1. Introduction

Depression represents a profound global public health challenge, affecting millions of individuals and imposing substantial personal, societal, and economic burdens (World Health Organization [WHO], 2017). As a complex and multifaceted condition, depression is shaped by an array of social determinants that interact to either exacerbate or mitigate its severity (Marmot, 2005; Lorant et al., 2003). To capture this complexity, the present study employs the Centre for Epidemiological Studies Depression Scale (CES-D8) as a standardized measure of depressive symptoms. Using data from the European Social Survey (ESS), this analysis explores how key predictors—namely, age, alcohol consumption, self-rated health, social connections, and life satisfaction—are associated with the prevalence of depressive symptoms. The aim of this study is to provide valuable insights into how social determinants impact mental health. This will establish an empirical basis for developing public health interventions and policies to reduce the burden of depression. This study focuses on a comparative European context. The study looks at how the social and cultural systems of three countries influence the link between socio demographic factors and depressive symptoms. The study adds to the growing literature identifying social determinants as central to mental health (Patel et al., 2018) and offers insights for global policymaking.

2. Literature/ Theoretical Framework

A substantial body of literature underscores the importance of several social determinants for mental health outcomes. Based on this evidence, the following key predictors are identified:

Age

Older adults often face challenges such as physical decline, social isolation, and ageism, which have been associated with increased vulnerability to depression (Fiske, Wetherell, & Gatz, 2009; WHO, 2017).

• Hypothesis 1 (H1): Older adults exhibit higher levels of depression compared to younger individuals.

Alcohol Consumption

While alcohol may be used as a coping mechanism, excessive or frequent consumption has been linked to poorer mental health outcomes (Boden & Fergusson, 201). This bidirectional relationship suggests that alcohol consumption not only reflects underlying distress but may also exacerbate depression over time.

• Hypothesis 2 (H2): Higher alcohol consumption correlates with higher depression levels.

Self-Rated Health

Self-rated health has been demonstrated to be a robust predictor of both physical and mental health outcomes. Poor self-rated health has been linked to increased depressive symptoms, as it encapsulates individuals’ overall perception of their physical well-being and its psychological impact (Jylhä, 2009).

• Hypothesis 3 (H3): Individuals with poorer self-rated health report higher levels of depression.

Social Connections

Social connections and the frequency of social interactions have been demonstrated to serve as critical buffers against mental health decline. A strong social network has been shown to be capable of mitigating feelings of loneliness and isolation, both of which are significant risk factors for depression. Conversely, infrequent social contact has been found to leave individuals more vulnerable to depressive symptoms (Cacioppo et al., 2006).

• Hypothesis 4 (H4): Less frequent social meetings are associated with higher depression levels.

Life Satisfaction

Elevated levels of life satisfaction have been demonstrated to be associated with diminished levels of depression, given that contentment with life can serve as a counteragent against negative mood states and thereby foster resilience (Pavot & Diener, 2008).

• Hypothesis 5 (H5): Higher life satisfaction is associated with lower depression levels

Country Contexts

The study focuses on three European countries: Norway, Germany and Spain. The national context significantly affects the impact of social determinants on depression. Variations in welfare systems, cultural norms and economic conditions can either mitigate or exacerbate depressive symptoms.

Norway:

• Strengths: Norway benefits from a comprehensive welfare system and universal healthcare, contributing to generally lower levels of depression among the population (Krokstad et al., 2013).

• Challenges: The sparse population can lead to social isolation, and alcohol consumption remains a salient issue, particularly among older adults

• Hypothesis 6a (H6a): In Norway, overall depression levels are relatively low; however, older adults experiencing higher alcohol consumption and social isolation will exhibit elevated depression levels.

