Between-group differences in CRI and DASS-21 scores
The whole sample’s mean CRIq score was 118, with a significant difference between the two age groups (t(93)=-5.21, p<0.001) because older adults showed higher CRIq scores compared to Middle-aged (see Figure 1).
The mean total score of DASS was 8.7 (SD=7.7), with middle agers reporting a higher number of mental health symptoms (mean=9.9, SD=7.7) than older participants (mean=6.8, SD=7.4), although this difference was not statistically significant (t(94)=1.97, p=0.051).
A significant difference between the two groups emerged when comparing the DASS-Stress scores (t(93)=3.04, p=0.003), with Middle-aged showing a higher score (i.e. higher amount of reported stress symptoms) relative to older participants (respectively 5 for middle agers and 2.8 for older adults). No difference emerged in the other subscales (see Table 1).
Table 1. Descriptive statistics of the whole sample and of the two sub-groups (Middle Agers and Older). Middle Agers: N=57; 40-59 years old; Older: N=39; 60-80 years old. Means and standard deviations in brackets. The CRI is classified into five levels: Low (<70), Medium-low (70-84), Medium (85-114), Medium-high (115-130) and High (>130). Higher DASS scores corresponded to more mental health reported symptoms (DASS range: 0-126; DASS subscales range: 0-42).
|
|
Age
|
CRI-tot
|
CRI-
Education
|
CRI-
Work
|
CRI-
Leisure
|
DASS
Total
|
DASS
Stress
|
DASS
Anxiety
|
DASS
Depression
|
Mean
(SD)
|
Global
|
55.6 (11)
|
118.5 (16.8)
|
119 (13)
|
108.9 (19.3)
|
114.8 (19.5)
|
8.7 (7.7)
|
4.1 (3.7)
|
1.5 (2)
|
3.1 (3.7)
|
Middle Age
|
48.2 (6)
|
112 (13.8)
|
114.2 (10.1)
|
104.6 (14.9)
|
108.9 (15.6)
|
9.9 (7.7)
|
5 (3.5)
|
1.7 (2.3)
|
3.2 (3.7)
|
Older
|
68.7 (6)
|
128.3 (16.4)
|
126.2 (13.7)
|
115.3 (23.2)
|
123.7 (21.5)
|
6.8 (7.4)
|
2.8 (3.5)
|
1 (1.3)
|
3 (3.9)
|
Range
|
Global
|
40-80
|
85-162
|
91-152
|
71-162
|
76-179
|
0-37
|
0-15
|
0-15
|
0-16
|
Middle Age
|
40-59
|
85-162
|
91-141
|
80-147
|
76-157
|
0-37
|
0-14
|
0-15
|
0-16
|
Older
|
60-80
|
90-160
|
91-152
|
71-162
|
85-179
|
0-29
|
0-15
|
0-4
|
0-16
|
Notes:
CRI= Cognitive Reserve Index (Nucci et al., 2012);
DASS=Depression Anxiety Stress Scale (DASS-21; Short Version; Lovibond and Lovibond, 1995; Henry and Crawford 2005).
Relationship between CRI and DASS-21 scores
A first correlation analysis based on the whole sample indicated that DASS-total score negatively correlated with the CRI score (ρ=-0.36, p<0.001), such as higher CR levels were associated with lower symptoms of depression, anxiety and stress, and therefore with better mental health.
An equivalent analysis yielded significant correlations between Age and DASS total score (ρ=-0.30, p=0.002), as well as Age and CRI (ρ=0.54, p<0.001), such as higher age was associated with both better mental health and higher CR levels.
Results of the regression analysis confirmed those based on the correlation approach, indicating that Age and CRI significantly predicted DASS total score, accounting for 11.5% (R2=0.115) of the variance across the entire sample (F(92, 2)=5.98, p=0.003). When analysing the contribution of the individual predictors in the model, we found that CRI was a significant predictor of DASS scores (βCRI=-0.14, t(92)=-2.61, p=0.011). Specifically, lower DASS scores (i.e. lower depression, anxiety and stress reported symptoms) emerged in older participants, as well as in participants with higher CRI levels. Although Age and CRI correlated significantly, no multicollinearity arose in the regression analyses (Spermann’s ρ = 0.54), indicating that the two variables independently contributed to the models.
Additional models were conducted for the independent CRIq components, with results showing that only CRI-WorkingActivity had a significant effect on DASS total score. Age and CRI-WorkingActivity significantly predicted DASS total score, accounting for 10.5% (R2=0.105) of the variance across the entire sample (F(92, 2)=6.49, p=0.002). Likewise, DASS scores were also significantly predicted by CRI-WorkingActivity (βCRI-WorkingActivity=-0.11, t(92)=-2.79, p=0.006), whereas CRI-Education and CRI-LeisureTime did not (respectively: F(92, 2)=3.03, p=0.278, and F(92, 2)=3.24, p=0.210).
Subsequent regression models with the three separate DASS sub scores as dependent variables indicated that CRI and Age significantly predicted only the scores of the DASS Stress subscale, such as lower DASS-Stress scores corresponded to higher age and higher CRI. In particular, Age and CRI significantly predicted DASS stress subscale, accounting for 17.1% (R2=0.171) of the variance across the entire sample (F(92, 2)=9.43, p<0.001). An analysis looking at the contribution of the individual predictors in the model showed that age and CRI were significant independent predictors (respectively: βage=-0.07, t(92)=-2.13, p=0.036 and βCRI=-0.05, t(92)=-2.19, p=0.031).
Strikingly, the CRI subscale with the strongest effect on the DASS-Stress scale was the CRI-Working Activity. In particular, Age and CRI significantly predicted DASS stress subscale accounting for 19.4% (R2=0.194) of the variance across the entire sample (F(92, 2)=10.98, p<0.001). Both age and CRI-WorkingActivity were significant independent predictors (βage=-0.08, t(92)=-2.76, p=0.006, and βCRI-WorkingActivity=-0.05, t(92)=-2.74, p=0.007).
Given the significant effect of age in the total sample, separate linear regression models for Middle-age and Older groups were subsequently run to investigate potential age-related differences. Results of these models revealed that while CRI total score significantly predicted DASS total score in both groups (Middle Age: F(48, 1)=3.06, R2=0.066; βCRI=-0.145, t(48)=-1.75, p=0.05; Older Adults: F(36, 1)=4.11, R2=0.103; βCRI=-0.144, t(48)=-2.03; , p=0.049), CRI-WorkingActivity significantly predicted DASS-Stress scores only in the Middle-age group (Middle Age: F(48, 1)=4.81, R2=0.091; βCRI-WorkingActivity=-0.07, t(48)=-2.19, p=0.033; Older Adults: F(36, 1)=3.43, R2=0.087; βCRI-WorkingActivity=-0.045, t(48)=-1.85; p=0.07; see Figure 1).