Final Paper
2025-29-06
Authors: Olga Kricsfalusi, Lotte Pijpers, Lina-Marie Schwietale
Alcohol Consumption and Smoking Behavior as Influencing Factors of Depressive Symptoms and Gender Differences: A Study on the German Population
1. Introduction and Literature Review
Depression, also known as depressive disorder, is a significant mental health condition that profoundly affects individuals’ wellbeing and daily lives worldwide (WHO, 2023a). It arises from a combination of biological and social factors. According to the WHO (2023a), depressive disorders are approximately 50% more prevalent in women than in men. Simultaneously, lifestyle factors – particularly alcohol and tobacco consumption - pose major public health challenges due to their association with various physical and psychological illnesses. In Germany, these concerns are especially pronounced, with over 2 million people experiencing depressive symptoms in 2024, placing the country 16th in global depression prevalence among 180 nations (World Population Review, 2023).
Alcohol consumption plays a particularly significant role in this context. Globally, approximately one-third (32.5%) of the population has a history of alcohol use (Qi et al., 2024). Germany ranks fifth worldwide in alcohol consumption (Statista, 2025). Numerous studies have demonstrated a direct link between alcohol consumption and depressive symptoms (Qi et al., 2024; Burlaka et al., 2024). One study highlights the importance of not only the quantity but also the frequency of alcohol consumption in determining its impact on depression. The findings suggest that consumption frequency is a key factor influencing the relationship between alcohol use and depressive symptoms (Qi et al., 2024). Interestingly, Qi et al. (2024) also found that individuals who consume alcohol in small quantities have a lower likelihood of developing depression compared to non-drinkers. However, a recent German study contradicts this, emphasizes that even small amounts of alcohol can result in negative health outcomes (Maisch, 2024).
Similarly, tobacco consumption remains a pressing public health challenge, contributing significantly to the global disease burden. The world is currently facing a tobacco epidemic, with over 8 million deaths annually attributed to smoking (WHO, 2023b). Alarmingly, approximately 1.3 million of these deaths result from passive smoking, underscoring the urgency of addressing this issue (WHO, 2023b). Cigarette smoking is the primary form of tobacco consumption. A systematic review by Fluharty et al. (2016) examined the association between cigarette smoking, depression, and anxiety, revealing mixed results. While some studies found evidence linking smoking status to later depression, others did not. A cross-sectional study by Wu et al. (2023) further explored this relationship among US adults. By adjusting for potential confounders and examining smoking status, smoking volume, and smoking cessation, the researchers found that smoking behavior overall increases the risk of depression. Moreover, higher smoking frequency and volume correlated with an increased risk of depression, while smoking cessation was associated with a reduced risk. While strong evidence supports this association, some research remains inconclusive. For instance, Burlaka et al. (2024) investigated the connection between substance use - including tobacco, alcohol and cannabis - and depressive symptoms but did not find a specific link between tobacco use and depression.
Gender differences further influence the relationship between substance use and mental health outcomes. Burlaka et al. (2024) reported that female college students exhibited significantly higher depressive symptoms compared to their male counterparts. Heavy alcohol consumption appears to have a stronger impact on inducing depressive symptoms in women (Kim et al., 2024). Notably, gender disparities extend beyond alcohol, as smoking-related depression and anxiety also differ between males and females (Fluharty et al., 2016).
Despite the extensive research on alcohol consumption and its effects in Germany, there remains a significant gap in studies specifically investigating the association between alcohol use and depression. Given Germany’s high alcohol consumption rates, this study aims to examine the link between alcohol use and depressive symptoms, thereby contributing to a deeper understanding of alcohol’s impact on mental health within the German population. Additionally, while many studies have explored smoking and its effects, there is still limited research on the relationship between tobacco use and depressive symptoms. This paper will address this gap by analyzing the connection between tobacco consumption and depression, with the goal of informing preventive measures.
Understanding gender-based differences in substance use and depressive symptoms is crucial, as these disparities may influence how alcohol and tobacco consumption impact mental health. Preventive strategies targeting alcohol-related issues should account for these differences (Zhan et al., 2012). Ultimately, this study aims to contribute to global depression prevention efforts by identifying some of its underlying causes, reinforcing the idea that tackling depression is not merely an individual issue but a collective societal responsibility (Kim et al., 2024). Based on prior research, we formally hypothesize the following (see also Fig.1):
HA-1: The less alcohol people consume, the fewer depressive symptoms they experience.
HA-2: The less tobacco people consume, the fewer depressive symptoms they experience.
HA-3: The strength of the association between alcohol/tobacco consumption and depressive symptoms is stronger for females than males.
Fig. 1. Hypothesized model of the relationships between study variables.
The path diagram illustrates the proposed effects of alcohol consumption, smoking behaviour, gender, and their interactions on depression, informed by initial model estimates.