Germany:

• Strengths: Germany benefits from a robust healthcare system, characterized by high accessibility and quality of medical services, which contributes to better overall health outcomes (Busse & Blümel, 2014)

• Challenges: a significant challenge in the German context is the cultural normalization of alcohol consumption. Alcohol use is widely accepted as part of social life, which can lead to an under-recognition of problematic drinking patterns, especially among older adults. This normalization may hinder early detection and intervention for harmful alcohol use, thereby exacerbating depressive symptoms within this vulnerable group (World Health Organization, 2018).

• Hypothesis 6b (H6b): In Germany, older adults with high alcohol consumption are expected to report elevated depression levels.

Spain

• Strengths: A strong, family-centric culture in Spain tends to provide substantial social support, which can protect against depression. Research has shown that robust family and social support networks in Spanish older adults are associated with better mental health outcomes (Zunzunegui et al., 2003)

• Challenges: On the other hand, economic pressures and related stressors— exacerbated by the recent economic crisis—may increase vulnerability to depressive symptoms among the population. Evidence from time-series analyses has demonstrated that economic downturns in Spain are linked to deteriorating mental health (Gili et al., 2013).

• Hypothesis 6c (H6c): In Spain, while close family ties may serve as a mitigating factor for depression, it is anticipated that economic stressors will have a potentiating effect on depression risks.

Research Gap

Numerous studies have explored the social determinants of depression. However, there is a relative paucity of cross-country analyses that leverage large-scale datasets such as the European Social Survey (ESS11). This study aims to address this gap by examining how cultural and social factors differentially influence depressive symptoms across Norway, Germany, and Spain.

Summary of Hypotheses

• H1: Older adults report higher depression levels than younger individuals.

• H2: Higher alcohol consumption correlates with higher depression levels.

• H3: Individuals with poorer self-rated health report higher depression levels.

• H4: Less frequent social meetings are associated with higher depression levels.

• H5: Higher life satisfaction is associated with lower depression levels.

• H6a: In Norway, despite a generally lower level of depression due to a comprehensive welfare system, older adults with higher alcohol consumption and social isolation will exhibit increased depression.

• H6b: In Germany, older adults with high alcohol consumption are expected to report elevated depression levels.

• H6c: In Spain, robust family bonds may mitigate depression, whereas economic stressors will amplify the risk of depressive symptoms.

3. Methods

3.1 Data Source and Sample

# Our Selected Countries
# Filter for Norway, Germany, and Spain
df <- df[df$cntry %in% c("Norway", "Germany", "Spain"), ]

The present study utilises secondary data from the ESS Round 11. The sample was filtered to include respondents from Norway, Germany, and Spain, yielding a final sample of n=5601 observations.

3.2 Variables and Measurement

Dependent Variable:

• Depression Scale: A composite measure of depressive symptoms was constructed using eight self-reported items related to emotional well-being (fltdpr, flteeff, slprl, wrhpp, fltlnl, enjlf, fltsd, and cldgng). The ‘enjlf’ item was reverse-coded so
that all items are oriented in the same direction, with higher scores reflecting greater depressive symptoms. The final scale was computed as the row-wise mean of these items, and its internal consistency was confirmed using Cronbach’s alpha.

# MULTI-ITEM SCALE: DEPRESSION
# Create background variable "depression scale" from D20-D27
# Variables:
# fltdpr: Felt depressed D20
# flteeff: Felt everything was an effort D21
# slprl: Sleep was restless D22
# wrhpp: Could not get going D23
# fltlnl: Felt lonely D24
# enjlf: Enjoyed life D25
# fltsd: Felt sad D26
# cldgng: Felt discouraged D27

# Convert variables to numbers
df$d20 = as.numeric(df$fltdpr)    # D20
df$d21 = as.numeric(df$flteeff)   # D21
df$d22 = as.numeric(df$slprl)     # D22
df$d23 = as.numeric(df$wrhpp)     # D23
df$d24 = as.numeric(df$fltlnl)    # D24
df$d25 = as.numeric(df$enjlf)     # D25
df$d26 = as.numeric(df$fltsd)     # D26
df$d27 = as.numeric(df$cldgng)    # D27

