CO-OCCURRENCE OF DEPRESSIVE SYMPTOMS AND ALCOHOL USE: IMPACT OF SOCIODEMOGRAPHIC FACTORS

Since depression disorders have been found to be the leading cause of disability (Friedrich, 2017), affecting 17.3 million adults in the United States alone (National Institute of Mental Health, 2017). Various scales have been developed to measure depression. The National Survey on Drug Use and Health uses the Major Depression Episode module to measure lifetime depression symptoms, this is the scale the current work aims to validate. The selected model had adequate fit. It was also found to be invariant at the strictest level across gender, race, income, and education level. The validated scale was then used to study the relationship between depression and alcohol use. The comorbidity of depression and alcohol use has been common in our society for many years (Floud et al., 2015). The purpose of the current study is to investigate the role that race, gender, and socioeconomic status have on the comorbidity between depression and alcohol use. There were differences between depression symptoms and alcohol use on the socioeconomic demographics. However, there were no moderation effects. This indicates that sociodemographics do play a role in both depression symptoms and alcohol use, even though they do not change the relationship. The implications of these studies are two-fold, it sets up future research to be done people who may not have depression but screen into a depression scale to determine how these factors effect.

. In the United States, approximately 17.3 million (7.1%) adults have experienced at least one major depressive episode (MDE) in their lifetime (National Institute of Mental Health "Major Depression", 2017). Depression is defined as experiencing five or more depressive symptoms within a two-week period of time, with at least one symptom being either "depressed mood" or "loss of interest or pleasure" (American Psychiatric Association (APA), 2013). There are many ways that researchers can measure depression, one scale used is the National Comorbidity Survey -Replication (NCS-R) depression module (Kessler et al., 2003;Fava et al., 2010;. lifetime. The adapted NCS-R depression module from the NSDUH is one of the primary sources used to present population-based estimates of MDE .
SAMHSA uses these data to update the MDE estimates on a national level annually. It can also be used to shape policies at the national, state, and local levels. The validation of a scale utilized in a national sample is important for population estimates and policy work at the state and federal level.
The purpose of the current paper is to provide support for using the MDE module in the NSDUH as a one-factor latent variable. Since the data collected are from all 50 states, D.C., and Puerto Rico, the sample is consistently diverse in gender, race, and socioeconomic status (SES). Accordingly, a psychometric validation of the module will be done to compare gender, racial groups, and SES.
Depression affects people of any gender, race, or SES . Past research provides support for depression scales being a 1-factor model in various scales (Ryan et al., 2013;. Ryan et al. (2013) found a 1-factor model for the Patient Health Questionnaire 9-item.  investigated fit on various depression and anxiety models and found that fit between a 1-factor and 2-factor depression scale was not significant. Ensuring that a depression module in a nationwide survey records the same information across these different groups is important. Differences in depression are prevalent between males and females. Females are more likely than males to report depressive symptoms (Angst et al., 2002;Kuehner, 2013;Major Depression, 2019). The prevalence of depression was found to be 8.7% among females while in males it was only 5.3% in the NSDUH data from 2017 (Major Depression, 2019). According to Angst et al. (2002), the reasoning behind this difference is very complex. One theory proposed was, females report more symptoms on the depression scale than males do, meaning males could be impaired and report fewer symptoms even with the same level of depression (Angst et al., 2002). Angst et al. (2002) theorized that this could indicate that males have a lower threshold of depression symptoms than females do (Angst et al., 2002). There was no evidence to suggest that depression symptoms would be different between males and females, therefore there should be invariance across the sexes for the NCS-R. Past research Fonseca-Pedrero et al., 2010) suggests that various depression scales are invariant across sexes as well, increasing support that the NCS-R is invariant. Investigating the invariance of the NCS-R between females and males can provide support that the module is measuring depression symptom accurately between sexes.
Past literature  also supports that White people are more likely than other race to report having depressive symptoms.
There are multiple explanations for this, including how symptoms present Escobar & Gorey, 2018; and cultural background Escobar & Gorey, 2018;. Studies have found that the prevalence of depression among Caucasians is 17.9% compared to African Americans whose prevalence is 10.4% . Although the prevalence seems to be higher in the Caucasian sample, the African American sample has been more effected by chronic depression 56% compared to 38.6% of the Caucasian sample . There are various reasons for this difference, one possibility for the different levels of prevalence is the rate at which seeking treatment happens. African Americans tend to seek treatment less often than Caucasian's leading to depression in African American's being undiagnosed or misdiagnosed more frequently . However, aside from these differences other work found that depression symptoms do not vary be race . It is theorized that somatic symptoms are more common in the Black community because it is more culturally acceptable for somatic symptoms to appear than cognitive symptoms . Cultural barriers also impact the way Hispanic people report or seek treatment for depression as well (Escobar & Gorey, 2018). Escobar and Gorey (2018) conducted a meta-analysis investigating the Hispanic cultures impact on depression treatment in the United States. It was found that throughout the eight randomized control trials and one quasi-experimental study, that adapting depression treatment to account for Hispanic culture reduced depression levels over the course of treatment (Escobar & Gorey, 2018). With cultural differences and the possibility of symptoms being reported differently, testing the invariance of the depression module used in the NSDUH could be valuable.
Lastly, the difference in reporting depression symptoms is also found between different SES levels, although less research has been done with the factor. Research has found that people in higher SES report more depression symptoms (Inaba et al. 2005;. People in high SES levels may be more likely to report depression symptoms because they are able to worry about their mental wellbeing beyond physical needs (Inaba et al., 2005). Inaba et al. (2005) found a negative correlation between depressive symptoms and income level, indicating that people with lower incomes reported more depressive symptoms. For people who experience food insecurity, housing security, or physical safety, their mental state may not be at the forefront of their thoughts. Studying invariance between different levels of SES will be done by looking at two factors, income levels and education levels. It is important to account for multiple SES factors if possible . Due to the differences in depression levels and lack of research on this topic, invariance testing would be valuable.
There is a lack of research investigating the factor structure and invariance of the MDE module in the NSDUH. In order to compare depression symptoms across demographics, it is important to ensure consistency of the measurement. The current study will investigate the factor structure of the MDE module within the NSDUH and then test the invariance of the factor structure across various sociodemographics. The MDE module is expected to be a 1-factor structure. Once the structure is validated, multiple invariance tests will be conducted to determine if comparison across groups is possible.

