\documentclass[11pt, letterpaper]{article}
\usepackage{latexsym}
\usepackage[pdftex]{graphicx}
\usepackage{amsfonts}
\usepackage{lscape}
\usepackage{pdflscape}
\usepackage{amsmath, amsthm, amssymb}
\usepackage{indentfirst}
\usepackage{booktabs}
\usepackage{url}
\usepackage[top=1in, left=1in, right=1in,
bottom=1in]{geometry}

\begin{document}

\section*{\bf{Online Appendix 1: Interviewing Procedures}}

\bigskip

\noindent The Primary Investigators (PIs) for this study were not directly involved in data collection. Rather, the investigators oversaw sample collection, enumerator training, and other administrative tasks that are normally checked when conducting a survey in the developing world. SEDCO (a reputable Pakistani survey firm) was hired to conduct the survey. While SEDCO was involved in the actual collection of participant data, the PIs were ultimately responsible for the design, report, and oversight of survey procedures.\\

\noindent Initial field data collection was conducted in the four main Pakistani provinces. Data collection in the tribal areas (FATA) was initially postponed due to severe winter weather and the seasonal migration of residents to warmer areas of the country.\\

\noindent Following the winter weather, Professor XXXX traveled to Pakistan to work with SEDCO to finalize the survey and train the survey enumerators. Field supervisors recruited the enumerators approximately two weeks prior to training. Enumerator training focused on human subjects training (informed consent, debrief), how to work through the survey and the varied questions' meaning, and practice explaining each question in Pakistan's other vernacular languages.\\

\noindent The following Guidelines for Training in the Protection of Human Subjects were followed:
\begin{itemize}
\item Interviewers must identify themselves and offer to provide the contact information of the Principal Investigator if subjects have questions that the interviewer cannot resolve on the spot.
\item All subjects must be told that the information they are giving will be treated as confidential, that their names will never be mentioned and they will not be identified in any way in the publications that come out of the research.
\item Subjects can refuse to answer any particular question in the interview and must be told that they can do so at the outset of the interview.
\item Subjects must be informed that they can stop the interview at any time.
\item The identity of subjects and the firms they work for are completely confidential. The researchers will not mention the name of any subject or of any firm to anyone outside the research team, for any reason.
\item If handwritten or computer notes are made during interviews, those notes must not contain the name of a person or firm. Only ID codes and dates should be used on these records or notes. Those records should be kept in a locked cabinet when not being used. Those notes may not be shown to anyone outside of the research team.
\item When data are entered into a computer, those computer records should not contain any names of persons or of firms. ID numbers and dates may be used if needed.
\item All information gathered from research subjects, whether from individuals or firms, is strictly confidential and may not be shared with anyone, or used for any purpose other than the current research, without obtaining the written permission of the Principal Investigator at Georgetown University.
\end{itemize}

\noindent Following initial training, the PIs continued to provide oversight at key stages of the effort. Professors XXXX and YYYY met in-person with the survey firm director prior to the fielding of the survey and had extensive interactions during the various stages of pretesting and revising the instrument. The survey firm director and the PIs continued to communicate via email and phone as the survey was in the field, and the survey firm director immediately reported any potential problems or concerns. Based on significant social and research networks the PIs have developed within Pakistan over a number of years, as well as effective collaborations on previous large-scale surveys with local partners, the PIs were confident that SEDCO abided by all applicable local, international and U.S. standards of human subjects protection.\\

\noindent Survey Administration:\\

\noindent Surveys were administered by the enumerator reading a paper script verbatim in a place of the respondent's choosing. This is standard practice for face-to-face household surveys. During enumerator training those enumerators selected for work in local vernaculars (Punjabi, Sindhi, etc.) practiced translating the survey form from Urdu. For the KPK and FATA, we initially sought to prepare a Pashto script, as this is the language in which most respondents were fluent. However, given the number of non-standard Pashto dialects, the research team concluded (in consultation with the client) that an Urdu script would be best for purposes of standardization and quality control. Previous work conducted by the research team has demonstrated that most respondents in all areas of Pakistan understand Urdu regardless of their preference to speak it or their vernacular language.\\

