3.3. Hierarchical regression analysis
Drawing upon the stress and coping model, hierarchical linear
regression was used to analyze our data. Two analyses were carried
out, considering burden and mental health as dependent variables.
In both analyses, contextual variables (caregiver’s age, hours spent
caring per day and time spent as a caregiver) were included in the
first step. Variables usually considered as indicators of stressors or
substantial demands on caregivers are included in the second step
(functional and cognitive status of the care recipient). Social
support, one of the most widely analyzed mediating variables, is
included in the third step; finally, caregiver’s leisure time was
included in the last step. The mediating variables assessed were
included in two separate steps in order to ascertain whether
caregivers’ perception of their free time adds significant explained
variance to the dependent variables analyzed (burden and mental
health) once social support has been controlled. The results of
these regression analyses are shown in Table 4.