Methods
National Survey on Drug Use and Health (NSDUH) The data for the current study will come from the NSDUH (SAMHSA, 2019), which is used to track specific substance use, mental health problems, treatments, and other health-related information. Data are collected annually using an independent, multistage area probability sample within each of the 50 states and the District of Columbia. People are excluded from the study if they are active military employees.
There are no requirements for having used substances or experienced mental health issues. The NSDUH collects data from people 12 and older, however due to differences in the depression questionnaire, the current study will use data from only participants 18 and older. The survey is administered via an audio computer-assisted self-interviewing system . For a more detailed description of how data was collected  , 1994).

Sex
For the multiple group CFA, gender was asked in the NSDUH but recoded into sex. For the current study we will be using sex with male and female being the categories.

Race
For the multiple group CFA investigated validation across racial groups. Racial groups were categorized by non-Hispanic White, non-Hispanic Black/ African American, non-Hispanic Native American/ Alaskan Native, non-Hispanic Native Hawaiian/ Other Pacific Islander, non-Hispanic Asian, non-Hispanic more than one race, and Hispanic.
For the current study, we recoded the racial categories into non-Hispanic White, non-Hispanic Black/ African American, and Hispanic, and all other because they were the three largest racial/ ethnicity groups and the rest were below 5% so they were grouped together.

Education Level
To investigate the validation of the MDE module across education levels, the categories included, (1) less than high school, (2) high school graduate, (3) some college/ associate degree, and (4) college graduate.