# Reverse scale of d25 (enjlf)
# Reverse-coding aligns d25 (enjlf) with the depression scale (higher = more depressed)
df$d25 = 6 - df$d25

# CRONBACH'S ALPHA: DEPRESSION SCALE
# Compute variances of individual items
item_variances = sum(apply(df[, c("d20", "d21", "d22", "d23", "d24", "d25", "d26", "d27")], 2, var, na.rm = TRUE))

# Compute variance of the total scale
# The total variance is the variance of the sum of all depression-related items (D20-D27).
total_variance = var(rowSums(df[, c("d20", "d21", "d22", "d23", "d24", "d25", "d26", "d27")], na.rm = TRUE), na.rm = TRUE)

# Cronbach's alpha calculation
n_items = 8  # Number of items (D20-D27)
item_variances = sum(apply(df[, c("d20", "d21", "d22", "d23", "d24", "d25", "d26", "d27")], 2, var, na.rm = TRUE))
cronbach_alpha = (n_items / (n_items - 1)) * (1 - item_variances / total_variance)

To assess the internal consistency of the depression scale, Cronbach’s alpha was calculated. An optimal value lies between 0.8 and 0.92, indicating that the items reliably measure the same underlying construct without being redundant. Values below 0.6 suggest weak reliability, while values close to 1 may indicate redundancy among items.

In our case, the calculated Cronbach’s alpha is 0.6346298, which indicates moderate internal consistency. While this value is not ideal, it suggests that the eight items of the CES-D8 scale are reasonably related, though there may be room for improvement in the coherence of the scale.

• Age: Age in years is treated as a continuous variable.

# Convert factors to characters before converting to numeric to avoid undeclared factor levels issues
df$agea <- as.numeric(as.character(df$agea))

• Alcohol Consumption (alcfreq): Recoded into a numeric scale from 0 (“Never”) to 6 (“Every day”)

# Convert Alcohol Frequency
df$alcfreq <- as.numeric(factor(df$alcfreq, levels = c(
  "Never", "Less than once a month", "Once a month", "Several times a month", 
  "Once a week", "Several times a week", "Every day"), labels = c(0,1,2,3,4,5,6)))

• Self-Rated Health (health): Recoded on a 5-point scale from 1 (“Very bad”) to 5 (“Very good”)

# Convert Health (Self-Rated)
df$health <- as.numeric(factor(df$health, levels = c(
  "Very bad", "Bad", "Fair", "Good", "Very good"), labels = c(1,2,3,4,5)))

• Social Connections (sclmeet): Frequency of social meetings on a 7-point scale (0 “Never” to 6 “Every day”)

# Convert Social Connections (sclmeet)
df$sclmeet <- as.numeric(factor(df$sclmeet, levels = c(
  "Never", "Less than once a month", "Once a month", "Several times a month",
  "Once a week", "Several times a week", "Every day"), labels = c(0,1,2,3,4,5,6)))

• Life Satisfaction (stflife): Measured on an 11-point scale from 0 (“Extremely dissatisfied”) to 10 (“Extremely satisfied”)

# Convert Life Satisfaction (stflife)
df$stflife <- as.numeric(factor(df$stflife, levels = c(
  "Extremely dissatisfied", "2", "3", "4", "5", "6", "7", "8", "9", "10", "Extremely satisfied"), labels = c(0,1,2,3,4,5,6,7,8,9,10)))

3.3 Data Preparation and Processing

The dataset was cleaned several times for preparing the same on a statistical analysis platform. In specific:

• Selection of Countries: The dataset had been filtered to include only the respondents pertaining to Norway, Germany, and Spain.

# My Selected Countries
# Filter for Norway, Germany, and Spain
df <- df[df$cntry %in% c("Norway", "Germany", "Spain"), ]

• Handling Missing Data: It identified the missing values and did median imputation to retain maximum observations without significant distortion of the data.

