3. Results
3.1. Normality and outlying data
Prior to the analysis, and on the basis of Tabachnick and Fidell’s
(1996) criteria, data were screened for univariate and multivariate
outliers and for normality. Functional status scores were negatively
skewed, so that a reflect and square root transformation of
this variable was carried out. Negative kurtosis was found for the
cognitive status scores, and a reflect and logarithm transformation
was performed. Mental health scores were positively skewed, and
a square root transformation was carried out. As regards outlying
data, one univariate outlier was found for the leisure time variable,
and was deleted. No multivariate outliers were found (Mahalanobis
distance at p < 0.001). Sociodemographic data for the
resulting sample are shown in Table 1 and the descriptive statistics
for the assessed variables are in Table 2.