Psychometric properties of the original and short versions of the Falls Efficacy Scale-International (FES-I) in people with Parkinson’s disease

Table 4 Psychometric comparison of the original and short FES-I, n = 101

	Original FES-I	Short FES-I
Missing item responses	0.9%	0.7%
Corrected item-total correlation, min-max	0.59 (item 6)–0.85 (items 1 and 7)	0.63 (item 16)–0.78 (item 7)
Total scores, n	92	96
Mean (SD)	30 (12.0)	13 (5.1)
Min-Max	16–59	7–25
Skewness	0.72	0.71
Floor/ceiling effects	10/0	19/0
Cronbach’s alpha	0.96	0.89
ICC (95% CI)	0.92 (0.88–0.95)	0.91 (0.86–0.94)
SEM^a (% of possible scoring range)	3.4 (7)	1.6 (7)
SEM^b (% of possible scoring range)	2.5 (5)	1.7 (8)
SDD^c (% of possible scoring range)	9.6 (20)	4.3 (20)
SDD^d (% of possible scoring range)	6.8 (14)	4.7 (21)

FES-I Falls Efficacy Scale-International, SEM Standard Error of Measurement, SDD Smallest Detectable Difference
Possible scoring ranges: Original FES-I 16–64, Short FES-I 7–28; higher = worse
^aBased on ICC, using the formula SEM = SD _baseline \( \times \sqrt{1- ICC} \)
^bBased on Cronbach’s alpha, using the formula SEM = SD _baseline \( \times \sqrt{1- Cronbac{h}^{\prime } s\ alpha} \)
^cBased on the ICC-based SEM, using the formula \( SDD= SEM\times 1.96\times \sqrt{2} \)
^dBased on the Cronbach’s alpha-based SEM, using the formula\( SDD= SEM\times 1.96\times \sqrt{2} \)

ISSN: 1477-7525