Furthermore, this study investigates the association between alcohol consumption and smoking behavior and the likelihood of experiencing clinically significant depression, while also examining whether gender moderates these relationships. The same conceptual model, as illustrated in Fig. 1, was applied to analyze also these variable relationships. Accordingly, we formally hypothesize the following:
HA-4: Higher alcohol consumption increases the odds of clinically significant depression.
HA-5: Higher tobacco consumption increases the odds of clinically significant depression.
HA-6: The relationship between alcohol/tobacco consumption and clinically significant depression is moderated by gender.
2. Methods
In this research, we conducted the analysis using secondary data from the European Social Survey (ESS) Round 11, released in November 2024. The ESS provides high-quality, cross- national data on social and health-related behaviors, including alcohol consumption, smoking behavior, and depressive symptoms, which are central to this study. The data collection for ESS Round 11 employed a probability sampling method, representing individuals aged 15 and older (European Social Survey, 2023). Probability sampling ensures a non-zero probability of selection and involves random selection, producing unbiased estimates of population parameters, such as the sample mean and standard deviation (Bhattacherjee, 2012). Furthermore, the data underwent comprehensive quality assessment procedures and has received ethical approval, ensuring compliance with data protection and confidentiality standards (European Social Survey, 2023). Therefore, we consider the use of this data both meaningful and appropriate, as it directly aligns with our research objectives of analyzing the influence of alcohol and tobacco consumption on depressive symptoms in Germany, with a focus on gender differences.
2.1. Variables
To investigate the hypotheses in this study, the following variables were examined: depression, alcohol consumption frequency and cigarette smoking behavior. Additionally, covariates such as age, education level, and subjective general health were included in the analysis model. Further details on each variable are provided below.
2.1.1. Depressive symptoms
The CES‐D8 Depression Scale was employed to operationalize depression as the dependent variable. The CES-D8 is an 8-item questionnaire designed to measure the frequency of depressive symptoms over the past week. Respondents rated how often they experienced each symptom during the last week using a Likert-type scale, ranging from 0 (None or almost none of the time) to 3 (All or almost all of the time).
2.1.2. Alcohol consumption
Frequency of alcohol consumption was examined as one of the independent variables in this study. Participants from the ESS survey were asked how often they had a drink containing alcohol in the past 12 months. Responses were measured on a 7-point scale ranging from 1 (Every day) to 7 (Never).
2.1.3. Smoking behaviour
The independent variable cigarette smoking behavior was examined depending on the frequency. Participants responded using a 6-point scale, ranging from 1 (I smoke daily, 10 or more cigarettes) to 6 (I have never smoked).
2.1.4 Confounding variables
The confounder age was measured as a continuous variable, representing the participant’s self-reported age in years. Education level was assessed using the International Standard Classification of Education (ISCED), ranging from ISCED 1 (completed primary education) to ISCED 6 (doctoral degree). Subjective general health was measured using a Likert-type scale from 1 (Very Good) to 5 (Very Bad).
2.2. Statistical analysis
All statistical analyses were performed using the statistical software R, specifically R Studio. Information about the sample and key variables was summarized using descriptive statistics. Two of the eight questions from the depression scale (D23 and D25) were negatively worded and were therefore reverse coded for data analysis. Subsequently, the eight items of the questionnaire were computed into a single scale, and Cronbach’s alpha was calculated to measure internal consistency. The Kruskal-Wallis H test was used to examine the association between alcohol/tabacco consumption and depressive symptoms. This test was chosen because the dependent variable, depression, is ordinal (Likert-type scale), while the independent variables, alcohol consumption and smoking behavior, are categorical (Laerd Statistics, 2014). Additionally, it allows for testing statistical differences between groups. Linear regression model was conducted to test hypotheses HA-1, HA-2 and HA-3. Models 2 and 3 examined the association between alcohol and tobacco consumption and depressive symptoms while adjusting for confounding variables. Models 4 and 5 incorporated interaction terms to assess whether the relationship between alcohol consumption or smoking behavior and depressive symptoms is stronger among females than males. Finally, a logistic regression model was used to test hypotheses HA-4, HA-5, and HA-6. Clinically significant depression was defined using a CES-D-8 cut-off score of 9 out of 24, a threshold considered reliable according to Knaul et al. (2018).
