8210 wk 10 discc

Cherie McElroy-Burch 

RE: Discussion - Week 10 

COLLAPSE

Top of Form

Wagner  (2020) explains that a dummy variable is a dichotomy normally coded as a value of “1” to designate the existence of a specific attribute and are helpful when including nominal-level variables in research that require interval- or ratio-level variables. Dummy variables are used in regression analysis and represent the various subgroups within the sample study. The Regression model is a prediction model that uses one or more independent variables to predict the values of a dependent variable (Frankfort-Nachmias et al., 2020). The multiple regression model examines how the value of a dependent variable is impacted by multiple independent variables (Frankfort-Nachmias et al., 2020).

General Social Survey Dataset’s Mean of Age

The General Social Survey monitors and studies the trends in behavior, attitudes, and practices of United States residents.(NORC, n.d.). After receiving funding from The National Science Foundation in 1972 the General Social Survey was launched. Data collected includes information such as respondent’s age, citizenship, sexual orientation and more. The average age of participants in the survey is 46.02. Respondents ages range from 18-89. A mean age of 46.02 signifies that the participants are considered middle age, thus the results from this survey may not be generalizable to the greater population. 

Research Question

Using The General Social Survey dataset, a regression test was conducted and a research question was constructed to answer:

To what extent does marital status affect socioeconomic status? 

The null hypothesis is: “There is no relationship between socioeconomic status and marital status?”

The alternate hypothesis is: “There is a relationship between socioeconomic status and marital status?”

Interpretation of the Coefficients of the Model

Table 1 displays the coefficients of respondents’ marital status and socioeconomic status. The reference category in Table 1 is ‘Married’. Each uncategorized coefficient in Table 1 represents a value in reference to ‘Married’. There is a difference in means between Married and Widowed of - 9.663. The difference between means for Divorced and Married is – 6.683 units. The difference between means for Separated and Married is – 13.079, while the difference between means for Never Married and Married is – 11.597. The level of significance is p is less than .001, which is below the standard alpha level of .05 commonly used by researchers in social science (Wagner, 2020). Since statistical significance has been found the null hypothesis can be rejected. It is concluded that respondents’ marital status is a significant predictor of respondents’ socioeconomic status” (Walden University, 2016). Used in multiple regression testing, the Variance Inflation Factor (VIF) measures the amount of multicollinearity in a set of variables (Walden University, 2016). Values calculated as 10 or above indicate serious multi-collinearity in the model. The calculated VIF for this dataset are 1.083 (Widowed), 1.134 (Divorced), 1.036 (Separated), and 1.157 (Never Married). Each of these are well below 10 and indicated there is no correlation between the variables. It can also be assumed that this model meets the assumption that marital status is a predictor of socioeconomic status

Table 1

Coefficients

  Interpretation of the Diagnostics for the Regression Model

A multiple regression model analyzes the impact multiple independent variables have on the values of a dependent variable (Frankfort-Nachmias et al., 2020). Table 2 displays the Cook's Distance values. The Cook’s Distance statistic provides information about undue influence and certain outliers of one or more of the chosen variables that may be causing undue influence on the model and may have a significant impact (Walden University, 2016). The Cook’s Distance calculated for this model displays a range from .000 to .012. This range suggests no undue influence because it is well below 1.0. The general rule for Cooks Distance is anything calculated at 1.0 or greater is considered problematic (Walden University, 2016).

Table 2

Residual Statistics

Analysis and Summary of the Regression Model

The Durbin-Watson statistic provides information on independence of errors and has values ranging from zero to four (Walden University, 2022).  The calculated Durbin-Watson for this dataset is 1.783 which signifies no correlation between the residuals. The R Square statistic measures the proportion of variance in the dependent variable that is explained by two or more independent variables (Frankfort-Nachmias et al., 2020).  The R Square is .054 which suggests a minuscule and weak effect size which is not meaningful. Higher effect sizes indicates the research finding has practical significance, while a weak effect size indicates limited practical applications (Warner, 2012).  Possible implications for social change for this dataset start with research to determine if lower socioeconomic status has any correlation to individuals deciding not to marry or to separate or divorce. Programs could be organized to assist with marital and financial counseling.

References

 Frankfort-Nachmias, C., Leon-Guerrero, A., & Davis, G. (2020). Social statistics for a diverse society (9th ed.). Sage Publications.

NORC at The University of Chicago. (n.d.). Get GSS Data | NORC. Https://Gss.Norc.Org. Retrieved June 12, 2022, from https://gss.norc.org/get-the-data

Wagner, III, W. E., (2020). Using IBM® SPSS® statistics for research methods and social science statistics (7th ed.). Thousand Oaks, CA: Sage Publications.

Walden University, LLC. (Producer). (2016). Regression diagnostics and model evaluation [Video file]. Author.

 

Bottom of Form