	0.04
8 Less than 15% dropouts	64.8	.30	fair	36.6	0.13	-0.20
9 Intention-to-treat analysis	59.7	.15	none to slight	17.1	-0.43	-0.21
10 Between-group statistical comparisons	87.9	.39	fair	82.9	0.88	0.07
11 Point measures and variability data	92.6	.55	moderate	87.2	0.88	0.01
Quality discrimination^a	73.5	.41	moderate	21.1	-0.42	-0.05

^aBased on PEDro cut score = 6 (see main text for details)
ICC Intraclass Correlation Coefficient, CI confidence intervals, SEM Standard Error of the Measurement, P + + agreement on the presence of the item, Sa symmetry expressions for agreement, Sd symmetry expressions for disagreement

ISSN: 2057-0082