\noindent The standard practice was to allow the respondent to select the location for the interview. Typically this was the foyer of a home, but some respondents (especially females) invited enumerators into the kitchen or requested using a courtyard or other outdoor space.\\

\noindent No personal identifying information (i.e. names, relationships, addresses, etc.) was recorded on the interview forms. No references to the subject were made on the interview forms if the subject happened to divulge his/her name during the course of the survey. Likewise, during the survey debriefing, subjects were given the opportunity to have their data omitted from inclusion. In cases where this occurred, the enumerator thanked the subject for their time, and the field supervisor destroyed (either by shredding or burning) the data at the field office. To avoid the possibility of creating a lasting link between the subject and their participation, survey forms were not left with the subjects unless they specifically asked for them.\\

\noindent Two sets of informed consent and debrief forms were used. One set was for the Punjab, Balochistan, and Sindh. The second set was for the Pashtun-dominant areas of Khyber Pakhtunkhwa and FATA. The materials for Pashtun-dominant areas included information for a local contact at Peshawar University.\\

\noindent Mixed gender teams travelled to each Primary Sampling Unit (PSU). Female enumerators travelled in groups accompanied by the field supervisor, while male enumerators traveled in teams of two. Female interviewers conducted surveys with female respondents, while male interviewers conducted surveys with male respondents. This procedure was intended to accommodate concerns women may have about safety and respectability.\\

\noindent Each survey team included members fluent in Urdu and the appropriate local vernacular (Punjabi, Sindhi, Baloch, etc.). This procedure, as well as the translation ability of the enumerators, was verified by the PIs during training and site visits.  The enumerators who worked in local vernaculars were well experienced through their work with SEDCO and other survey firms. Note that Pakistan is a well-surveyed country and there exists a large pool of well-educated and highly experienced enumerators used by a wide array of Pakistani firms such as Gallup Pakistan, AC Nielson and SEDCO among others.\\

\noindent Only enumerators fluent in a subject's native tongue administered the survey, unless the subject was fluent in Urdu (as are many native Punjabi, Sindhi, Baloch, and even Pashto speakers). When an enumerator encountered a respondent who did not speak Urdu, the enumerator contacted the field supervisor and an enumerator fluent in the participant's native tongue was sent to conduct the survey.

 \clearpage

\section*{\bf{Online Appendix 2: Knowledge and Education Measures}}

\bigskip

\noindent \textbf{Political Knowledge Quiz}

\bigskip

\noindent Now we're going to ask you some questions about recent political events:\\

\noindent Q500.  Did Imran Khan, the head of the Pakistan Tehreek-e-Insaf, declare, to some extent, his personal financial assets in December 2011?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\item Yes
\item No
\end{enumerate}

\noindent Q502. What was the most recent year in which the government made a peace deal with militants in Swat?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\item 1994
\item 2004
\item 2006
\item 2009
\end{enumerate}

\noindent Q505. What party heads the ruling Coalition in Parliament? PPP, JI, ANP, BNP, JUI-F, PML-N, PML-Q, PTI, or MQM? [Read full name of party if respondent does not know what acronym means]
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\item PPP: Pakistan People's Party
\item JI: Jamaat-e-Islami
\item ANP: Awami National Party
\item BNP: Baloch National Party
\item JUI-F: Jamiat Ulema-e-Islam (Fazlur Rehman faction)
\item PML-N: Pakistan Muslim League-Nawaz Sharif
\item PML-Q: Pakistan Muslim League-Qaid
\item PTI: Pakistan Tehreek-e-Insaf
\item MQM: Muttehida Qaumi Mahaz
\end{enumerate}

