psc 5940: interactions as multi-level models
DESCRIPTION
PSC 5940: Interactions as Multi-Level Models. Session 3 Fall, 2009. Workshop: PRCs. Load EE data Run a simple model: Willingness to pay for an alternative energy tax Use randomly assigned “price” as IV Plot to relationship (use jitter) Now add: Income, Ideology - PowerPoint PPT PresentationTRANSCRIPT
September 8, 2009 Session 3 Slide 1
PSC 5940: Interactions as Multi-Level Models
Session 3
Fall, 2009
September 8, 2009 Session 3 Slide 2
Workshop: PRCs• Load EE data
• Run a simple model:
• Willingness to pay for an alternative energy tax
• Use randomly assigned “price” as IV
• Plot to relationship (use jitter)
• Now add: Income, Ideology
• Change in price variable? (Why?)
September 8, 2009 Session 3 Slide 3
Model Elaboration
• EE09 & NS09 Data: research thinking• Analysis of residuals
• Additions to the ERDF model:
• Belief in anthropogenic climate change
• Recodes?
• Understanding of GCC science
• Recode “What scientists’ believe…” variables
• Turn in 1 page summaries
September 8, 2009 Session 3 Slide 4
Dummy Intercept Variables• Dummy variables allow for tests of the differences
in overall value of the Y for different nominal groups in the data (akin to a difference of means)
• Coding: 0 and 1 values (e.g., men versus women)
Y
X1
X2,0
X2,1
September 8, 2009 Session 3 Slide 5
Modeling Belief in GCC as a function of knowledge and Republican Party Identification
Belief in GCC systematically lower for those who identify as Republicans
Call:lm(formula = gcc_bel ~ R_id + gcc_knowl)
Residuals: Min 1Q Median 3Q Max -16.158 -2.582 1.842 3.842 12.460
Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) -1.4467 0.5529 -2.617 0.00897 ** R_id -3.0135 0.3110 -9.688 < 2e-16 ***gcc_knowl 1.5210 0.1310 11.613 < 2e-16 ***---Residual standard error: 5.526 on 1507 degrees of freedom (188 observations deleted due to missingness)Multiple R-squared: 0.1468, Adjusted R-squared: 0.1457 F-statistic: 129.7 on 2 and 1507 DF, p-value: < 2.2e-16
September 8, 2009 Session 3 Slide 6
Dummy Variable Applications• Implies a comparison (the omitted group)
– Be clear about the “comparison category”• Multinomial Dummies
– When categories exceed 2• Importance of specifying the base category
• Examples of Category Variables– Experimental treatment groups– Race and ethnicity– Region of residence– Type of education– Religious affiliation– “Seasonality”
• Adds to modeling flexibility
September 8, 2009 Session 3 Slide 7
Interaction Effects• Interactions occur when the effect of one X is
dependent on the value of another• Modeling interactions:
– Use Dummy variables (requires categories)– Use multiplicative interaction effect
• Multiply an interval scale times a dummy (also known as a “slope dummy”)
• Example: the effect of GCC knowledge (gcc_knowl) on belief in climate change (gcc_bel) may be affected by whether the respondent identifies with the Republican Party (R_id)– Re-code the interaction; run it.
September 8, 2009 Session 3 Slide 8
Modeling belief in Climate Change with a Dummy Slope Variable: Knowledge*
Republican ID (=1) Call:lm(formula = gcc_bel ~ R_id + gcc_knowl + gcc_kn_R)
Residuals: Min 1Q Median 3Q Max -15.971 -2.610 1.691 4.029 13.826
Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) -0.6831 0.6696 -1.020 0.3079 R_id -5.1433 1.1007 -4.673 3.24e-06 ***gcc_knowl 1.3308 0.1613 8.252 3.37e-16 ***gcc_kn_R 0.5563 0.2758 2.017 0.0439 * ---Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 5.521 on 1506 degrees of freedom (188 observations deleted due to missingness)Multiple R-squared: 0.1491, Adjusted R-squared: 0.1474 F-statistic: 87.99 on 3 and 1506 DF, p-value: < 2.2e-16