constructs of the 4 questionnaires showed acceptable consistency
(Cronbach's alpha ¼ 0.615). Next, the non-parametric KruskaleWallis
(KW) test was used to assess statistical differences on
the sample means. The option for the KW test is justified by the fact
that the sample is small, and there is no certainty regarding normal
distribution of data. The KW test does not require a normal distribution
of the residuals as the analogous one-way ANOVA does
(Corder and Foreman, 2009). Tukey's test was used to point out the
sectors that are significantly different from the others. Questions
related to workstation, work organization and company constructs
reflected EDIs particular to each sector, do not enabling direct
comparison among them. Comparative results presented in this
article were therefore drawn based on EDIs from physical environment,
work content and discomfort/pain constructs that were
common to all four evaluated sectors. Managers and supervisors
were not interviewed and did not answer questionnaires, therefore
their opinions regarding work performed in the sectors as well as
the possible solutions for improvement were obtained through
dialogues and discussions held on many occasions during the
interview process.