3. Results
3.1. Sample Description
Table 1 presents the total number of participants included in the study (2420), with 50.2% male and 49.8% female, indicating an almost equal gender distribution. The mean age of participants is 50.4 years, with a standard deviation of 19.0. Additionally, Table 1 includes a description of key variables categorized into three groups: education level, income, smoking behavior, and alcohol consumption frequency. Missing responses are reported for each variable group.
| Variable | Category | n (Total=2420) | %(Total=100) |
|---|---|---|---|
| Gender | Male | 1214 | 50.2 |
| Female | 1206 | 49.8 | |
| Age (Mean ± SD) | 50.4±19 | ||
| Education Level | Less than Secondary Education | 42 | 1.7 |
| Upper Secondary Education | 1755 | 72.5 | |
| Tertiary Education and Beyond | 609 | 25.2 | |
| NA | 14 | 0.6 | |
| Income group | Low income | 470 | 19.4 |
| Middle income | 878 | 36.3 | |
| High income | 821 | 33.9 | |
| NA | 251 | 10.4 | |
| Smoking behavior | Current smokers | 558 | 23.1 |
| Ex-smokers | 869 | 35.9 | |
| Never smoked | 991 | 41 | |
| NA | 2 | 0.1 | |
| Alcohol Consumption | Frequent drinkers | 1124 | 46.4 |
| Occasional drinkers | 555 | 22.9 | |
| Rare/Never drinkers | 736 | 30.4 | |
| NA | 5 | 0.2 | |
| Data Source: European Social Survey (ESS), Round 11 |
3.2. Data Analysis
The research followed a two-stage analysis, first assessing the reliability of the dependent variable and then evaluating the hypotheses. Reliability was tested using Cronbach’s alpha, yielding a result of 0.794, indicating acceptable consistency and a reliable scale.
3.3. Kruskal-Wallis test results
The Kruskal-Wallis test was conducted to examine the association between alcohol consumption frequency/smoking behavior and depressive symptoms. This is a reliable method for determining whether there are statistical differences between two groups. The results in Table 2 show very small p-values for both alcohol consumption p < 0.001 and smoking behavior p < 0.001 with respect to depressive symptoms. Additionally, the large H values (53.5 for alcohol consumption and 44.1 for smoking behavior) further confirm that these variables are significantly associated with depression levels.
3.4. Visualization of the associations
| Independent Variable | Chi-Square (H) | df | p-value |
|---|---|---|---|
| Alcohol Consumption Frequency | 53.529 | 6 | 9.18e-10 |
| Smoking Behavior | 44.063 | 5 | 2.25e-08 |
| Data Source: European Social Survey (ESS), Round 11 |
Boxplots 1 and 2 were generated to visualize the association between alcohol consumption/smoking behavior and depressive symptoms. As the Kruskal-Wallis test indicated a statistically significant difference between groups, the boxplots provide a visual representation of the distribution and central tendency of depressive symptoms across alcohol and smoking consumption categories.
Boxplot 1 illustrates that median depression levels are lowest among individuals who consume alcohol several times a week or every day, while the highest median levels are observed among those who never drink or drink less than once a month. Depression level variation is generally wider among non-drinkers and infrequent drinkers. Additionally, outliers are present in all groups but appear more frequently among non-drinkers and occasional drinkers.
Boxplot 2 shows that median depression levels are highest among individuals who smoke daily, especially those who consume 10 or more cigarettes per day. In contrast, the lowest median levels are seen among those who have never smoked or have only smoked a few times. Depression levels also tend to be more varied among current and former smokers. Outliers are present across all groups, but are particularly frequent among those who used to smoke or have never smoked.
3.5. HA-1, HA-2 and HA-3 testing results
A linear regression model was conducted with depressive symptoms as the outcome variable, as presented in Table 3. The following reference groups were defined: daily alcohol consumption, never smoked, female gender, very good subjective health, and low educational level. In the baseline model (Model 1), potential confounders, age, gender, educational level and subjective health, were included as covariates in order to statistically control for their confounding influence on the relationship between the main variables and depressive symptoms. In the following models (models 2-5), the main effects as well as the interaction effects of alcohol and tobacco consumption with gender were also taken into account, while the confounders remained controlled.
