Table 1 represents the descriptive results for the indicators of this study. As observed, the mean scores for positive emotions (M = 6.71) and quality of life (M = 7.63) present a mean above the median of the scale range assessed as opposed to negative emotions (M = 5.18), fatalism (M = 2.68) and loneliness (M = .58), whose values have a lower mean compared to the range median of these indicators. Comparing variable means according to the sex of participants, no significant differences are observed between both groups; except in the case of negative emotions, where the mean for females (M = 5.38) is statistically higher than for males (M = 4.82).
Table 1
Descriptive analysis of study variables (N = 1036)
| | | Range | | Overall Mean (SD) | Male Mean (SD) | Female Mean (SD) | p-value |
Quality of life | | | (1–10) | | 7.63 (1.46) | 7.69 (1.41) | 7.59 (1.48) | .28 |
Negative emotions | | | (1–10) | | 5.18 (2.31) | 4.82 (2.30) | 5.38 (2.28) | .00 |
Positive emotions | | | (1–10) | | 6.71 (1.85) | 6.63 (1.85) | 6.75 (1.84) | .33 |
Fatalism | | | (1–5) | | 2.68 (.80) | 2.68 (.86) | 2.69 (0.77) | .95 |
Loneliness | | | (1–3) | | .58 (.28) | .59 (.29) | .57 (.28) | .47 |
Prior to the linear regression and moderation analyses, a correlation analysis was performed between the variables in the hypothesis of this study. The results of Pearson correlations are presented in Table 2. As observed, all the relationships between variables are significant. (p < .01). However, based on the standard interpretation of effect size (Cohen, 1988), this is high in the association between quality of life and positive emotions (r = .54), while correlation coefficients are moderate in the associations between loneliness and quality of life (r = − .49), loneliness and positive emotions (r = − .38), loneliness and negative emotions (r = .39), fatalism and negative emotions (r = .30), and quality of life and positive emotions (r = − .39). The rest of associations present a small association level.
Table 2
Analysis of correlations between the variables of the model
| 1 | 2 | 3 | 4 | 5 |
1. Loneliness | - | .23** | − .49** | − .38** | .39** |
2. Fatalism | | - | − .26** | − .21** | .30** |
3. Quality of life | | | - | .54** | − .39** |
4.Positive emotions | | | | - | − .16** |
5. Negative emotions | | | | | - |
**. The correlation is significant at the 0.01 level (bilateral). |
Table 3 presents the results of the analysis and control variable regressions over quality of life. Model 1 only comprises the control variables of sex and age. For this model, only the age variable results (β = .03, p < .05). In turn, Model 2 employs the variables of the first model but also included the independent variables of fatalism about COVID-19 and loneliness separately. In the case of fatalism, β = − .29 and p < .001 are reported, while for loneliness these values are β = -2.32, and p < .001.
Model 3 conducted calculations through Hayes’ PROCESS macro in order to assess the moderating effect of loneliness over the relationship between fatalism about COVID-19 and the quality-of-life indicator. The combined effect of fatalism and loneliness was negative (β = −.43, p < .01), and the regression coefficient for fatalism about COVID-19 indicator was not significant (β = −.03, p > .05) but it was for loneliness (β = −1.19, p < .01) The slope test in Table 4 indicates that the impact of fatalism on quality of life is strengthened as reported loneliness increases.
Table 3
Linear regression and Hayes’ linear regression analyses considering loneliness a moderator (dependent variable = quality of life)
| Model 1 | Model 2 | Model 3 |
β | t | β | t | Β | t |
Age | .03* | 2.22 | .003 | .25 | .01 | .50 |
Sex | .08 | .83 | .13 | 1.59 | .14 | 1.66 |
Fatalism | | | − .29*** | -5.66 | − .03 | − .27 |
Loneliness | | | -2.32*** | -16.35 | -1.19** | -2.61 |
Fatalism vs loneliness | | | | | − .43** | -2.60 |
R2 | .04 | .26 | .27 |
F(df1, df2) | 3.00(2,1029) | 91.79 (4,1021) | 75.2(5,1020) |
*p < .05, **p < .01, ***p < .001
Table 4
Slope test analysis of the conditional effects of the moderator (loneliness)
Conditional effects of moderator at M ± 1 SD (slope test) | Effect | SE | T | p |
Loneliness Low − 1 SD (-0.28) | − .16 | .07 | -2.21 | .03 |
Loneliness Medium M (0.00) | − .28 | .05 | -5.57 | .00 |
Loneliness High + 1SD (0.28) | − .40 | .07 | -5.97 | .00 |
Likewise, the graphic analysis of the moderation in Fig. 1 indicates that the relationship between fatalism and quality of life presents a more pronounced slope with high levels of loneliness compared to the slope of this relationship when a low value of loneliness is used as a reference.