\noindent Q510. As you may know the 18th Amendment to the Constitution was passed in 2010. What did this amendment do? Please respond `yes' to all that apply:
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\item Renamed the NWFP to Khyber Pakhtunwa
\item Devolved power to the Provinces
\item Renamed the Northern Areas to Gilgit/Baltistan
\item Require the teaching of math and sciences in grade school
\item Created a new court of appeals for civil affairs
\item Gives to the President's right to dissolve the parliament
\item Reforms policing in Pakistan
\end{enumerate}

\noindent Now we're going to ask you some questions about public figures.\\

\noindent [FOR Q515- Q535 READ RESPONDENTS THE FOLLOWING OPTIONS:]
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\item Asif Ali Zardari
\item Yusaf Rhaza Gilani
\item Iftikar Chaudry
\item Ashraf Parvez Kiyani
\item Shabaz Sharif
\item Syed Qaim Ali Shah
\item Ameer Haider Khan Hoti
\item Aslam Raisani
\item Altaf Hussain
\end{enumerate}

\noindent Q515. Who is the President of Pakistan?\\

\noindent Q520. Who is the Prime Minister of Pakistan?\\

\noindent Q525. Who is the Chief Minister of (INSERT PROVINCE OF RESPONDENT)?\\

\noindent Q530. Who is the Chief Justice of the Supreme Court?\\

\noindent Q535. Who is the Chief of Army Staff?

\bigskip \bigskip

\noindent \textbf{Knowledge of Policies in Endorsement Experiment}

\bigskip

\noindent Now we want to ask about whether you have heard of some specific political debates going on right now.

\bigskip

\noindent Q540. Have you heard about debates over whether to deploy the Army to deal with violence in Karachi?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\item Yes
\item No
\end{enumerate}

\noindent Q550. Have you heard about the Frontier Crimes Regulation (sarhad main kavanin) and plans to revise it?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\item Yes
\item No
\end{enumerate}

\noindent Q560. Have you heard about discussions between the Governments of Pakistan and Afghanistan to use peace jirgas to resolve their disputes for example the location of the border?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\item Yes
\item No
\end{enumerate}

\noindent Q570. Have you heard about debates over ongoing efforts between the Indian and Pakistani governments to resolve their difference through dialogue?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\item Yes
\item No
\end{enumerate}

\clearpage

\noindent \textbf{Education}

\bigskip

\noindent What is the highest level of education that you have completed?
\begin{enumerate}
\setlength{\itemsep}{1pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\setlength{\parskip}{0pt}
\item Less than Primary
\item Primary
\item Middle
\item Matriculate
\item Intermediate (F.A/F.Sc)
\item Graduate (B.A/B.Sc.)
\item Professionals (M.S.C., M.A., Ph.D. or other professional degree)
\end{enumerate}

\bigskip

\clearpage

\section*{\bf{Appendix 1: Figures and Tables}}

\begin{figure}[h]
\begin{center}
\includegraphics[width=\textwidth]{endorsebalance.pdf} \\
\textbf{Appendix Figure 1: Balance of the Endorsement Experiment} \\
\end{center}
\end{figure}

\begin{figure}
\begin{center}
\includegraphics[width=\textwidth]{povbalance.pdf} \\
\textbf{Appendix Figure 2: Balance of the Poverty Experiment} \\
\end{center}
\end{figure}

\begin{figure}
\begin{center}
\includegraphics[width=\textwidth]{violbalance.pdf} \\
\textbf{Appendix Figure 3: Balance of the Violence Experiment} \\
\end{center}
\end{figure}

\begin{figure}
\begin{center}
\includegraphics[width=4in]{knowquiz.pdf} \\
\textbf{Appendix Figure 4: Distribution of Scores on the Knowledge Quiz} \\
\end{center}
\end{figure}

\begin{figure}
\begin{center}
\includegraphics[width=4in]{knowpolicy.pdf} \\
\textbf{Appendix Figure 5: Distribution of Scores on the Policy Knowledge Index} \\
\end{center}
\end{figure}