Confirmatory Factor Analysis
Since the MDE module used in the NSDUH was adapted from the NCS-R, a factor analysis was performed to investigate the properties of the measure. A confirmatory factor analysis (CFA) was conducted to validate the use of the MDE module as a latent variable. Using the MDE as a latent construct will increase the reliability, variability, and the correlations with other factors, which will make it easier to work with. Other benefits of determining the factor structure include and reducing confounding variance. A 1-factor model was selected due to past literature that supports all items being part of one scale through a random sample of people both with and without depression (Kessler et al., 2003), it is also the most parsimonious factor structure.
When looking at the factor loadings on the 1-factor model, item 1 was the only factor to load below .3, therefore it was dropped. A CFA was then conducted using diagonal weighted least squares (DWLS) estimation on the remaining 8-item 1-factor solution. The DWLS estimation was used since this method accounts for categorical items. After the CFA was conducted, three separate multiple group CFAs (MG-CFA) were implemented using the DWLS estimation. The model fit was evaluated by the chi-square value, comparative fit index (CFI), root mean square error of approximation (RMSEA) with a 90% confidence interval, and the standardized root mean square residual (SRMR) for each group separately to ensure the same factor structure held. In order to have acceptable fit, a CFI of .90 or greater , RMSEA of .06 or less , and a SRMR. In order to test the factorial invariance across racial groups, gender, income level, and education level, a multiple group CFA (MGCFA) was conducted. The test was conducted based on the common four-step approach, which entails testing configural invariance (i.e., testing overall factor structure with no equality constraints), metric invariance (i.e., factor loadings constrained to be equal across groups), scalar invariance (i.e., threshold constrained to be equal across groups), and strict (i.e., residuals constrained to be equal across groups). If these requirements were met than little to no measurement bias exists when comparing across these groups and latent means can be meaningfully compared (Milfont & Fischer, 2010 Since the MDE module is binary (yes/ no responses), diagonally weighted least squares will be used as the estimator instead of maximum likelihood estimation.

Reliability
Reliability was measured using Cronbach alphas estimation on the scale.

CFA
A CFA was conducted on the 8-item 1-factor solution. Normality was tested, each item was found to be normally distributed. Each item was correlated with each other (see

Reliability/ Validity
Alpha coefficient for the overall MDE module was .605.

Discussion
This study aimed to fill a gap in research by examining the factor structure of the MDE module used in the NSDUH, which is administered across the United States. The structure of the MDE module was examined first followed invariance tests between different sociodemographics. This was done to determine if the MDE module was comparable across groups. Specially, gender, race, income level, and education levels were tested. It was found that the factor structure of the MDE module was a 1-factor solution. This indicates that all items were measuring on one construct, depression. One item did need to be dropped in order to increase fit, however the study moved forward with the 8-item model. As expected, all remaining items were related to each other. The item that was removed from the model was general feelings of sadness. The 1-factor model held up when used on a different sample of the NSDUH.
The MDE module used in the NSDUH was found to be a 1-factor model after removing one item from the module. This 1-factor model held consistent when tested by sex, race/ethnicity, income level, and education level. Invariance testing proceeded using the 1-factor, 8-item model.

