Voting to Tell Others Online Appendix

Size: px

Start display at page:

Download "Voting to Tell Others Online Appendix"

Elvin Martin
5 years ago
Views:

1 Voting to Tell Others Online Appendix Stefano DellaVigna UC Berkeley and NBER John A. List UChicagoandNBER Gautam Rao UC Berkeley This version: January 13, 214 Ulrike Malmendier UC Berkeley and NBER 1

2 Online Appendix Figure 1. Number of Times Asked about Voting CDFs of number of times asked about turnout Probability <= value Number of times asked 21 Congr. Elections 28 Presid. Elections Note: Online Appendix Figure 1 plots the cumulative distribution function of the self-reported number of times asked among the respondents to the 211 door-todoor survey. The continuous line refers to the 21 Congressional election, and the dotted line refers to the 28 Presidential election. 1

3 Online Appendix Figures 2a-b. Sensitivity of Estimated Parameters to the Moments: Willingness to Complete a Survey µ survey and σ survey for voters 2 Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive µ_survey_voters σ_survey_voters µ survey and σ survey for non-voters Effect on parameter in units of asymptotic standard deviation PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive µ_survey_nonvoters σ_survey_nonvoters

4 Note: Online Appendix Figures 2a and 2b present the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the mean value of completing a 1-minute survey (in blue) and the standard deviation of this willingness to complete a survey (in orange) respectively for voters (Online Appendix figure 2a) and for non-voters (Online Appendix Figure 2b). 3

5 Online Appendix Figure 3. Sensitivity of Estimated Parameters to the Moments: Value of time Value of Time for voters and non-voters Note: Online Appendix Figure 3 presents the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the value of time for voters (in blue) and for non-voters (in orange). 4 Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive timeval_nonvoters timeval_voters

6 .5 Online Appendix Figure 4. Sensitivity of Estimated Parameters to the Moments: Probability of Being at Home Baseline probability of being home (h) for voters and non-voters Note: Online Appendix Figure 4 presents the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the baseline probability of being at home for voters (in blue) and for non-voters (in orange). 5 Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive h_nonvoters h_voters

7 Online Appendix Figure 5. Sensitivity of Estimated Parameters to the Moments: Probability of Observing the Flyer Probability of seeing the flyer (r) for voters and non-voters Note: Online Appendix Figure 5 presents the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the probability of observing the flyer for voters (in blue) and for non-voters (in orange). 6 Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive r_nonvoters r_voters

8 Online Appendix Figures 6a-b. Sensitivity of Estimated Parameters to the Moments: Sorting Elasticity and Social Pressure Elasticity of sorting (eta) and Social Pressure Cost (S) of refusing survey for voters Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive eta_voters S_svy_voters.4 Elasticity of sorting (eta) and Social Pressure Cost (S) of refusing survey for non-voters PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive Effect on parameter in units of asymptotic standard deviation eta_nonvoters S_svy_nonvoters

9 Note: Online Appendix Figures 6a-b present the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the cost of sorting (in blue) and for the social pressure cost (in orange) first for voters (Online Appendix Figure 6a) and then for non-voters (Online Appendix Figure 6b). 8