# check missing values
colSums(is.na(df[, c("agea", "alcfreq", "health", "sclmeet", "stflife")]))
##    agea alcfreq  health sclmeet stflife 
##      16     776       4       3      36
# replace missing values with median
df$agea[is.na(df$agea)] <- median(df$agea, na.rm = TRUE)
df$alcfreq[is.na(df$alcfreq)] <- median(df$alcfreq, na.rm = TRUE)
df$health[is.na(df$health)] <- median(df$health, na.rm = TRUE)
df$sclmeet[is.na(df$sclmeet)] <- median(df$sclmeet, na.rm = TRUE)
df$stflife[is.na(df$stflife)] <- median(df$stflife, na.rm = TRUE)
df$depression[is.na(df$depression)] <- median(df$depression, na.rm = TRUE)

# Final Check
str(df[, c("agea", "alcfreq", "health", "sclmeet", "stflife")])
## 'data.frame':    5601 obs. of  5 variables:
##  $ agea   : num  34 42 41 58 40 46 56 63 24 57 ...
##  $ alcfreq: num  2 6 3 5 6 3 3 2 5 1 ...
##  $ health : num  4 5 2 4 5 4 5 3 4 4 ...
##  $ sclmeet: num  6 4 3 5 5 5 5 3 6 4 ...
##  $ stflife: num  8 7 11 11 8 8 7 9 9 9 ...
summary(df[, c("agea", "alcfreq", "health", "sclmeet", "stflife")])
##       agea         alcfreq          health         sclmeet     
##  Min.   :15.0   Min.   :1.000   Min.   :1.000   Min.   :1.000  
##  1st Qu.:35.0   1st Qu.:2.000   1st Qu.:3.000   1st Qu.:4.000  
##  Median :51.0   Median :3.000   Median :4.000   Median :6.000  
##  Mean   :49.8   Mean   :3.586   Mean   :3.774   Mean   :5.168  
##  3rd Qu.:65.0   3rd Qu.:5.000   3rd Qu.:4.000   3rd Qu.:6.000  
##  Max.   :90.0   Max.   :7.000   Max.   :5.000   Max.   :7.000  
##     stflife      
##  Min.   : 1.000  
##  1st Qu.: 7.000  
##  Median : 8.000  
##  Mean   : 7.805  
##  3rd Qu.: 9.000  
##  Max.   :11.000

• Recoding and Scaling: Some of the categorical variables were converted to numeric scales for regression analysis.

# Convert variables to numbers
df$d20 = as.numeric(df$fltdpr)    # D20
df$d21 = as.numeric(df$flteeff)   # D21
df$d22 = as.numeric(df$slprl)     # D22
df$d23 = as.numeric(df$wrhpp)     # D23
df$d24 = as.numeric(df$fltlnl)    # D24
df$d25 = as.numeric(df$enjlf)     # D25
df$d26 = as.numeric(df$fltsd)     # D26
df$d27 = as.numeric(df$cldgng)    # D27

# Reverse scale of d25 (enjlf)
# Reverse-coding aligns d25 (enjlf) with the depression scale (higher = more depressed)
df$d25 = 6 - df$d25

• Construction of Depression Scale: The items measuring symptoms of depression have been summed for an average score, taking into consideration reverse coding whenever the items were not in the direction of the scale.

# Compute the depression scale
# Score = mean of item values row-wise = sum of item values / number of items 
df$depression = rowSums(df[, c("d20", "d21", "d22", "d23", "d24", "d25", "d26", "d27")], na.rm = TRUE) / 8

# replace of the remaining missing values in depression
df$depression[is.na(df$depression)] <- median(df$depression, na.rm = TRUE)

# Done. Show descriptives:
summary(df$depression)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   1.625   1.875   1.908   2.125   3.875

During data preparation, some challenges arose, such as undeclared factor levels and warnings about missing values. These were resolved by painstaking recording and data type changes to make them suitable for the proposed analyses.