\clearpage

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{l c c c c}
\multicolumn{5}{p{10.5cm}}{\textbf{Appendix Table 1: Descriptive Statistics for Policy Support by Experimental Condition}}\\[1em]
\hline

& Mean & S.D. & \# Obs & Non-response \\
& & & & rate\\[.25em]

\underline{Violence in Karachi (1-5)} & & & & \\[.25em]
Militancy Treatment & 3.59 & 1.29 & 8899 & 8.57\%\\
Edhi Treatment & 3.77 & 1.20 & 3108 & 5.36\%\\
Control & 3.65 & 1.20 & 2993 & 7.54\%\\[1em]

\underline{Mainstreaming FATA (1-5)} & & & & \\[.25em]
Militancy Treatment & 3.40 & 1.31 & 8339 & 14.32\%\\
Edhi Treatment & 3.47 & 1.27 & 2892 & 11.94\%\\
Control & 3.46 & 1.22 & 2782 & 14.06\%\\[1em]

\underline{Durand Line (1-5)} & & & & \\[.25em]
Militancy Treatment & 3.38 & 1.35 & 8497 & 12.70\%\\
Edhi Treatment & 3.50 & 1.29 & 2927 & 10.87\%\\
Control & 3.43 & 1.27 & 2818 & 12.94\%\\[1em]

\underline{Dialogue with India (1-5)} & & & & \\[.25em]
Militancy Treatment & 3.63 & 1.32 & 8659 & 11.03\%\\
Edhi Treatment & 3.78 & 1.19 & 3026 & 7.86\%\\
Control & 3.66 & 1.26 & 2899 & 10.44\%\\[.25em]
\hline
\multicolumn{5}{p{10.5cm}}{Note: Questions are on a five point scale ranging from 1 = ``Not at all'' to 5 = ``A great deal''}\\

\end{tabular}
\end{center}
\end{table}

\clearpage

\begin{landscape}

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{2.275cm} c c c c c c c c c c}
\multicolumn{11}{l}{\textbf{Appendix Table 2: Support for Militant Groups as Measured by the Endorsement Experiment}}\\[.25em]
\hline
& \multicolumn{2}{c}{\underline{Group Average}} & \multicolumn{2}{c}{\underline{SSP}}
& \multicolumn{2}{c}{\underline{Pakistan Taliban}} & \multicolumn{2}{c}{\underline{Afghan Taliban}}
& \multicolumn{2}{c}{\underline{Edhi}}\\[.5em]
Endorsement & -0.010 & -0.016  & -0.002 & -0.007 & -0.009 & -0.016 & -0.020 & -0.024 & 0.017 & 0.010\\
Condition & (0.014) & (0.014) & (0.018) & (0.017) & (0.018) & (0.018) & (0.019) & (0.018) & (0.017) & (0.016)\\[1em]
Constant & 0.638*** & 0.543***  & 0.638*** & 0.536*** & 0.638*** & 0.537*** & 0.638*** & 0.534*** & 0.638*** & 0.542***\\
& (0.012) & (0.023) & (0.012) & (0.029) & (0.012) & (0.030) & (0.012) & (0.030) & (0.012) & (0.027)\\[1em]
N & 10485 & 10485 & 5286 & 5286 & 5244 & 5244 & 5207 & 5207 & 5376 & 5376\\
R$^2$ & 0.00 & 0.03  & 0.00 & 0.03 & 0.00 & 0.03 & 0.00 & 0.04 & 0.00 & 0.05\\
Dem. Controls & N & Y  & N & Y & N & Y & N & Y & N & Y\\[.15em]
\hline
Endorsement & 0.031 & 0.027  & 0.052 & 0.054 & 0.016 & 0.011 & 0.026 & 0.022 & 0.004 & 0.016\\
Condition & (0.041) & (0.040) & (0.058) & (0.057) & (0.049) & (0.048) & (0.057) & (0.055) & (0.049) & (0.048)\\[1em]
Political & 0.093* & 0.026  & 0.093* & 0.035 & 0.093* & 0.038 & 0.093* & 0.021 & 0.093* & 0.021\\
Knowledge & (0.054) & (0.053) & (0.054) & (0.054) & (0.054) & (0.053) & (0.054) & (0.054) & (0.054) & (0.053)\\[1em]
Endorsement x & -0.068 & -0.071 & -0.089 & -0.100 & -0.042 & -0.044 & -0.075 & -0.074 & 0.018 & -0.009\\
Knowledge & (0.069) & (0.067) & (0.097) & (0.095) & (0.083) & (0.082) & (0.094) & (0.090) & (0.083) & (0.081)\\[1em]
Constant & 0.581*** & 0.523***  & 0.581*** & 0.512*** & 0.581*** & 0.515*** & 0.581*** & 0.517*** & 0.581*** & 0.531***\\
& (0.032) & (0.035) & (0.032) & (0.038) & (0.032) & (0.039) & (0.032) & (0.039) & (0.032) & (0.037)\\[1em]
N & 10485 & 10485 & 5286 & 5286 & 5244 & 5244 & 5207 & 5207 & 5376 & 5376\\
R$^2$ & 0.00 & 0.04 & 0.00 & 0.03 & 0.00 & 0.03 & 0.00 & 0.04 & 0.01 & 0.05\\
Dem. Controls & N & Y  & N & Y & N & Y & N & Y & N & Y\\
\hline
\multicolumn{11}{l}{Notes: OLS regressions predicting policy support. Robust standard errors clustered by PSU in parentheses.}\\
\multicolumn{11}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}

