- Mod 3 - Multiple Reg
- 3.1 Overview
- 3.2 The Model
- 3.3 Assumptions
- 3.4 Modelling LSYPE Data
- 3.5 Model 1: Ordinal Explanatory Variables
- 3.6 Model 2: Dichotomous Explanatory Variables
- 3.7 Model 3: Nominal Variables
- 3.8 Predicting Scores
- 3.9 Model 4: Refining the Model
- 3.10 Comparing Coefficients
- 3.11 Model 5: Interaction Effects 1
- 3.12 Model 6: Interaction Effects 2
- 3.13 Model 7: Value Added Model
- 3.14 Diagnostics and Assumptions
- 3.15 Reporting Results
- Mod 3 - Multiple Reg
Multiple Linear Regression Module Exercise
The following questions are slightly different in style to the ones you encountered in the previous module. We asked you to run a full analysis last time but we are now dealing with far more complex models and a single worked example would be way too big! Instead we have broken the process down into smaller questions which are easier to digest and more pleasant for your statistical palette. Don’t worry; we will still be testing you! You will still need to perform a full multiple linear regression analysis. Note that you will also need to use the skills you learnt in previous modules.
Use the LSYPE 15,000 dataset to work through each question. As before we recommend that you answer them in full sentences with supporting tables or graphs where appropriate as this will help when you come to report your own research. The answers are on the next page.
Note: The variable names as they appear in the SPSS dataset are listed in brackets.
There is a variable in the LSYPE data (singlepar) which indicates whether the pupil lives in a single parent family (value=1) or not (value=0). What percentage of pupils in the sample live in single parent families (singlepar)?
Use Frequencies to answer this question.
Does the percentage of pupils with single parents (singlepar) vary across different ethnic groups (ethnicity) and is the association statistically significant?
Use chi-square for this analysis.
Is living in a single parent family (singlepar) related to educational attainment at age 14 (ks3stand), our outcome variable? Graphically display the relationship.
Use a bar chart.
Is the relationship between age 14 attainment (ks3stand) and single parent family (singlepar) statistically significant if you add singlepar to model 7 as an explanatory variable? What is the importance of the single parent family variable relative to the other explanatory variables in the model?
Run model 7 (Page 3.13) adding ‘single parent family’ (singlepar) as an explanatory variable.
Does adding the single parent family variable improve the predictive power of model 7 by a statistically significant increment?
Calculate R2 change using two blocks for your regression analysis.
Does adding the single parent family (singlepar) variable cause any issues for the assumption of homoscedasticity of variance?
Check the scatterplot of predicted score and standardised residual.
All finished? Perhaps it is time to...