An important issue that has yet to be resolved in pain measurement lit
erature concerns the number of levels needed to assess self-reported p
ain intensity. An examination of treatment outcome literature shows a
large variation in the number of levels used, from as few as 4 (e.g.,
4-point Verbal Rating scales (VRS)) to as many as 101 (e.g., 101-point
Numerical Rating scales (NRS)). The purpose of this study was to prov
ide an empirically derived guideline for determining the number of lev
els needed. Chronic pain patients (n = 124) provided pre- and post-tre
atment measures of pain intensity using 101-point NRS for least, most,
current, and average pain. The patients' responses to the measures we
re examined closely to determine the actual number of levels used. In
addition, their responses to the 101-point scales were recoded to form
7 scales of varying levels (2- to 101-point scales). The sensitivity
of the 7 recoded scales was examined. The results indicated that littl
e information is lost if 101-point scales are coded as 11- or 21-point
scales. Moreover, examination of the actual responses to the 101-poin
t measure showed that almost all patients treated it as a 21-point sca
le by providing responses in multiples of 5 or 10, while a substantial
number of patients treated it as an ii-point scale, providing respons
es in multiples of 10 only. The results suggest that 10- and 21-point
scales provide sufficient levels of discrimination, in general, for ch
ronic pain patients to describe pain intensity.