Table 3.

Data quality, scaling assumptions, acceptability, and reliability of the FSS

Psychometric property	FSS total
Data quality (n = 553)
Subjects with missing items, n (%)	2 (0.4)
Number of missing items, n (%)	2 (<0.01)
Computable scale scores, n (%)	551 (99.6)
Scaling assumptions (n = 551)
Item mean score, range	3.9–5.2
Item SD, range	1.8–2.2
Item skewness: range	−0.940–0.020
Item to total correlation, range	0.626–0.875
Item 1	0.626
Item 2	0.707
Item 3	0.846
Item 4	0.875
Item 5	0.854
Item 6	0.837
Item 7	0.866
Item 8	0.771
Item 9	0.795
Acceptability
Possible score range	1–7
Observed score range	1–7
Score, mean (SD)	4.5 (1.7)
Floor, n (%)	14 (2.5)
Ceiling, n (%)	19 (3.5)
Skewness	−0.5
Reliability
Cronbach's alpha (95% CI)
Entire sample	0.949 (0.942–0.955)
Cronbach's alpha when one item deleted: range	0.939–0.951
Cronbach's alpha (95% CI)
Age groups
<40 years	0.950 (0.930–0.966)
40–49 years	0.941 (0.924–0.956)
50–59 years	0.951 (0.939–0.961)
60–69 years	0.959 (0.947–0.969)
≥70	0.942 (0.912–0.964)
Gender groups
Female	0.950 (0.942–0.957)
Male	0.946 (0.930–0.960)
EDDS groups
0–3	0.952 (0.943–0.961)
4–6.5	0.935 (0.922–0.947)
7–9	0.957 (0.942–0.970)
Disease phenotype groups
Relapsing‐remitting (95% CI)	0.944 (0.932–0.954)
Secondary progressive (95% CI)	0.940 (0.925–0.953)
Primary progressive (95% CI)	0.957 (0.943–0.969)
Unknown disease phenotype (95% CI)	0.954 (0.934–0.971)

CI, confidence interval; FSS, Fatigue Severity Scale; SD, standard deviation.