Abstract
The comorbidity of depression and alcohol use has been common in our society for many years (Floud et al., 2015). The purpose of the current study is to investigate the role that race, gender, and socioeconomic status have on the comorbidity between depression and alcohol use. This study was using a national dataset, the National Survey of Drug Use and Health. The results indicated that sociodemographics did not change the relationship between depression and alcohol use. However, there were differences between depression symptoms and alcohol use on the socioeconomic demographics. This indicates that sociodemographics do play a role in both depression symptoms and alcohol use, even though they do not change the relationship. The implications of this study are two-fold, it sets up future research to be done people who may not have depression but screen into a depression scale to determine how these factors effect depression and alcohol use and it indicates research using national data sets is feasible and important.
Co-occurrence of Depressive Symptoms and Alcohol Use: The Impact of Sociodemographic Factors Depression and alcohol use have commonly co-occurred in our society for many years (Flouds et al., 2015;. The direction of the relationship between depression and alcohol use has not been consistent in research  and the role of sociodemographic variables (e.g., gender, race, and socioeconomic status (SES)) have not been thoroughly investigated. Large national data sets exist that can be leveraged to study this relationship (e.g. National Survey on Drug Use and Health; Treatment Episode Data Set). The purpose of the current study is to investigate how different sociodemographic variables (e.g., gender, race, SES) can change the relationship between depressive symptoms and alcohol use. This will be done using a national dataset that is collected annually and funded by the Substance Abuse and Mental Health Administration (SAMHSA).
The co-occurrence of alcohol use and depression has been a prevalent issue in our society for many years . Past research  has attempted to explain the directionality of the relationship between depression and alcohol use in multiple ways. Many researchers  have found, people with depression utilize alcohol as a method to cope with negative affect and it can create a cyclical cycle with the relationship between depression and alcohol use. Due to the nature of the relationship, selecting which one should be the focal outcome variable has varied based on the sample type. The current study will follow the  study, which used alcohol as the outcome variable since their study focused on people who were not required to be currently diagnosed with depression.
Since co-occurrence of depression and alcohol use prevalent in the United States, it is important to study multiple factors that could change the relationship between the two.
Past research has studied the impact that various sociodemographic factors have on the co-occurrence between depression and alcohol use, including race, gender, and SES. However, most studies have limited the number of factors included in their sample and model.  is one of the few that accounted for multiple factors and found that education level plays a significant role in the relationship between mental health and substance abuse symptoms when considering gender, race, and other health factors, within a community sample. It was found that people who did not have a high school diploma prior to 21 were more likely to have comorbidity of mental health issues, including depression, and substance use issues later in life . While race and gender were included as covariates, they were not the focal point of  results.
When the past research (Salas et al., 2015) has investigated the co-occurrence between depression and alcohol use with race as the primary variable, it was found that White people were more likely to have co-occurrence than Black/ African Americans.  found similar results, African American people were not more likely to have comorbidity disorders than White people. Other studies  have found that ethanol was more likely to be in the system of White people who had committed suicide than any other race. While the co-occurrence does appear more frequently in the White samples, literature finds that the co-occurrence in African American's is linked to worse outcomes (Dagher & Green, 2014). These outcomes include, low educational attainment, poor work outcomes, high unemployment levels, and welfare dependency (Dagher & Green, 2014). This heightens the importance of accounting for multiple sociodemographic factors within one model.
Other studies have investigated the role that gender plays in the relationship between depressive symptoms and substance use . Through the various research methods implemented, the overall findings seem to consistently find that the co-occurrence between depression and alcohol use is stronger in women .  found that depression symptoms are linked to later drinking behavior, after accounting for gender. Multiple studies have also supported that women are at greater risk of having depressive symptoms than men  however, overall men are at greater risk of substance use disorders .
This increases the importance of studying how gender can change the relationship between depression/ depressive symptoms and alcohol use. When investigating the impact of gender as a moderator in a longitudinal study, Moscato et al. (1997) found that depressive symptoms predicted alcohol problems among females, over a three-year period, there was no significance in predicting alcohol problems in males.
When the relationship between depression and substance use has been studied used income levels to measure SES. They used Gini scores to measure the SES of the overall neighborhoods that participants lived in . They found that as neighborhood Gini scores increased (or median income level became higher) alcohol use increased .  investigated the relationship between SES status and substance use. He defined SES status as income and education level of parents. Families with higher household income level were more likely to engage in binge drinking and marijuana use, although these were small findings .
It was also found that people with college educated parents were more likely to engage in binge drinking, use marijuana, and cocaine versus people who had high school educated parents . Dagher and Green (2014) found that young adults with longer periods of unemployment had a higher chance of having substance use issues and/or depressive symptoms. This result indicates that it might not be the SES a person is in but what actions they are taking in life (Dagher & Green, 2014).  found that in adolescents, that lower SES, measured by education level of a parent and income level, was negatively related to cigarette smoking but positively related to alcohol consumption. Adolescents who lived in higher income houses, consumed more alcohol .
Other studies investigated the relationship between drug use and depression within one level of SES.  investigated the relationship between co-occurring drug use and depression symptoms. Using 336 adults from Baltimore City who had co-occurring depressive symptoms and drug use, multiple stressors are related to the co-occurrence . However, for our purpose of looking at this study, it was found that drug-related financial stress was related to increase substance use. Disordered problems within the neighborhood was related to increased drug use.
Other factors were related, however, those did not involve money or neighborhood environment . Research on the co-occurrence between depression and substance use with SES as a predictor or moderating third variable was not very strong within the literature.
While many past studies have investigated the impact of race, gender, and SES not many have studied them together. The few studies which included multiple sociodemographic factors found that taken into account together does affect the relationship. Assari (2018) found that when race and SES were in one model, there was a higher risk of depression among African Americans which is contradictory to Mezuck et al. (2010) finding that depression was more common in White people. Tormohlen et al.