| term | statistic | Model 1: Baseline | Model 2: + Alcohol | Model 3: + Tabacco | Model 4: Alcohol*Gender | Model 5: Tabacco*Gender |
|---|---|---|---|---|---|---|
| (Intercept) | estimate | 1.741*** | 1.684*** | 1.638*** | 1.722*** | 1.637*** |
| std.error | (0.036) | (0.057) | (0.057) | (0.093) | (0.058) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| agea_num | estimate | -0.003*** | -0.003*** | -0.003*** | -0.003*** | -0.002*** |
| std.error | (0.000) | (0.000) | (0.001) | (0.001) | (0.001) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| gndrMale | estimate | -0.091*** | -0.077*** | -0.081*** | -0.185* | -0.079** |
| std.error | (0.017) | (0.018) | (0.018) | (0.091) | (0.027) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (0.042) | (0.004) | |
| edu_groupedMedium | estimate | -0.106*** | -0.076* | -0.077* | -0.074* | -0.077* |
| std.error | (0.031) | (0.032) | (0.031) | (0.031) | (0.031) | |
| p.value | (<0.001) | (0.016) | (0.014) | (0.019) | (0.014) | |
| edu_groupedHigh | estimate | -0.143*** | -0.108** | -0.096** | -0.092** | -0.096** |
| std.error | (0.034) | (0.034) | (0.034) | (0.034) | (0.034) | |
| p.value | (<0.001) | (0.002) | (0.005) | (0.008) | (0.005) | |
| healthGood | estimate | 0.169*** | 0.171*** | 0.165*** | 0.166*** | 0.165*** |
| std.error | (0.024) | (0.024) | (0.024) | (0.024) | (0.024) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| healthFair | estimate | 0.368*** | 0.366*** | 0.353*** | 0.354*** | 0.353*** |
| std.error | (0.027) | (0.027) | (0.027) | (0.027) | (0.027) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| healthBad | estimate | 0.732*** | 0.721*** | 0.700*** | 0.700*** | 0.700*** |
| std.error | (0.036) | (0.036) | (0.036) | (0.037) | (0.036) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| healthVery bad | estimate | 0.904*** | 0.880*** | 0.860*** | 0.856*** | 0.860*** |
| std.error | (0.085) | (0.085) | (0.085) | (0.085) | (0.085) | |
| p.value | (<0.001) | (<0.001) | (<0.001) | (<0.001) | (<0.001) | |
| alcfreqSeveral times a week | estimate | -0.024 | -0.021 | -0.106 | -0.022 | |
| std.error | (0.040) | (0.040) | (0.089) | (0.040) | ||
| p.value | (0.554) | (0.590) | (0.233) | (0.589) | ||
| alcfreqOnce a week | estimate | -0.017 | -0.011 | -0.105 | -0.012 | |
| std.error | (0.041) | (0.041) | (0.087) | (0.041) | ||
| p.value | (0.682) | (0.777) | (0.231) | (0.775) | ||
| alcfreq2-3 times a month | estimate | 0.002 | 0.016 | -0.082 | 0.016 | |
| std.error | (0.043) | (0.043) | (0.088) | (0.043) | ||
| p.value | (0.964) | (0.717) | (0.353) | (0.720) | ||
| alcfreqOnce a month | estimate | -0.003 | 0.009 | -0.112 | 0.009 | |
| std.error | (0.046) | (0.046) | (0.091) | (0.046) | ||
| p.value | (0.956) | (0.852) | (0.219) | (0.852) | ||
| alcfreqLess than once a month | estimate | 0.052 | 0.064 | 0.008 | 0.064 | |
| std.error | (0.043) | (0.043) | (0.088) | (0.043) | ||
| p.value | (0.224) | (0.133) | (0.931) | (0.134) | ||
| alcfreqNever | estimate | 0.100* | 0.112** | 0.023 | 0.112** | |
| std.error | (0.043) | (0.043) | (0.088) | (0.043) | ||
| p.value | (0.020) | (0.009) | (0.794) | (0.009) | ||
| smok_groupedCurrent smokers | estimate | 0.099*** | 0.099*** | 0.101** | ||
| std.error | (0.023) | (0.023) | (0.032) | |||
| p.value | (<0.001) | (<0.001) | (0.002) | |||
| smok_groupedEx-smokers | estimate | 0.024 | 0.024 | 0.026 | ||
| std.error | (0.020) | (0.020) | (0.028) | |||
| p.value | (0.233) | (0.227) | (0.358) | |||
| alcfreqSeveral times a week × gndrMale | estimate | 0.101 | ||||
| std.error | (0.099) | |||||
| p.value | (0.307) | |||||
| alcfreqOnce a week × gndrMale | estimate | 0.117 | ||||
| std.error | (0.099) | |||||
| p.value | (0.234) | |||||
| alcfreq2-3 times a month × gndrMale | estimate | 0.131 | ||||
| std.error | (0.102) | |||||
| p.value | (0.198) | |||||
| alcfreqOnce a month × gndrMale | estimate | 0.185+ | ||||
| std.error | (0.108) | |||||
| p.value | (0.086) | |||||
| alcfreqLess than once a month × gndrMale | estimate | 0.027 | ||||
| std.error | (0.102) | |||||
| p.value | (0.788) | |||||
| alcfreqNever × gndrMale | estimate | 0.114 | ||||
| std.error | (0.101) | |||||
| p.value | (0.261) | |||||
| smok_groupedCurrent smokers × gndrMale | estimate | -0.003 | ||||
| std.error | (0.045) | |||||
| p.value | (0.938) | |||||
| smok_groupedEx-smokers × gndrMale | estimate | -0.004 | ||||
| std.error | (0.039) | |||||
| p.value | (0.921) | |||||
| R2 | 0.209 | 0.217 | 0.224 | 0.226 | 0.224 | |
| R2 Adj. | 0.206 | 0.213 | 0.218 | 0.218 | 0.218 | |
| Source: European Social Survey Round 11 (ESS11). |
Confounders: The analysis of the baseline model (Model 1) shows that the subjective level of health in particular is a strong predictor of depressive symptoms. People with “good” health report significantly higher depression scores compared to those with “very good” health (est. = 0.169,p=<.001), with this correlation increasing further as health becomes less good. Age also has a significant negative correlation with depression symptoms (est. = -0.003, p = <.001), which indicates a tendency for depressive symptoms to decrease as age increases. In addition, men report significantly lower levels of depression symptoms compared to women (est. = -0.091, p = <.001), which could indicate gender-specific differences in depressive symptoms. In contrast, there were no significant effects for the level of education; it does not appear to play a significant role in the severity of depressive symptoms in this sample.