\end{landscape}

\begin{landscape}

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{2.275cm} c c c c c c c c c c}
\multicolumn{11}{l}{\textbf{Appendix Table 2: Support for Militant Groups as Measured by the Endorsement Experiment (cont)}}\\[.25em]
\hline
& \multicolumn{2}{c}{\underline{Group Average}} & \multicolumn{2}{c}{\underline{SSP}}
& \multicolumn{2}{c}{\underline{Pakistan Taliban}} & \multicolumn{2}{c}{\underline{Afghan Taliban}}
& \multicolumn{2}{c}{\underline{Edhi}}\\[.5em]
Endorsement & -0.028 & -0.031 & -0.019 & -0.019 & -0.036 & -0.040 & -0.029 & -0.031 & 0.007 & 0.010\\
Condition & (0.021) & (0.021) & (0.027) & (0.027) & (0.027) & (0.027) & (0.029) & (0.028) & (0.025) & (0.024)\\[1em]
Policy-Specific & 0.136*** & 0.118***  & 0.136*** & 0.122*** & 0.136*** & 0.120*** & 0.136*** & 0.112*** & 0.136*** & 0.116***\\
Knowledge & (0.025) & (0.026) & (0.025) & (0.026) & (0.025) & (0.026) & (0.25) & (0.026) & (0.025) & (0.026)\\[1em]
Endorsement x & 0.023 & 0.021  & 0.016 & 0.011 & 0.035 & 0.035 & 0.014 & 0.011 & 0.003 & -0.007\\
Knowledge & (0.031) & (0.031) & (0.039) & (0.039) & (0.039) & (0.039) & (0.043) & (0.042) & (0.037) & (0.036)\\[1em]
Constant & 0.571*** & 0.538*** & 0.571*** & 0.523*** & 0.571*** & 0.527*** & 0.571*** & 0.526*** & 0.571*** & 0.528***\\
& (0.017) & (0.025) & (0.017) & (0.031) & (0.017) & (0.032) & (0.017) & (0.031) & (0.017) & (0.029)\\[1em]
N & 10485 & 10485  & 5286 & 5286 & 5244 & 5244 & 5207 & 5207 & 5376 & 5376\\
R$^2$ & 0.05 & 0.07  & 0.05 & 0.06 & 0.05 & 0.07 & 0.04 & 0.07 & 0.05 & 0.07\\
Dem. Controls & N & Y  & N & Y & N & Y & N & Y & N & Y\\[.15em]
\hline
Endorsement & -0.009 & -0.012 & 0.006 & 0.005 & -0.010 & -0.013 & -0.021 & -0.024 & 0.015 & 0.014\\
Condition & (0.018) & (0.017) & (0.022) & (0.021) & (0.023) & (0.022) & (0.024) & (0.023) & (0.021) & (0.021)\\[1em]
Level of & 0.105*** & 0.069** & 0.105*** & 0.093*** & 0.105*** & 0.047 & 0.106*** & 0.052* & 0.104*** & 0.039\\
Education & (0.025) & (0.027) & (0.025) & (0.030) & (0.025) & (0.029) & (0.025) & (0.030) & (0.025) & (0.029)\\[1em]
Endorsement x & -0.017 & -0.018 & -0.041 & -0.044 & -0.011 & -0.013 & 0 & 0.001 & -0.004 & -0.012\\
Education & (0.031) & (0.030) & (0.040) & (0.040) & (0.039) & (0.038) & (0.042) & (0.041) & (0.039) & (0.037)\\[1em]
Constant & 0.610*** & 0.539***  & 0.610*** & 0.530*** & 0.610*** & 0.535*** & 0.609*** & 0.534*** & 0.611*** & 0.540***\\
& (0.014) & (0.024) & (0.014) & (0.029) & (0.014) & (0.031) & (0.014) & (0.031) & (0.014) & (0.028)\\[1em]
N & 10485 & 10485 & 5286 & 5286 & 5244 & 5244 & 5207 & 5207 & 5376 & 5376\\
R$^2$ & 0.01 & 0.03 & 0.01 & 0.03 & 0.02 & 0.03 & 0.02 & 0.04 & 0.02 & 0.05\\
Dem. Controls & N & Y  & N & Y & N & Y & N & Y & N & Y\\
\hline
\multicolumn{11}{l}{Notes: OLS regressions predicting policy support. Robust standard errors clustered by PSU in parentheses.}\\
\multicolumn{11}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}