---------------------------------Please insert Fig. 1 here------------------------------------------
Table 5 presents the regression coefficients of the variables considered in the hypotheses of this study about the negative emotions’ indicator. For Model 1, which considers only the control variables, both variables are significant. In the case of age, β = − .07, p < .01, while for sex β = − .50, p < .01. Model 2 adds the variables of fatalism about COVID-19 and loneliness. The regression coefficient for the age variable in this model is β = − .04, p < .05, while for the sex variable is β = − .56, p < .001. Regarding independent variables, fatalism about COVID-19 exhibits a significant relationship (β = .62, p < .001), as well as loneliness (β = 2.69, p < .001).
Finally, Model 3, based on Hayes’ linear regression analysis, shows that most indicators are significant, except for the variable age (β = −.03, p > .05). This model incorporates the interaction variable of fatalism about COVID-19 vs. loneliness which yields a significant coefficient of β = − .53, p < .05. Table 6 presents the conditional effects of the moderating variable, where the impact of fatalism about COVID-19 on experienced negative emotions is weakened with the increase of the loneliness indicator.
Table 5
Linear regression and Hayes’ linear regression analyses considering loneliness a moderator (Dependent variable: negative emotions)
VD = negative emotions | Model 1 | Model 2 | Model 3 |
β | t | β | t | Β | t |
Age | − .07** | -3.27 | − .04* | .02 | − .03 | -1.77 |
Sex | − .50** | -3.38 | − .56*** | -4.17 | − .53** | -4.13 |
Fatalism | | | .62*** | 7.61 | .66*** | 7.68 |
Loneliness | | | 2.69*** | 11.60 | 2.58*** | 11.46 |
Fatalism vs loneliness | | | | | − .53* | -1.40 |
R2 | .02 | .21 | .21 |
F(df1, df2) | 12.13(2,1028) | 68.80(4,1021) | 55.48(5, 1020) |
*p < .05, **p < .01, ***p < .001
Table 6
Slope test corresponding to the conditional effects of the moderator (loneliness)
Conditional effects of moderator at M ± 1 SD (slope test) | Effect | SE | t | p |
Soledad Low − 1 SD (-0.29) | .81 | .12 | 7.03 | .00 |
Soledad Medium M (0.00) | .66 | .08 | 8.12 | .00 |
Soledad High + 1SD (0.29) | .51 | .11 | 4.74 | .00 |
Finally, Fig. 2 presents a moderation graph. The relationship between fatalism about COVID-19 and negative emotions has a steeper slope when lower levels of loneliness exist, compared with a straight line when there are higher levels of loneliness, for this case, the line is less steep.
---------------------------------Please insert Fig. 2 here------------------------------------------
The balance of affects presents both negative and positive emotions. Table 7 shows the results of the regression analyses when positive emotions are considered the dependent variable. When they are considered control variables (Model 1), both are not significant. Likewise, when incorporating the dependent variables separately into Model 1 (Model 2), the regression coefficient for the fatalism about COVID-19 indicator is negative and significant (β = − .29, p < .001), as is the case of loneliness (β = -2.27, p < .001).
Finally, Model 3, based on Hayes analysis, indicates that interaction of independent variables is not significant; however, the regression coefficients are still significant for fatalism about COVID-19 and loneliness at the individual level. In the case loneliness about COVID-19, these are β = -2.37, p < .001, while the values are β = − .31, p < .01 for fatalism about COVID-19.
Table 7
Linear regression and Hayes’ linear regression analyses considering loneliness as a moderator (Dependent variable: positive emotions)
| Model 1 | Model 2 | Model 3 |
Β | t | β | t | β | t |
Age | .02 | 0.72 | − .02 | − .89 | − .01 | − .89 |
Sex | − .07 | -0.03 | − .07 | − .62 | − .07 | − .62 |
Fatalism | | | − .29*** | -4.40 | − .31** | .04 |
Loneliness | | | -2.27*** | -11.98 | -2.37*** | .00 |
Fatalism vs loneliness | | | | | .03 | .91 |
R2 | .00 | .16 | .16 |
F(df1, df2) | .75(2,1028) | 49.49(4,1021) | 39.56(5, 1020) |
*p < .05, **p < .01, ***p < .001