H1: The analysis shows no linear relationship between less frequent alcohol consumption and lower depressive symptoms. People who never consume alcohol report significantly more depressive symptoms compared to the reference group of daily consumers (est. = 0.1, p = 0.0198, Model 2; est. = 0.023, p = 0.794, Model 4). All other consumption categories do not differ significantly from the reference group. The results therefore speak against H1.
H2: There is a significant positive correlation between current smoking behavior and depressive symptoms (est. = 0.099, p = <.001, Model 3), i.e. current smokers report more symptoms than people who have never smoked. There is no significant difference between former smokers and pepole who never smoked (est. = 0.024, p = 0.233, Model 3). This partly confirms H2.
H3: To test for gender-specific differences, interaction effects with the characteristic “male” were included in Model 4 (alcohol × gender) and Model 5 (tobacco × gender). The results show that none of these interaction effects are statistically significant (all p > 0.05). A weak tendency can only be seen for the category “drink once a month × male” (est. = 0.185, p = 0.0859). Overall, there is no evidence for a significantly stronger association between alcohol or tobacco consumption and depressive symptoms in women compared to men. This does not support H3.
The variance of the models is between R² = 0.209 (Model 1) and R² = 0.226 (Model 4). A relevant proportion of the variance (approx. 20.9%) is therefore already explained by the confounders in the baseline model. The adjusted R² shows small differences between the models, with values between 0.206 (Model 1) and 0.218 (Model 5), which suggests that the additional predictors and interaction effects provide only limited additional explanatory value for depressive symptoms.
3.5. HA-4, HA-5 and HA-6 testing results
The first crude model included the main predictors alcohol consumption frequency and smoking behavior and was adjusted only for age. This model tested whether alcohol and tobacco consumption increase the odds of clinically significant depression while controlling for age. The results showed that smoking behavior was significantly associated with a higher likelihood of depression (OR = 2.38, 95% CI: 1.89 – 3.00) while higher alcohol consumption frequency was associated with a lower likelihood of depressive symptoms (OR = 0.49, 95% CI: 0.39 – 0.62).
The second model examined whether gender moderates the relationship between alcohol and tobacco consumption and clinically significant depression. The interaction terms—alcohol × gender (OR = 0.99, 95% CI: (95% CI: 0.63 – 1.56) and tobacco × gender (OR = 0.80, 95% CI: 0.51 – 1.27) were close to 1 and not statistically significant, indicating no strong evidence of gender-based moderation effects.
The third, fully adjusted model included the main predictors, their interactions with gender, and potential confounders: age, education level, and subjective health. Smoking remained a significant predictor (OR = 2.26, 95% CI: 1.59 – 3.22), while alcohol consumption retained a negative association (OR = 0.56, 95% CI: 0.40 – 0.79). Notably, poorer subjective health was also strongly associated with higher odds of clinically significant depression (OR = 0.23, 95% CI: 0.18 – 0.29), where higher values represent better self-rated health).
The pseudo R² values for the models showed a progressive increase in explained variance. McFadden’s R² values were 0.039, 0.043, and 0.118 for the crude, interaction, and fully adjusted models, respectively. Corresponding Nagelkerke’s R² values were 0.109, 0.121, and 0.334. for the crude, interaction, and fully adjusted models, respectively. Corresponding Nagelkerke’s R² values were 0.069, 0.101, and 0.290, indicating improved model fit with additional variables and interactions.
Based on the statistical analysis, HA-4 is not supported by the evidence and therefore cannot be accepted. While a relationship between alcohol consumption and clinically significant depression was observed, it was in the opposite direction, higher alcohol consumption was associated with lower odds of clinically significant depression, not higher as hypothesized. In contrast, the findings support HA-5, as tobacco use was significantly associated with a higher likelihood of clinically significant depression. Regarding HA-6, there is no strong evidence from the current analysis to suggest that gender moderates the relationship between alcohol or tobacco use and depression, thus, HA-6 is also not supported.
3.6. Comparison of the results from linear and logistic regression
The results of the linear and logistic regression show a high degree of consistency, although the outcome variables are different. While the linear regression examines the severity of depressive symptoms as a continuous variable, the logistic regression focuses on the presence of clinical depression (dichotomous variable). In both models, smoking and alcohol consumption show significant effects: smoking increases and alcohol consumption decreases the probability and severity of depressive symptoms. Gender shows a significant influence in both models, with female gender tending to be associated with higher depression scores or a higher probability of clinical depression. Overall, the findings confirm that the underlying predictors are robustly associated with depressive symptoms, regardless of whether these are recorded as a continuous expression or as a clinical diagnosis.