\end{landscape}


\clearpage

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{2.7cm} c c c}
\multicolumn{4}{p{8.25cm}}{\textbf{Appendix Table 3: Effects of Self-Reported Socio-Economic Status on Support for Militant Groups}}\\[.25em]
\hline
& \multicolumn{3}{c}{\underline{Self-Reported SES}}\\[.75em]
Endorsement & -0.011 & -0.015 & --- \\
Condition & (0.017) & (0.016) & \\[.75em]
SES: & -0.014 & -0.006 & -0.029*\\
Bottom 20\% & (0.016) & (0.015) & (0.018)\\[.75em]
SES: & 0.008 & 0.002 & 0.011\\
Top 20\% & (0.016) & (0.016) & (0.019)\\[.75em]
Endorsement x & -0.032* & -0.031* & -0.014\\
Low SES & (0.020) & (0.020) & (0.020)\\[.75em]
Endorsement x & 0.034** & 0.031* & 0.023\\
High SES & (0.019) & (0.019) & (0.018)\\[.75em]
Constant & 0.639*** & 0.578*** & 0.633***\\
& (0.014) & (0.024) & (0.013) \\[.75em]
N & 10485 & 10485 & 10502 \\
R$^2$ & 0.01 & 0.03 & 0.14 \\
Dem. Controls & N & Y & N \\[.75em]
Endorsement x Province FE & N & N & Y \\
Clustering & PSU & PSU & District \\
\hline\\
\multicolumn{4}{p{8.25cm}}{Notes: OLS regressions predicting policy support. One-tailed test conducted for the prediction that relative poverty will decrease support for militancy. Robust standard errors clustered by PSU or District in parentheses.}\\
\multicolumn{4}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}