(2019) found that when gender is included in a study the impact of SES is not as strong.
This supports the need to incorporate multiple sociodemographic variables into one study even more.
The current study will investigate how gender, race, and SES can influence the relationship between depressive symptoms and alcohol use for people who screen into taking a depression measure. Using a national dataset, National Survey on Drug Use and Health (NSDUH), we will gain knowledge on how the co-occurrence of depressive symptoms and alcohol use can change across the nation. It also provides insights on how to utilize a national dataset that is collected annually in a way that can predict health behaviors. Not only will this study provide insight on the relationship between depressive symptoms and alcohol use it will expand the abilities of the utilization of national datasets that use federal funds to collect.

Method National Survey on Drug Use and Health (NSDUH)
The data for the current study will come from the NSDUH (SAMHSA, 2019), which is used to track specific substance use, mental health problems, treatments, and other health-related information. Data are collected annually using an independent, multistage area probability sample within each of the 50 states and the District of Columbia. People are excluded from the study if they are active military employees.
There are no requirements for having used substances or experienced mental health issues. The NSDUH collects data from people 12 and older, however due to differences in the depression questionnaire, the current study will use data from only participants 18 and older. The survey is administered via an audio computer-assisted self-interviewing system . For a more detailed description of how data was collected view the "Results from the 2018 National Survey on Drug Use and Mental Health: Summary of National Findings" (SAMHSA, 2019).

Sample
The current study will use a sample of all participants 18 and older from the 2018

Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (American Psychiatric
Association (APA), 2000). In order to measure depression in adults, the NSDUH uses an adapted version of the National Comorbidity Survey -Revised (NCS-R) . Revisions were made to reduce the length. For the current study, we will use the endorsement of major depressive symptoms during the lifetime, all participants answered yes to a series of questions in order to be screened into taking the major depressive episode (MDE) module. If they said no to, "Have you ever in your life had a period of time lasting several days or longer when most of the day you felt sad, empty or depressed?"; "Have you ever had a period of time lasting several days or longer when most of the day you were very discouraged about how things were going in your life?"; or "Have you ever had a period of time lasting several days or longer when you lost interest in most of the things you usually enjoy like work, hobbies, and personal relationships?" they were not screened into taking the depression scale. The scale was investigated in a previous study and found a 1-factor, 8-item structure, that will be used in the current study. There are eight questions asked in the lifetime MDE module. If a participant responded yes to five out of the nine symptoms than they fell into the category of having a MDE in their lifetime. For the current study, we will not classify people as yes or no for MDE in lifetime, instead we will create a sum score (0-8) in order to measure number of depressive symptoms in the lifetime per participant (SAMHSA, 2019).

Alcohol Use
Alcohol use per month is measured by a composite variable of amount of alcohol the participant has consumed in the past 30 days and number of drinks consumed on days drinking. The question asked was "What is your best estimate of the number of days you drank alcohol during the past 30 days?" There were nine options for past 30 days drinking, (1) 1 or 2 days, (2) 3 to 5 days, (3) 6 to 9 days, (4) 10 to 19 days, (5) 20 to 29 days, (6) all 30 days, (7) never used alcohol, (8) did not use alcohol in the past 30 days, and (9) don't know (SAMHSA, 2019). The number of drinks that you drank on the days you drank ranged from 1 to 90. These two were multiplied to get the approximate number of drinks in past 30 days.
We also wanted to measure high-risk drinking. In order to do this, we will be using a binge drinking measure. Participants were asked "During the past 30 days, that is since ___, on how many days did you have [4 or more]/[5 or more] drinks on the same occasion? By "occasion," we mean at the same time or within a couple of hours of each other." Participants could report anywhere from 0 to 30 days.