10 Online Appendix Figure 7. Sensitivity of Estimated Parameters to the Moments: Lying Cost (Full Estimation) Cost of Lying (L) Note: Online Appendix Figure 7 presents the sensitivity of the estimates to the individual moments following Gentzkow and Shapiro (213). The plotted lines indicate the local sensitivity of the given parameter estimate to an individual moment. A positive bar indicates that a (local) increase in the moment would increase the estimated value of the parameter. Higher bars indicate more influential moments for the identification of the parameter. For each moment, we plot the influence estimate for the lying cost. These estimates are for the full estimation case, which requires that on this subsample voters and non-voters have the same key parameters, allowing for estimation of the lying cost. 9 Effect on parameter in units of asymptotic standard deviation PH_V_NW_d5m PH_V_Nw_1d1m PH_V_Nw_1d5m PH_V_W_d5m PH_V_W_1d1m PH_V_W_1d5m PH_V_We_d5m PH_V_We_1d1m PH_V_We_1d5m PH_V_Oo_d5m PH_V_Oo_1d1m PH_V_Oo_1d5m PH_V_Ooe_d5m PH_V_Ooe_1d1m PH_V_Ooe_1d5m PH_NV_NW_d5m PH_NV_Nw_1d1m PH_NV_Nw_1d5m PH_NV_W_d5m PH_NV_W_1d1m PH_NV_W_1d5m PH_NV_We_d5m PH_NV_We_1d1m PH_NV_We_1d5m PH_NV_Oo_d5m PH_NV_Oo_1d1m PH_NV_Oo_1d5m PH_NV_Ooe_d5m PH_NV_Ooe_1d1m PH_NV_Ooe_1d5m PSV_V_NW_d5m PSV_V_Nw_1d1m PSV_V_Nw_1d5m PSV_V_W_d5m PSV_V_W_1d1m PSV_V_W_1d5m PSV_V_We_d5m PSV_V_We_1d1m PSV_V_We_1d5m PSV_V_Oo_d5m PSV_V_Oo_1d1m PSV_V_Oo_1d5m PSV_V_Ooe_d5m PSV_V_Ooe_1d1m PSV_V_Ooe_1d5m PSV_NV_NW_d5m PSV_NV_Nw_1d1m PSV_NV_Nw_1d5m PSV_NV_W_d5m PSV_NV_W_1d1m PSV_NV_W_1d5m PSV_NV_We_d5m PSV_NV_We_1d1m PSV_NV_We_1d5m PSV_NV_Oo_d5m PSV_NV_Oo_1d1m PSV_NV_Oo_1d5m PSV_NV_Ooe_d5m PSV_NV_Ooe_1d1m PSV_NV_Ooe_1d5m POO_V_Oo_d5m POO_V_Oo_1d1m POO_V_Oo_1d5m POO_V_Ooe_d5m POO_V_Ooe_1d1m POO_V_Ooe_1d5m POO_NV_Oo_d5m POO_NV_Oo_1d1m POO_NV_Oo_1d5m POO_NV_Ooe_d5m POO_NV_Ooe_1d1m POO_NV_Ooe_1d5m PSV_V_Nw_NoInfo PSV_V_Nw_Info PSV_V_W_NoInfo PSV_V_W_Info PSV_V_We_NoInfo PSV_V_We_Info PSV_V_Oo_NoInfo PSV_V_Oo_Info PSV_V_Ooe_NoInfo PSV_V_Ooe_Info PSV_NV_Nw_NoInfo PSV_NV_Nw_Info PSV_NV_W_NoInfo PSV_NV_W_Info PSV_NV_We_NoInfo PSV_NV_We_Info PSV_NV_Oo_NoInfo PSV_NV_Oo_Info PSV_NV_Ooe_NoInfo PSV_NV_Ooe_Info PL_V_5m_Control PL_V_5m_Incentive PL_V_1m_Control PL_V_1m_Incentive PL_NV_5m_Control PL_NV_5m_Incentive PL_NV_1m_Control PL_NV_1m_Incentive

11 Online Appendix Table 1. Survey Treatments, Robustness Specification: Dependent Variable: Group: $1/1min Treatment $1/5min Treatment Simple Flyer Treatments Flyer Treatments with Opt-out Mention of Election in Flyer Voters Informed at Door of Election Topic Omitted Treatment Fixed Effects for Solicitor, Date- Location, and Hour (Benchmark) Fixed Effects for Solicitor-Date- Location, and Hour R2 N OLS Regressions Indicator for Answering the Door Indicator for Completing Survey Voters Non-Voters Voters Non-Voters (1) (2) (3) (4) (5) (6) (7) (8).364**.314* ***.266*** (.15) (.16) (.15) (.16) (.1) (.11) (.9) (.9).596***.518*** ***.638***.467***.47*** (.17) (.18) (.15) (.17) (.13) (.14) (.9) (.1) ***.948***.496***.51*** (.18) (.2) (.18) (.19) (.13) (.14) (.1) (.11) ***.731***.325***.349*** (.19) (.21) (.18) (.19) (.13) (.14) (.1) (.11) ** -.274* -.194* -.238** -.238*** -.216** (.13) (.14) (.14) (.15) (.11) (.12) (.8) (.9) (.9) (.1) (.8) (.8) No Flyer, $/5min Treatment No Flyer, $/5min, Not Informed Treatment X X X X X X X X ,873 6,873 6,324 6,324 6,873 6,873 6,324 6,324 Notes: Estimates for a linear probability model with standard errors, clustered by solicitor-date, in parentheses. The omitted treatment is the Baseline No-Flyer $-5 minutes survey. The regressions include fixed effects for the solicitor, for the date-town combination, and for the hour of day in Columns 1,3, 5, 7. The regressions include in addition fixed effects for solicitor-date-town location in Columns 2, 4, 6, 8. * significant at 1%; ** significant at 5%; *** significant at 1% \ 1