\clearpage

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{2.5cm} c c c c c c}
\multicolumn{7}{p{13cm}}{\textbf{Appendix Table 4: Effects of Experimental Treatments on Support for Militant Groups}}\\[.25em]
\hline
& \multicolumn{3}{c}{\underline{Poverty Experiment}}
& \multicolumn{3}{c}{\underline{Violence Experiment}}\\[.75em]
Endorsement & 0.015 & 0.006 & --- & 0.025 &  0.024 & ---\\
Condition & (0.020) & (0.020) & & (0.030) & (0.029) & \\[.75em]
Poverty & 0.031* & 0.022 & 0.024* & --- & --- & ---\\
Treatment & (0.024) & (0.023) & (0.018) & & & \\[.75em]
Endorsement x & -0.052** & -0.046** & -0.036** & --- & --- & --- \\
Poverty & (0.029) & (0.028) & (0.021) & & & \\[.75em]
Violence & --- & --- & --- & 0.086*** & 0.091*** & 0.062*** \\
Treatment & & & & (0.032) & (0.031) & (0.025) \\[.75em]
Endorsement x & --- & --- & --- & -0.107*** & -0.112*** & -0.104***\\
Violence & & & & (0.039) & (0.038) & (0.041)\\[.75em]
Constant & 0.623*** & 0.532*** & 0.631*** & 0.597*** & 0.482*** & 0.627***\\
& (0.017) & (0.025) & (0.013) & (0.026) & (0.037) & (0.016) \\[.75em]
N & 10485 & 10485 & 10492 & 5300 & 5300 & 5309 \\
R$^2$ & 0.03 & 0.04 & 0.14 & 0.01 & 0.05 & 0.14 \\
Dem. Controls & N & Y  & N & N & Y & N \\[.75em]
Endorsement x Province FE & N & N & Y & N & N & Y \\
Clustering & PSU & PSU & District & PSU & PSU & District\\
\hline\\
\multicolumn{7}{p{13cm}}{Notes: OLS regressions predicting policy support. One-tailed tests conducted for the following predictions: relative poverty will decrease support for militancy, and higher perceived levels of violence will decrease support for militancy. Robust standard errors clustered by PSU or District in parentheses.}\\
\multicolumn{7}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}

\clearpage

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{3.75cm} c c c}
\multicolumn{4}{p{10cm}}{\textbf{Appendix Table 4: Effects of Experimental Treatments on Support for Militant Groups (cont)}}\\[.25em]
\hline
& \multicolumn{3}{c}{\underline{Poverty x Violence}}\\[.75em]
Endorsement & 0.034 & 0.031 & ---\\
Condition & (0.042) & (0.041) & \\[.75em]
Relatively Poor, & 0.109*** & 0.105*** & 0.075**\\
More Violence (RPMV) & (0.046) & (0.045) & (0.037)\\[.75em]
Relatively Wealthy, & 0.078** & 0.084** & 0.049*\\
More Violence (RWMV) & (0.046) & (0.045) & (0.031)\\[.75em]
Relatively Poor, & 0.017 & 0.008 & 0\\
Less Violence (RPLV) & (0.050) & (0.048) & (0.040)\\[.75em]
Endorsement x & -0.146*** & -0.145*** & -0.123***\\
RPMV &  (0.056) & (0.054) & (0.048)\\[.75em]
Endorsement x & -0.087* & -0.095** & -0.091**\\
RWMV & (0.055) & (0.054) & (0.051)\\[.75em]
Endorsement x & -0.020 & -0.016 & -0.006\\
RPLV & (0.058) & (0.056) & (0.044)\\[.75em]
Constant & 0.590*** & 0.480*** & 0.629***\\
& (0.038) & (0.045) & (0.016)\\[.75em]
N & 5300 & 5300 & 5303\\
R$^2$ & 0.01 & 0.05 & 0.14\\
Dem. Controls & N & Y  & N\\[.75em]
Endorsement x Province FE & N & N & Y\\
Clustering & PSU & PSU & District\\
\hline\\
\multicolumn{4}{p{10cm}}{Notes: OLS regressions predicting policy support. One-tailed tests conducted for the following predictions: relative poverty will decrease support for militancy, and higher perceived levels of violence will decrease support for militancy. Robust standard errors clustered by PSU or District in parentheses.}\\
\multicolumn{4}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}


