Does the present misuse of alcohol have a major influence on the impacts of growing up with an alcoholic or living with an alcoholic? For our empirical project, our group would like to examine the hypothesis that having an alcoholic parent or growing up in an alcoholic environment influences alcohol abuse as an adult. To avoid prejudice when analyzing the data, we will include additional explanatory factors such as the state of the workforce, age, years of education, geographic area, as well as ethnicity affecting alcohol misuse. We will show this data through multivariate regression models to present our findings. Our group seeks to determine whether there is a significant correlation between childhood exposure to alcoholism and present alcohol abuse through the use of regression models and relevant literature study. Literature: Wooldridge, J.M. (2016). "Introductory Econometrics: A Modern Approach", 6th edition. Data Set Handbook. Terza, J.V. (2002). "Alcohol Abuse and Employment: A Second Look". Journal of Applied Econometrics, 17, 393-404. Retrieved from http://qed.econ.queensu.ca/jae/ Perkins WH, Berkowitz AD. (1991). "Collegiate COAS and alcohol abuse: Problem drinking in relation to assessments of parent and grandparent alcoholism". Journal of Counseling & Development, 69, 237-240. Data source We used one dataset from the 1988 Alcohol Supplement of the National Health Interview Survey, by Mullahy and Sindelar (1996) (M&S), with the population of our sample comprising of the civilian, non-institutionalized population of the United States and the District of Columbia. The variables we are using are abuse which measures if there is current abuse of alcohol where 1 = is abuse of alcohol and 0 = no abuse of alcohol. The variable liveale measures if the respondent lived with an alcoholic, 1 = lived with an alcoholic 0 = did not live with an alcoholic. Age is measured by the respondent's age. Status measures if the respondents are in the work force, 1 = out of the work force, 2 = unemployed, and 3 = employed. Race uses variable white, where 1 = white and 0 nonwhite. The variable educ measures the number of years of education the respondents have completed. An issue we found in our dataset is that the information sourced is heavily processed, posing a challenge. The mean age of the respondents is 39.17 years old. The mean amount of education is 13.31 years. The number of people who responded to alcohol abuse was 974. 8848 people do not abuse alcohol. 8379 respondents are Caucasian, and 1443 respondents are not. 684 respondents are out of the work force. 316 respondents are unemployed. 8822 respondents are employed. 1848 of respondents lived with an alcoholic and 7974 did not. Ran a regression on current abuse of alcohol and lived with an alcoholic and there was very little correlation between them. The r squared value is 0.003283. Planned variables (Y and Xs) Dependent Variable (Y) = Current abuse of alcohol today Explanatory Variables (X) = Raised by a mother or father that was an alcoholic, age, status in the workforce, race identity, years of education and geographic location. Our group has determined the above explanatory variables (X) as we found these given circumstances to be the variables in which influence and predicts the results within our dependent variable (Y). Our report will go into depth on the role and impact that each variable possesses on the overall correlation between variables X and variable Y. Methodology We will employ a multiple linear regression model to investigate the idea that having an inebriated parent or growing up in an alcoholic setting affects alcohol misuse as an adult. We can examine the connection between adult alcohol abuse (the dependent variable) and a number of independent factors, such as early alcoholism exposure, the condition of the labour market, age, years of schooling, place of residence, and ethnicity. The linear regression model can be specified as: Alcohol Abuse_i = Bo+ B(Alcoholic Parent_i) + 2(WorkforceStatus_i) + 3(Age_i) + B4(Education_i) + s(Geographic Area_i) + B6(Ethnicity_i) +&_i Where: AlcoholAbuse i is a measure of alcohol abuse for individual i (e.g., alcohol consumption, frequency of alcohol-related problems, etc.) Alcoholic Parent i is a binary variable indicating whether individual i had an alcoholic parent or grew up in an alcoholic environment (1 = Yes, 0 = No) WorkforceStatus_i represents the state of the workforce for individual i (e.g., employed, unemployed, etc.) Age i is the age of individual i Education_i represents the years of education for individual i Geographic Area_i is a categorical variable representing the geographic area for individual i Ethnicity_i is a categorical variable representing the ethnicity of individual i Bo, 1, ..., 6 are the parameters to be estimated _i is the error term for individual i Sensitivity Analysis: Through this analysis, we can assess the reliability of our conclusions. To make sure that our findings are accurate and unaffected by possible problems with the data or the model, we run a number of tests and reviews. Multicollinearity: Check for multicollinearity by comparing the correlations between two or more separate factors. Because of this, it may be challenging to pinpoint the precise impact of each variable on the dependent variable. To determine whether there is multicollinearity, we use the variance inflation factor (VIF). To solve the problem, we might need to combine or eliminate correlated factors if the VIF numbers are high (above 10). Model description: By adding, deleting, or altering independent factors, we can evaluate various iterations of our regression model. This aids us in determining whether or not our findings hold true for various model stipulations. If the findings change considerably when we change the model, we may need to reconsider our choice of variables or investigate additional factors that could influence alcohol abuse. Influential observations and outliers: An outlier is a data point that greatly deviates from the bulk of observations. Data points that strongly influence the hypothesized connection between the independent and dependent factors are known as influential observations. To find possible anomalies and significant observations, we look at leverage values (a measure of each observation's impact) and residuals (the gap between observed and projected values). If any are discovered, we can evaluate how they will affect our findings and determine whether to eliminate them or modify our model appropriately. Heteroskedasticity: Heteroskedasticity refers to the fact that our model's error terms' variation varies depending on the values of the independent variables. This may result in inaccurate standard deviations and ineffective predictions. To find heteroskedasticity, we use the Breusch-Pagan test or the White test. If heteroskedasticity is discovered, we can use robust standard deviations to account for it and boost the accuracy of our predictions. By conducting these sensitivity analyses, we hope to make sure that our results are dependable and consistent, giving a strong basis for our conclusion regarding the link between exposure to alcoholism as a kid and adult alcohol addiction. Another model we hope to use in our analysis will be the logistic regression model. With this model, we hope to examine if the state of the workforce, age, ethnicity, and geographic area all influence a person's alcohol misuse and how strongly correlated they are. Additionally, the current abuse of alcohol will be our dependent variable while at the same time we will be adjusting for other variables such as age, ethnicity, state of the workforce and geographic area to ascertain whether some explanatory variables presented here may yield a higher causal effect on the misuse of alcohol versus others that may not strongly influence the misuse of alcohol in individuals. Further, results stemming from the factors will be categorized in a 0 or 1 range, with 0 causing little to no effect and 1 being a strong indicator of causal effect. Also, with the findings of this model which will be presented later, a better understanding of why really persons become alcoholics will be investigated and further, which factors can be mostly attributed to this cause and which factors do not play such a significant role in the overall cause of persons becoming alcoholics. The results of this model will further provide a baseline for the appropriate recommendations to take when assessing the topic of alcohol misuse in adults.