4. Results
Table 1 presents the background variables, the descriptive statistics for the mediators and the independent and dependent variables of the sample. The majority of the participants were women, aged between 65 and 95 (mean = 74.3, SD = 6.3) with moderate health status (Mean = 3.61, SD = 1.15). Their education ranged from 6 to 21 years (Mean = 10.0, SD = 4.0) and most had a partner (71.1%). About 50% were Arabs and the others, Jews. No differences were found between the two ethnic groups (data not shown).
The feelings of loneliness of the participants were medium (Mean = 2.2, SD = 1.1); depressive symptoms were fairly low (Mean = 4.8, SD = 3.2); malnutrition was low-to-medium (Mean = 6.8, SD = 4.3) and subjective age during the COVID-19 pandemic was medium (Mean = 2.94, SD = 1.73).
The first hypothesis is supported by the results of the paired t-test. A difference was indicated between the two times, and the participants felt older during the pandemic (see Table 2).
The results of the bivariate tests of the association between the research variables and subjective age during the COVID-19 pandemic reveal that the only demographic variable that correlated significantly with subjective age during the crisis was marital status (having a partner correlated positively with lower subjective age, see Table 3).
The results reveal that all the independent and mediation variables were significantly related to the dependent variable. High levels of loneliness, depressive symptoms, and malnutrition correlated positively with subjective age during the pandemic.

The Mediation Analyses
Using the PROCESS model 4 [62,63], we tested the second and third hypotheses that during the crisis, feelings of loneliness would be associated directly with subjective age, and indirectly associated with subjective age through depressive symptoms and malnutrition, controlling for covariates (see Table 4 and Figure 1). The results indicate a significant total direct effect (path c; without mediators) of loneliness on subjective age, and a significant indirect effect through malnutrition as a mediator. No indirect connection was found between the second mediator (depressive symptoms) and subjective age. No significant associations were found between the background variables and subjective age.