#checking warnings
warnings()

3.4 Statistical Analysis

The hypotheses were tested using the following statistical procedures:

• Descriptive Statistics: Means, medians, and frequency distributions were determined for all important variables.

• Bivariate analysis: Using Pearson’s product-moment correlation tests, an exploratory look into the association between depression with each independent variable.

• Multiple Linear Regression: A set of regression models examining the age, alcohol consumption, self-rated health, social connections, and life satisfaction effects on the level of depression.

• Moderation analysis: Analyzed whether the association between depression and key independent variables varied across countries by introducing interaction terms into the regression models.

All analyses were performed in R, and variable distributions were evaluated with the use of histograms. The level of significance was set at p < 0.05.

4. Results

This section presents the descriptive statistics, bivariate associations, and multivariate regression results based on the ESS data from Norway, Germany, and Spain. The primary focus of the analyses was on the relationships between depression and the following predictors: age, alcohol consumption, self-rated health, social connections, and life satisfaction.

4.1 Sample Description

• Age (agea):

Mean = 49.8009284;

Median = 51

hist(df$agea, 
     main = "Age Distribution", 
     xlab = "Age", 
     col = "lightgreen")

• Depression Scale (depression):

Mean = 1.9075835;

Median = 1.875

hist(df$depression, 
     main = "Distribution of Depression Scale", 
     xlab = "Depression Score", 
     col = "lightblue", 
     breaks = 8)

• Alcohol Consumption (alcfreq):

Mean = 3.5859668;

Median = 3

• Self-Rated Health (health):

Mean = 3.774326;

Median = 4

• Social Connections (sclmeet):

Mean = 5.1676486;

Median = 6

• Life Satisfaction (stflife):

Mean = 7.8053919;

Median = 8

4.2 Multivariate Models and Moderation by Country

In order to examine the combined effects of the predictors and to test whether the relationships vary by country, several regression models with interaction terms were estimated.

# REGRESSION ANALYSIS: PREDICTING DEPRESSION
model_depression <- lm(depression ~ agea + alcfreq + health + sclmeet, data = df)
summary(model_depression)
## 
## Call:
## lm(formula = depression ~ agea + alcfreq + health + sclmeet, 
##     data = df)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1.58094 -0.22813 -0.05037  0.17422  1.92985 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  2.9572560  0.0316627  93.399  < 2e-16 ***
## agea        -0.0037441  0.0002641 -14.176  < 2e-16 ***
## alcfreq     -0.0148245  0.0024150  -6.138  8.9e-10 ***
## health      -0.1675063  0.0053750 -31.164  < 2e-16 ***
## sclmeet     -0.0344117  0.0033781 -10.187  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.3495 on 5596 degrees of freedom
## Multiple R-squared:  0.1901, Adjusted R-squared:  0.1895 
## F-statistic: 328.3 on 4 and 5596 DF,  p-value: < 2.2e-16
# focussing on coeffients
coefficients(model_depression)
##  (Intercept)         agea      alcfreq       health      sclmeet 
##  2.957256006 -0.003744137 -0.014824459 -0.167506324 -0.034411699

5. Limitations

The study is subject to several limitations that should be acknowledged. Firstly, the cross sectional nature of the ESS data limits the capacity for causal inferences to be made. Secondly, the unexpected negative associations observed for age and alcohol consumption indicate the potential involvement of unmeasured variables (e.g. socioeconomic status, cultural factors). The collection of longitudinal data in future research would facilitate the exploration of these potential moderating factors and enhance the capture of depression dynamics over time.

In conclusion, while the analyses largely support the theoretical importance of self-rated health, social connections, and the impact of depression on life satisfaction, the findings regarding age and alcohol consumption warrant further investigation. The moderation effects by country further emphasise the role of cultural and contextual factors in shaping mental health outcomes.