4. Discussion
This study aimed to examine the association between alcohol and tobacco consumption and depressive symptoms as well as odds of clinically significant depression among the German population, as well as potential gender differences in this relationship. Additionally, the study included confounding variables, namely age, education level, and subjective general health. The findings aim to contribute to the existing body of knowledge and inform preventive measures to address the rising mental health challenges.
As reported in the results section, evidence was found supporting the HA-2 and HA-5, aligning with the studies discussed in the literature section. Wu et al. (2023) provided evidence supporting an association between smoking behavior and an overall increased risk of depression. Additionally, a longitudinal analysis by Park et al. (2024) clearly identified smoking as a significant factor in the development of depressive symptoms. Moreover, the results of our analysis showed that higher tobacco consumption increases the odds of clinically significant depression. This aligns with the findings from (Epstein et al., 2008), who specifically found that daily smokers are approximately 3 times more likely to develop clinically depressive symptoms than non-smokers. Nevertheless, some studies do not provide clear evidence of a straightforward relationship between smoking and depressive symptoms (Burlaka et al., 2024; Du et al. 2022).
Although this research did not support hypotheses HA-1 and HA-4, most of the literature shows contradicting evidence. Qi et al. (2024) identified a relationship between alcohol consumption frequency and depressive symptoms, while Li et al. (2022) highlighted excessive alcohol consumption as a risk factor for mental health. However, studies by Liang et al. (2021) and Qi et al. (2024) showed similar results to our findings, stating that low to moderate alcohol consumption was significantly associated with a lower risk of depressive symptoms compared to never drinkers in both men and women. To the best of our knowledge, limited research exists on the association between alcohol consumption frequency and the odds of clinically significant depression, indicating a high need for further explorations of the topic.
As shown in the results section, HA-3 and HA-6 could not be supported, meaning that this research does not support the hypotheses that the strength of the association between alcohol/tobacco consumption and depressive symptoms is stronger in females than in males as well as that the relationship between alcohol/tobacco consumption and clinically significant depression is moderated by gender. These findings contradict some studies mentioned in the literature review. For instance, Burlaka et al. (2024) and Kim et al. (2024) both found a gender difference in depressive symptoms, with females generally reporting significantly higher levels of depressive symptoms compared to males. Additionally, a cross-sectional study by Yue et al. (2015) found that the association between alcohol consumption and smoking was stronger in females than in males. However, since this study was conducted in a developing country with a different target group, the differences in the results may explain the variation when compared to Germany. Nevertheless, given the inconsistencies in the literature, further research is needed to explore the factors that may influence the relationship between alcohol consumption and gender differences in depressive symptoms.
Interestingly, our analysis revealed an additional observation regarding gender differences in depressive symptoms, indicating that females are more susceptible to depression than males. Previous research has provided evidence of gender differences in depressive symptoms, with studies showing that women tend to have higher mean scores on depression scales (Guo et al., 2024; Su et al., 2024).
In our research, we found a negative association between age and both smoking and alcohol consumption patterns, as well as depression, potentially influencing the relationships among these variables.Although the association between age and depressive symptoms was statistically significant in the linear regression model (p < 0.001), the effect size was very small (est. = –0.003), suggesting that age is not strongly associated with the severity of depressive symptoms in a practically meaningful way. This aligns with Gao et al. (2023), who found that chronological age is not significantly associated with depressive symptoms. However, education level in our sample did not have a strong influence on depressive symptoms, contradicting the findings of Zhao et al. (2024). Subjective general health showed a strong association with depressive symptoms, with poorer health linked to higher depression scores. While our findings suggest that poor subjective general health correlates with increased depression, it is also possible that depressive symptoms contribute to a more negative perception of health, even in the absence of physical health conditions (Ito et al., 2024).
5.1. Limitations
Our study has several limitations. We rejected hypotheses H0-2 and H0-5 based on contrary evidence however, we cannot definitively accept them, as future studies may yield different results (Bhattacherjee, 2022). Aditionally, there is a potential risk of confounding variables that were not included in the analysis, which could have influenced the results. Finally, while we found secondary data to be highly suitable for this research, data collected in a more systematic or scientific manner would have been even more purposeful for this study.
5.2. Quality criteria
It was crucial for this statistical study to ensure high-quality research on the associations between alcohol and tobacco consumption and depressive symptoms. This study took into account quality standards including reliability, validity, and objectivity, while also acknowledging its limitations. Reliability refers to how replicable the study would be to other researchers (Krefting et al., 2013). Statistical analyses were conducted with the use of R, a widely used software for statistical analysis, ensuring reproducibility of the results. Furthermore, the study made use of secondary data, making it easier to replicate. However, a potential reliability issue exists in the form of measurement bias, since the data were self-reported through a survey. When using the same dataset, reliability would be very high, while using other data collection methods could lead to discrepancies in the results.