12 Online Appendix Table 2. Survey Treatments, By Time Period Specification: Dependent Variable: Indicator for Answering the Door OLS Regressions Indicator for Completing Survey Indicator for Lie in Turnout Question Time Period: Summer Fall Summer Fall Summer Fall Summer Fall Summer Fall Summer Fall Group: Voters Non-Voters Voters Non-Voters Voters Non-Voters (1) (2) (3) (4) (5) (6) (7) (8) (9) (1) (11) (12) $1/1min Treatment.51*** **.271** *** (.18) (.26) (.2) (.23) (.14) (.16) (.12) (.12) $1/5min Treatment.69***.543* **.654***.7***.432***.534*** (.2) (.3) (.22) (.21) (.16) (.21) (.13) (.13) Simple Flyer Treatments ***.953***.928***.268*.815*** (.24) (.29) (.25) (.25) (.18) (.21) (.14) (.14) Flyer Treatments with Opt-out ***.545***.28.57*** (.24) (.31) (.24) (.26) (.17) (.21) (.15) (.15) Mention of Election in Flyer ** ** -.27* -.273** (.18) (.19) (.18) (.21) (.15) (.16) (.11) (.12) Voters Informed at Door of Election Topic (.12) (.13) (.11) (.11) Treatment with Incentive to Say that *** Did not Vote (.23) (.35) (.54) (.59) Omitted Treatment No Flyer, $/5min Treatment No Flyer, $/5min, Not Informed Treatment No Incentive to Lie Solicitor, Date-Location, Hour F.e. X X X X X X X X Date-Location F.e. X X X X R N 4,245 2,628 3,459 2,865 4,245 2,628 3,459 2, Notes: Estimates for a linear probability model with standard errors, clustered by solicitor-date, in parentheses. The regressions include fixed effects for the solicitor, for the date-town combination, and for the hour of day in Columns 1-8 and fixed effects for date-location in Columns * significant at 1%; ** significant at 5%; *** significant at 1% 11

13 Online Appendix Table 3. Survey Treatments, By Political Registration Specification: Dependent Variable: Indicator for Answering the Door OLS Regressions Indicator for Completing Survey Lie in Turnout Question Political Registration: Republican Democratic Other Republican Democratic Other RepublicanDemocratic Other (1) (2) (3) (4) (5) (6) (7) (8) (9) Panel A. Voters $1/1min Treatment.64** * (.25) (.23) (.25) (.18) (.17) (.17) $1/5min Treatment.544*.55**.887***.827***.612***.677*** (.28) (.25) (.29) (.21) (.19) (.21) Simple Flyer Treatments ** *** 265***.758*** (.33) (.25) (.32) (.23) (.21) (.24) Flyer Treatments with Opt-out -.769** *.687***.81***.564** (.35) (.25) (.35) (.24) (.21) (.23) Mention of Election in Flyer * -.344* (.25) (.2) (.25) (.21) (.17) (.19) Voters Informed at Door of Election * Topic (.17) (.14) (.17) Treatment with Incentive to Say that Did not Vote Omitted Treatment Solicitor, Date-Location, Hour F.e. Date-Location F.e. R2 N Panel B. Non-Voters $1/1min Treatment $1/5min Treatment Simple Flyer Treatments Flyer Treatments with Opt-out Mention of Election in Flyer Voters Informed at Door of Election Topic Treatment with Incentive to Say that Did not Vote (.43) (.27) (.54) No Flyer, $/5min Treatment No Flyer, $/5min, Not Informed Treatment X X X X X X X X X ,918 3,18 1,937 1,918 3,18 1, ** ** (.61) (.33) (.18) (.48) (.21) (.9) *.44*** (.72) (.39) (.17) (.51) (.26) (.1) *** (.93) (.47) (.2) (.65) (.27) (.11) ** (.91) (.44) (.19) (.65) (.28) (.11) *** ** (.69) (.34) (.15) (.43) (.22) (.9) (.43) (.19) (.9) * -.97** (.265) (12) (.45) Omitted Treatment No Flyer, $/5min Treatment No Flyer, $/5min, Not Informed Treatment Solicitor, Date-Location, Hour F.e. X X X X X X X X X Date-Location F.e. X X X R N 351 1,179 4, ,179 4, Notes: Estimates for a linear probability model with standard errors, clustered by solicitor-date, in parentheses. The regressions include fixed effects for the solicitor, for the date-town combination, and for the hour of day in Columns 1-8 and fixed effects for date-location in Columns * significant at 1%; ** significant at 5%; *** significant at 1%

14 Online Appendix Table 4. Moments and Estimates on Erat and Gneezy (212) Decision Number: Payoffs of A (Truth) Payoffs of B (Lie) Fraction Lying (Empirical) Fraction Lying (At Estimated Parameters) (2, 2) (19, 3) 33/11 (33%) 39% (2, 2) (21, 3) 49/11 (49%) 43% (2, 2) (3, 3) 66/12 (65%) 62% (2, 2) (21, 15) 38/14 (37%) 34% (2, 2) (3, 2) 57/19 (52%) 56% Parameter Estimates: Lying Cost Altruism Coefficient S.D. of error term 7. (1.4)***.29 (7)* 18.6 (4.)*** Notes: Estimates from minimum-distance estimator using the 5 moments shows above and weights given by the inverse of the variance of each moment. * significant at 1%; ** significant at 5%; *** significant at 1% 13

Online Appendix for The Importance of Being. Marginal: Gender Differences in Generosity

Online Appendix for The Importance of Being Marginal: Gender Differences in Generosity Stefano DellaVigna, John List, Ulrike Malmendier, Gautam Rao January 14, 2013 This appendix describes the structural