Selecting Variables into a Regression Equation - Assignment Example

Selecting Variables into a Regression Equation Selecting Variables into a Regression Equation To develop a regression equation an imaginary study has been considered which is aimed at determining the effect of socio-economic conditions on injuries amongst children. For this study, the dependent variable is the number of injuries amongst children and the independent variables (predictors) (Wang & Jain, 2003) and their assigned values are identified as parenting status (single, married, or divorced), parent education level (primary, college, or university), parent income level (Less than $20,000p.a., Between $20,001p.a. and $50,000p.a., Between $50,001p.a. and $100,000p.a., Above $100,000), density of population in neighborhood (low, medium, or high).

The data comprises of 31 observations for each variable. Using step wise approach to predicting the relationship between the dependent variable and independent variables, the correlations between these variables are determined as follows: Number of InjuriesNumber of Injuries1Parenting Status0.40Parent Education0.27Parent Income Level0.11Density of Population-0.30From the above table, the significant relationships are indicated between number of injuries and parenting status and density of population.

There is a positive relationship between number of injuries and parenting status and there is a negative relationship between number of injuries and density of population. Using MS Excel the following regression output is achieved.SUMMARY OUTPUTRegression StatisticsMultiple R0.24079R Square0.05798Adjusted R Square-0.08695Standard Error3.073421Observations31ANOVA dfSSMSFSignificance FRegression415.115853.7789630.4000630.806766Residual26245.59389.445916Total30260.7097    CoefficientsStandard Errort StatP-valueLower 95%Upper 95%Lower 95.

0%Upper 95.0%Intercept4.5126633.2794521.3760420.180547-2.2283511.25367-2.2283511.25367Parenting Status-0.478220.737032-0.648850.522125-1.993221.036767-1.993221.036767Parent Education0.1830280.7892770.2318930.818437-1.439351.805411-1.439351.805411Parent Income Level-0.10580.619629-0.170740.865748-1.379461.167868-1.379461.167868Density of Population0.737240.7049731.0457710.305292-0.711852.186333-0.711852.186333The regression model is only able to explain 5.8% of the total variations observed in 31 data entries (Kedem & Fokianos, 2002).

The above table indicates coefficients of correlation between dependent variable and independent variables. The following regression equation is achieved:Number of Injuries: 4.51 – 0.48 x Parenting Status + 0.18 x Parent Education - 0.11 x Income Level + 0.74 x Density of PopulationFrom the regression equation, it could be predicted that there is a negative relationship but not significant (p