The validity of the study consists of two forms: internal and external validity (Krefting et al., 2013). The internal validity of a study determines whether it actually measures what it intends to measure (Krefting et al., 2013). In this study, confounding variables were accounted for such as age, education level, and subjective general health. However, some other confounding variables that might have a played a role were not accounted for, which could have influenced the results. The use of appropriate statistical analysis methods further strengthened the internal validity of this study. External validity concerns the generalizability of the results, meaning how the studies’ results can be generalized to other contexts (Krefting et al., 2013). The study sample was merely focused on the German population, enhancing its application in this demographic setting. However, the study findings are not fully generalizable to different populations, for instance under other socioeconomic or cultural circumstances.
Objectivity refers mainly to how personal biases were removed and how information was gathered in the study (Krefting et al., 2013). The study upheld objectivity through the systematic dataset that was used as well as the statistical analyses that were conducted. Findings were presented in a transparent manner, by discussing supporting as well as contradicting literature. Furthermore, by mentioning the limitations, the study remains a neutral stance to further research. Overall, this study meets essential quality criteria including reliability, validity, and objectivity. However, future research should address other potential confounders to examine gender differences in different contexts and use primary data collection methods to strengthen the quality of the study.
References
Bhattacherjee, A. (2022). Social Science Research: Principles, Methods, and Practices. In Digital Commons. University of South Florida. https://digitalcommons.usf.edu/cgi/viewcontent.cgi?article=1002&context=oa_textbooks
Burlaka, J., Johnson, R. M., Marsack-Topolewski, C. N., Hughesdon, K., Owczarzak, J., Serdiuk, O., Bogdanov, R., & Burlaka, V. (2024). Association between Current Substance Use, Healthy Behaviors, and Depression among Ukrainian College Students. International Journal of Environmental Research and Public Health, 21(5), 586. https://doi.org/10.3390/ijerph21050586
Du, X., Wu, R., Kang, L., Zhao, L., & Li, C. (2022). Tobacco smoking and depressive symptoms in Chinese middle-aged and older adults: Handling missing values in panel data with multiple imputation. Frontiers in Public Health, 10. https://doi.org/10.3389/fpubh.2022.913636
Epstein, J. F., Induni, M., & Wilson, T. (2008). Patterns of Clinically Significant Symptoms of Depression Among Heavy Users of Alcohol and Cigarettes. Preventing Chronic Disease, 6(1), A09. https://pmc.ncbi.nlm.nih.gov/articles/PMC2644609/
European Social Survey. (2023). European Social Survey Round 11 Sampling Guidelines: Principles and Implementation. https://www.europeansocialsurvey.org/sites/default/files/2023-06/Round-11-ESS- sampling-guidelines.pdf
Fluharty, M., Taylor, A. E., Grabski, M., & Munafò, M. R. (2016). The Association of Cigarette Smoking With Depression and Anxiety: A Systematic Review. Nicotine & Tobacco Research, 19(1), 3–13. https://doi.org/10.1093/ntr/ntw140
Gao, S., Deng, H., Wen, S., & Wang, Y. (2023). Effects of accelerated biological age on depressive symptoms in a causal reasoning framework. Journal of Affective Disorders, 339, 732–741. https://doi-org.mci.idm.oclc.org/10.1016/j.jad.2023.07.019
Guo, S., Chu, C.-B., & Zheng, X.-Y. (2024). Changes in gender disparities of depressive symptoms among middle-aged and older adults in China: an age-period-cohort analysis. Social Psychiatry and Psychiatric Epidemiology: The International Journal for Research in Social and Genetic Epidemiology and Mental Health Services, 1–14. https://doi- org.mci.idm.oclc.org/10.1007/s00127-024-02747-6
Ito, N. T., Oliveira, D., Rodrigues, F. M. S., Castro-Costa, E., Lima-Costa, M. F., & Ferri, C. P. (2024). Depressive symptoms and self-rated health among Brazilian older adults: Baseline data from the ELSI-Brazil study. Brazilian Journal of Psychiatry, 46. https://doi- org.mci.idm.oclc.org/10.47626/1516-4446-2023-3331
Kim, Y., Kim, J., Oh, J. W., & Lee, S. (2024). Association between drinking behaviors, sleep duration, and depressive symptoms. Scientific Reports, 14(1), 5992. https://doi.org/10.1038/s41598-024-56625-x