Sociodemographics
The NSDUH recategorizes gender into sex, (1) male and (2) female (SAMHSA, 2019). Sex will be the measure used in the current study.
The measure for race/ ethnicity consists of (1) non-Hispanic White, (2) non-Hispanic Black/ African American, (3) non-Hispanic Native American/ Alaskan Native, (4) non-Hispanic Native Hawaiian/ Other Pacific Islander, (5) non-Hispanic Asian, (6) non-Hispanic more than one race, and (7) Hispanic (SAMHSA, 2019). The current study found that non-Hispanic White, Hispanic, and Black/ African American were the only racial categories that took up more than 5% of the data, therefore non-Hispanic Native American/ Alaskan Native, non-Hispanic Native Hawaiian/ Other Pacific Islander, non-Hispanic Asian, and non-Hispanic more than one race were all grouped together to be defined as all other races.

Proposed Analysis
In order to determine the effect of the sociodemographic variables on the relationship between depression and alcohol use in the last month, four moderation analyses will be conducted. The Hayes (2013) PROCESS package will be used in order to investigate this relationship. Cohen's technique (1988) for effect sizes will be used to measure the effect for t-tests and eta-squared (ω 2 ) will be used to measure effect size for analysis of variance (ANOVA). Each factor will be used as a focal moderator and the other three factors will be included in the moderation model as covariates. Using the regression model allows for multiple categories to be accounted for within each moderating variable. With regression moderation, the interaction between the IV and the moderating variable will first be investigated, if significant, than probing will be done on the interaction to determine where the change in the relationship between the IV and the DV exists, each significant change will then be reported. Next, we will run four moderation analyses to test if depressive symptoms and high-risk alcohol use is affected by the sociodemographic factors. Again, each moderation analysis will include one focal moderator with the remaining variables as covariates. This method will allow us take into account multiple sociodemographic factors at once.

Descriptive and Correlations
All variables were tested for normality, the number of alcoholic beverages consumed in a month was the only variable not found to be normally distributed (see table 2). There were 35,336 removed for either responding to less than 2 items on the 8item MDE module. There were 46 outliers removed, all outliers were 3 standard deviations above the mean for number of alcoholic beverages consumed in one month.
Once the outliers were removed, the number of alcoholic beverages consumed in a month was normally distributed. The adjusted MDE module with eight items was correlated to both the number of binge drinking days in a month and the number of alcoholic beverages consumed, and gender, race, income levels, and education level (see table 3).
Since number of alcoholic beverages was not correlated to the MDE module and wouldn't be used for the remainder of analyses, the 46 outliers were put back into the dataset, normality was tested again (see Another ANOVA found significant differences in the binge drinking and family income levels, F(3, 4684)= 8.768, p<.001, 2 =.01. An omega squared of .01 indicates that the overall effect was small even though there are significant differences. There were significant differences between families making less than $20,000 (M= 3.041,SD=5.290) and those $50, 999 (M=2.331,SD=4.464), less than $20,000 and more than $75,000 (M= 2.066,SD= 4.339),and $20,999 (M= 2.593,SD= 5.060) and more than $75,000. Those in families making less than $20,000 annually reported more binge drinking episodes than families who made between $50,000 and $74,999 and $75,000.

