4.5. Study Limitations. This observational studywas based on
a small number of participants. Accordingly, the data is not
necessarily representative for other individuals with SCI.
For reasons of comparison across the various aspects
observed and the documented changes over time parametric
descriptive statistics were used, that is, mean and 95%
confidence intervals. For the user satisfaction questionnaire
with its ordinal scale items equal distances between
response levels around the centre of 3 were assumed.
For the SF-12v2 data, the suggested normalization algorithm
was applied; the algorithm used data from the 1998
general US population while here German participants
had been enrolled.