Knaul, F. M., Farmer, P. E., Krakauer, E. L., De Lima, L., Bhadelia, A., Kwete, X. J., Arreola-Ornelas, H., Gómez-Dantés, O., Rodriguez, N. M., Alleyne, G. A. O., Connor, S. R., Hunter, D. J., & Rajagopal, M. R. (2018). Alleviating the access abyss in palliative care and pain relief—An imperative of universal health coverage: The Lancet Commission report. Journal of Global Health, 7(1), 010302. https://doi.org/10.1007/s41999-017-0016-0
Krefting, L., Kuper, A., Lingard, L., & Levinson, W. (2013). Nursing Research: Generating and Assessing Evidence for Nursing Practice. Academic Medicine, 88(4), 687–689. https://www.hopkinsmedicine.org/-/media/institute-excellence- education/documents/quality_criteria_in_research.pdf
Laerd Statistics. (2014). Kruskal-Wallis H Test in SPSS Statistics | Procedure, output and interpretation of the output using a relevant example. Laerd.com. https://statistics.laerd.com/spss-tutorials/kruskal-wallis-h-test-using-spss-statistics.ph
Liang, L., Hua, R., Tang, S., Li, C., & Xie, W. (2021). Low-to-Moderate Alcohol Intake Associated with Lower Risk of Incidental Depressive Symptoms: A Pooled Analysis of Three Intercontinental Cohort Studies. Journal of Affective Disorders, 286, 49–57. https://doi.org/10.1016/j.jad.2021.02.050
Li, Y., Zhang, C., Ding, S., Li, J., Li, L., Kang, Y., Dong, X., Wan, Z., Luo, Y., Cheng, A. S., Xie, J., & Duan, Y. (2022). Physical activity, smoking, alcohol consumption and depressive symptoms among young, early mature and late mature people: A cross- sectional study of 76,223 in China. Journal of Affective Disorders, 299, 60–66. https://doi-org.mci.idm.oclc.org/10.1016/j.jad.2021.11.054
Maisch, B. (2024). Alcohol consumption—None is better than a little. Herz, 49(6), 409 419. https://doi.org/10.1007/s00059-024-05280-z
Park, S. K., Oh, C.-M., Kim, E., Ryoo, J.-H., & Jung, J. Y. (2024). The longitudinal analysis for the association between smoking and the risk of depressive symptoms. BMC Psychiatry, 24(1). https://doi-org.mci.idm.oclc.org/10.1186/s12888-024-05828-7
Qi, P., Huang, M., & Zhu, H. (2024). Association between alcohol drinking frequency and depression among adults in the United States: A cross-sectional study. BMC Psychiatry, 24(1), 836. https://doi.org/10.1186/s12888-024-06296-9
Statista. (2025, January 3). Per capita alcohol consumption worldwide 2020, by country. Statista. https://www.statista.com/forecasts/1148811/per-capita-alcohol-consumption-by-country
Su, Z., Yang, X., Hou, J., Liu, S., Wang, Y., & Chen, Z. (2024). Gender differences in the co- occurrence of anxiety and depressive symptoms among early adolescents: A network approach. Journal of Psychiatric Research, 179, 300–305. https://doi- org.mci.idm.oclc.org/10.1016/j.jpsychires.2024.09.024
World Health Organization. (2023a, March 31). Depressive disorder (depression). World Health Organization; World Health Organization. https://www.who.int/news-room/fact- sheets/detail/depression
World Health Organization. (2023b, July 31). Tobacco. World Health Organization. https://www.who.int/news-room/fact-sheets/detail/tobacco
World Population Review. (2023). Depression Rates By Country 2023. World Population Review. https://worldpopulationreview.com/country-rankings/depression-rates-by- country
Wu, Z., Yue, Q., Zhao, Z., Wen, J., Tang, L., Zhong, Z., Yang, J., Yuan, Y., & Zhang, X. (2023). A cross-sectional study of smoking and depression among US adults: NHANES (2005– 2018). Frontiers in Public Health, 11, 1081706. https://doi.org/10.3389/fpubh.2023.1081706
Yue, Y., Hong, L., Guo, L., Gao, X., Deng, J., Huang, J., Huang, G., & Lu, C. (2015). Gender differences in the association between cigarette smoking, alcohol consumption and depressive symptoms: a cross-sectional study among Chinese adolescents. Scientific Reports, 17959. https://doi-org.mci.idm.oclc.org/10.1038/srep17959
Zhan, W., V. Shaboltas, A., & V. Skochilov, R. (2012). Gender Differences in the Relationship between Alcohol Use and Depressive Symptoms in St. Petersburg, Russia. Journal of Addiction Research & Therapy, 03(02). https://doi.org/10.4172/2155-6105.1000124
Zhao, R., Wang, J., Lou, J., Liu, M., Deng, J., Huang, D., & Fang, H. (2024). The effect of education level on depressive symptoms in Chinese older adults–parallel mediating effects of economic security level and subjective memory ability. BMC Geriatrics, 24(1). https://doi-org.mci.idm.oclc.org/10.1186/s12877-024-05233-5