Discussion
The purpose of the current study was to investigate the relationship between depression symptoms and alcohol use, and how sociodemographics can influence this relationship. The relationship between lifetime depression symptoms and past month alcoholic consumption were not related. This was a surprising finding given past literature supports the relationship between depression and alcohol use . However, the lack of significance between depression symptoms and alcohol consumption could be due to the fact that our participants had to screen into taking the MDE module and it measured lifetime depression. Therefore, they did not currently have to be experiencing depression symptoms during the month that they are reported binge drinking. However, when investigating the relationship between depression symptoms and binge drinking there was a relationship though. Once the sociodemographic variables were included in a moderation analysis, depression symptoms were no longer related to binge drinking episodes in a month. The sociodemographic variables did not moderate this relationship, but they were significantly related to depression symptoms and binge drinking episodes in a month.
To further the analysis, I investigated how binge drinking episodes and depression symptoms differed by gender, income level, and education level within the sample that screened into the depression module. Neither binge drinking nor depression symptoms differed by race. There were differences in all other sociodemographics for both depression symptoms and binge drinking episodes. Males reported more lifetime depression symptoms and more binge drinking episodes in one month than females. This contradicts past research  as it is commonly found that females experience more depressive symptoms than males. However, this could be due to the fact that a higher proportion of males did not screen into taking the depression module, the overall percentage of males in the NSDUH was 46% and the percentage of males that screened in took up 36%. It is important to note that for males who did screen into the depression module, both number of depression symptoms and number of binge drinking episode reported in a month was higher. This could indicate that when males are flagged for depression, they could report more depression symptoms than females and this could impact prevention work. Further investigation would be valuable to determine if this finding is consistent within other samples.
When investigating education levels further, the current findings provide support that the number binge drinking episodes in one month are more common among people with less education. People who screened into the depression module and had less than high school education reported more binge drinking episodes than those with some college and college degrees. People with high school degrees binge drank more in one month than those with some college and college degrees. People with some college binge drank more in one month than those with college degrees. Those with the college degrees binge drank less in a month than all other education levels, reporting binge drinking 2.02 days in one month on average. However, the reverse was found for depression symptoms in a lifetime. Those who screened into the depression scale and had a college degrees reported more lifetime depression symptoms than any other group. College graduates reported on average 2.14 lifetime depression symptoms out of the eight symptoms measured.
A similar pattern was found when investigating binge drinking and depression between income levels. Participants who were part of families that made less than $20,000 annually binge drank more than those who made $50,000-$75,000 and more than $75,000 annually. Reporting 3 binge drinking episodes a month on average. People who were part of families making $20,000 to $49,999 annually binge drank more than those making more than $75,000. People with lower income levels (less than $20,000) reported significantly less lifetime depression symptoms than all other income levels.
Participants who were in families that made between $20,000-$49,999 reported less symptoms than people who were part of families that made more than $75,000 annually.
People who were part of families that made more than $75,000 annually reported more depression symptoms than all other groups, reporting 2.02 depression symptoms.
These findings demonstrate that there are differences in drinking behaviors and reporting depression symptoms across different sociodemographics. However, more research is needed in order to determine what is influencing these different relationships in this population. Continued research could help tailor depression and/or alcohol use treatment to better fit the needs of different demographics and prevention work. One reason drinking levels could be lower among the college graduate sample could be an indication of the work that colleges and universities do to prevent substance use and dependency among their students. This finding could be useful for the public health field and targeting prevention work towards people who did not finish high school or those who didn't attend college. Targeting people in low income is showing to be a continued concern that public health officials should strategize programs and outreach plans to reach.
The findings also indicate that depression symptoms are reported more frequently reported in participants who have graduated college or are in high income brackets. This finding could be due to the fact that these symptoms were reported throughout a person's lifetime and not their current feelings, which indicate that a person with a college degree is more likely to have depression in their lifetime. It would not necessarily mean in the current time they are more likely to experience these symptoms. This finding could be used to remind college students about mental health resources, work on reducing mental illness stigma and encourage self-care.
As with all research, there are limitations. Within this study we used a national survey to conduct the analysis. This makes it difficult to use best practice scales to measure depression symptoms. It also encourages more skip patterns, which increases missing data for analyses. The data is also not longitudinal, so cross-sectional analysis was necessary within this dataset.
That being said, using a national sample has many benefits. It was representative of the United States and we were able to study depression and alcohol use within a nationwide sample of people 18 and older who screened into a depression module.
Moving forward including different mental/ behavioral health variables in the analysis might give us a better image as to what influences the relationship between depression and alcohol use. Incorporating other substances used, age of first use, and past treatment might change the outcomes of the current findings. All these variables are available in the NSDUH and could be valuable in targeting treatment and prevention moving forward.

Conclusion
The relationship between depression and alcohol use seems to be very complex.
Sociodemographics influence both depression symptoms and alcohol use however they do not change the relationship between depression and alcohol use, although they do reduce the relationship between depression and alcohol use when included as moderators.
Since the interactions are not significant though, depression and binge drinking is still related with race, gender, income, and education level are included as covariates.
Findings suggest that depression symptoms are higher among college graduates and people who are part of families making more than $75,000 annually. While binge drinking is higher among people who have not graduated college and make less than $20,000 annually. These findings indicate the treatment and prevention for alcohol use should be targeted towards people who have not graduated college or live in low SES. In the field of public health, these are already populations that are being targeted. It might be useful to increase mental health awareness and reduce stigma in people who have graduated college or make more than $75,000 as they reported more depressive symptoms.
between binge drinking, marijuana use, and depressive symptoms and