Reference no: EM132255220
Assignment - Mini Project
In this assignment, students are required to conduct linear and logistic regression models in SPSS and interpret them. For each set of models, students will need to create one table, and describe in 1-2 paragraphs. For both the linear and logistic models, students will choose four independent/predictor variables and using modified forward regression (which is to say, adding variables to the model).
Requirements for Mini Project:
- Table with 5 models for linear regression (include coefficients, sig values, r-squared, and standard errors).
- Table with 5 models for logistic regression (include OR, sig values, r-squared, and standard errors).
- Regression formula written with slopes and variables specified (linear regression)
- Regression formula written with slopes and variables specified (logistic regression)
- 1-2 paragraphs describing linear regression (describe coefficients, signs, significance values, r-squared), written clearly and concisely at a graduate level. Included in the interpretation should be a higher level explanation of the models, like which ones fit the best, and what happens when all are included at once.
- 1-2 paragraphs describing logistic regression (describe coefficients, signs, significance values, r-squared), written clearly and concisely at a graduate level.
- Included in the interpretation should be a higher level explanation of the models, like which ones fit the best, and what happens when all are included at once.
- Tables neat and tidy, labeled and referred to appropriately. NOT PASTED AS IMAGES.
Model building will be done with five models for linear and five for logistic. Each should utilize the same variables (with the exception of the logistic DV).
Model building:
Model 1: DV & IV1
Model 2: DV & IV2
Model 3: DV & IV3
Model 4: DV & IV4
Model 5: DV & IV1, IV2, IV3, IV4
This should be done for both linear regression and logistic regression. This means, that ultimately, you will need 5 commands for linear regression and 5 for logistic in your syntax. You will need to manually transfer over pertinent information from the regression output to a table that has space for the five models.
Write 1-2 paragraphs interpreting each table. This should include coefficients that were significant and their significance values in parenthesis. Also include observations about the differences in the coefficients in the bivariate models (models 1-4) and the full model (model 5). You also need to write the regression equation for each full model, in this form: Y' = a + bX1 + bX2 + bX3 + bX4.
A few things to note:
Models 1-4 represent unadjusted bi-variate relationships. It is often the case that variables will be related to a DV in unadjusted models and then NOT related to the DV (denoted by results that are not significant). When this happens (when a variable goes from being significant to being insignificant), we say that part of the original relationship in the adjusted model is accounted for in the full model. This happens a lot when there are relationships between IVs. For example, in our dataset, hardship and ICE_income are capturing a lot of the same stuff (poverty, inequality, and economic segregation in Chicago communities), so it is not unreasonable that including them both in the same model, may result in changes to either coefficient in the full model.
Describe results of tables. One thing to pay special attention to is how coefficients or odds ratios change when they are put together in the full model. The full model includes controls for the other variables (aka as: "fully adjusted"). In your explanations, you do not need to explain in great depth coefficients or odds ratios that are not significant. When variables are not significant in the univariate model, you can say simply, "Vars ___ and ___ were not significant in univariate models with the dependent variable... etc." When you are writing up your results, try to explain what happens when a variable goes from being significant in the univariate model to being insignificant in the fully adjusted model.
Attachment:- Assignment Files.rar