Testing For Multiple Regression Part 1

Image result for multiple regression analysis"
Variables Entered/Removeda
Model Variables Entered Variables Removed Method
1 Q1. Age, ADULTCT: Number of adults in householdb . Enter
a. Dependent Variable: Q46a. Level of democracy: today
b. All requested variables entered.
Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .025a .001 .001 2.880
a. Predictors: (Constant), Q1. Age, ADULTCT: Number of adults in household
ANOVAa
Model Sum of Squares Df Mean Square F Sig.
1 Regression 235.010 2 117.505 14.166 .000b
Residual 386074.042 46544 8.295    
Total 386309.052 46546      
a. Dependent Variable: Q46a. Level of democracy: today
b. Predictors: (Constant), Q1. Age, ADULTCT: Number of adults in household
Coefficientsa
Model Unstandardized Coefficients Standardized Coefficients t Sig.
B Std. Error Beta
1 (Constant) 5.319 .042   127.063 .000
ADULTCT: Number of adults in household .005 .005 .004 .886 .375
Q1. Age .005 .001 .024 5.249 .000
a. Dependent Variable: Q46a. Level of democracy: today
  • Based on the data given, what is the research question? Answer: Given all applicable data, how can we analyze real assumptions that may subsist between level of democracy today, and age of the respondents, and the number of adults in household?
  • The null hypothesis is that there seem to exist no relationship between respondents’ level of democracy toady, and age of the respondents, and number of adults in household.
  • My research design is a comparative design of methodology. This type of research design is composed of comparing and contrasting correlation or relationships among variables (Cantrell, 2011).
  • My dependent variable is the level of democracy today. It is measured as metric or interval variable to determine whether correlation exists between the variables.
  • My independent variable is the number of adults in the household. This is also measured as interval or ratio variable to determine correlation between the variables. The other variable that was added is the controlled independent variable of the respondents’ age. It is measure as ratio variable. The reason for adding age of the respondents is because of how multiple regression model is also referred to as fit a multiple regression model (Laureate Education, 2016g). Additionally, this type of regression model is an extended form of bivariate regression that tests more than one independent variable (Frankfort-Nachmias & Leon-Guerrero, 2015).  
  • Looking at the table of Model Summary, We can see the R correlation or multiple R of 0.025 as the R constant of the predictors. We can also see that the R Square and Adjusted R Square are the same. It’s always safer to use adjusted r square when you’re using multiple predictors (Wagner, 2016). This tells us the strength of the significant as the percent value of the adjusted R Square of the variability in respondents’ level of democracy today.
  • Our Anova table also shows a significance level of 0.000, which is below the conventional level of 0.05. The implication for positive social change allows us to review how relationship may exist in terms of Africa’s view on their democracy, the age of the democratic participants and the number of adults whom may participate in the democratic process. We can likewise review this implication in term of statistical Anova table. In Anova table, we can confirm the predictors of our significant level as 0.000. Here we see the significant level is 0.000. Therefore, we can say that, given the conventional threshold of 0.05, our model proved statistical significance. This also means that positive implications occur in terms of how adults can turn out to vote for democratic changes in Africa.
  • Looking at the coefficient table, we can explain the unstandardized coefficient while controlling other independent variables (Laureate Education, 2016g). Therefore, for each unit increase in our independent variable, our dependent variable will contrast by the value of the unstandardized coefficient (Laureate Education, 2016g). For example, for every increase in level of democracy today, our number of adults in household will vary by 0.005 units, controlling for numerical age of the respondents.
  • To further interpret our unstandardized coefficient, we would have to look into statistical significant level between the two independent variables. These statistical significant levels vary by moderate variance between the two independent variables. The number of adults in household is showing a P-value of 0.375 while age is showing a P-value of 0.00. To use the conventional 0.05 threshold, we can therefore accept the null hypothesis that there is relationship between the level of democracy today and the number of adults in household. However, we can reject the null hypothesis there is no relationship between level of democracy today and the age of dependents.
  • The Beta standardized coefficients can also be clarified similar to how we labelled the unit comparison of the unstandardized coefficients; the standardized coefficients allows us to say that for every unit increase in standard deviation, the dependent variable will vary by the beta’s value of standard deviations (Laureate Education, 2016g). For example, for every unit increase in standard deviation of the age of respondents, their level of democracy today will vary by the unit of 0.024, controlling for the number of adults in household.

Reference

              Cantrell, M. A. (2011). Demystifying the research process: Understanding a descriptive comparative research design. Pediatric Nursing, 37(4), 188-9.

             Frankfort-Nachmias, C., & Leon-Guerrero, A. (2015). Social statistics for a diverse society (7th ed.). Thousand Oaks, CA: Sage Publications.

            Laureate Education (Producer). (2016g). Multiple regression [Video file]. Baltimore, MD: Author.

            Wagner, W. E. (2016). Using IBM® SPSS® statistics for research methods and social science statistics (6th ed.). Thousand Oaks, CA: Sage Publications