\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{3.25cm} c c c c}
\multicolumn{5}{p{10.5cm}}{\textbf{Appendix Table 5: Effects of Experimental Treatments by Socio-economic Status on Support for Militant Groups}}\\[.25em]
\hline
& \multicolumn{2}{c}{\underline{Poverty Experiment}} & \multicolumn{2}{c}{\underline{Violence Experiment}}\\
Endorsement & 0.030* & 0.022 & 0.029 & 0.025 \\
Condition & (0.021) & (0.021) & (0.030) & (0.029) \\[1em]
Low SES: & 0.009 & 0.020 & -0.016 & -0.007 \\
Bottom 20\% & (0.021) & (0.020) & (0.032) & (0.032) \\[1em]
Endorsement x & -0.062** & -0.062*** & -0.013 & -0.010 \\
Low SES & (0.027) & (0.026) & (0.038) & (0.037) \\[1em]
Poverty & 0.042** & 0.037* & --- & --- \\
Treatment & (0.025) & (0.024) & & \\[1em]
Endorsement x & -0.062** & -0.058** & --- & --- \\
Poverty & (0.030) & (0.029) & & \\[1em]
Low SES x & -0.052* & -0.055** & --- & --- \\
Poverty & 0.032 & (0.031) & & \\[1em]
Endorsement x & 0.041 & 0.044 & --- & --- \\
Low SES x Poverty & (0.041) & (0.039) & & \\[1em]
Violence & --- & --- & 0.094*** & 0.097*** \\
Treatment & & & (0.033) & (0.032) \\[1em]
Endorsement x & --- & --- & -0.098*** & -0.101*** \\
Violence & & & (0.040) & (0.039) \\[1em]
Low SES x & --- & --- & -0.035 & -0.033 \\
Violence & & & (0.042) & (0.040) \\[1em]
Endorsement x & --- & --- & -0.046 & -0.051 \\
Low SES x Violence & & & (0.057) & (0.054) \\[1em]
Constant & 0.621*** & 0.555*** & 0.601*** & 0.517***\\
& (0.018) & (0.027) & (0.026) & (0.038) \\[1em]
N & 10485 & 10485 & 5300 & 5300 \\
R$^2$ & 0.01 & 0.03 & 0.02 & 0.05 \\
Dem. Controls & N & Y & N & Y \\
\hline
\multicolumn{5}{p{10.5cm}}{Notes: OLS regressions predicting policy support. One-tailed tests conducted for the following predictions: relative poverty will decrease support for militancy, and higher perceived levels of violence will decrease support for militancy. Robust standard errors clustered by PSU in parentheses. Middle and upper classes are pooled in these analyses.}\\
\multicolumn{5}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}

\clearpage

\begin{table}[ht]
\small
\begin{center}
\begin{tabular}{p{2.5cm} c c c c c c}
\multicolumn{7}{p{13cm}}{\textbf{Appendix Table 6: Effects of Poverty Experiment on Support for Militant Groups when Controlling for Potential Confounding Interactions}}\\[.25em]
\hline
& \multicolumn{2}{c}{\underline{Education}}
& \multicolumn{2}{c}{\underline{Literacy}}
& \multicolumn{2}{c}{\underline{TV Ownership}}\\[.75em]
Endorsement & 0.013 & 0.010 & 0.026 & 0.017 & 0.050* & 0.043\\
Condition & (0.022) & (0.022) & (0.023) & (0.023) & (0.028) & (0.027)\\[.75em]
Poverty & 0.027 & 0.022 & 0.028 & 0.022 & 0.033 & 0.024\\
Treatment & (0.023) & (0.023) & (0.023) & (0.023) & (0.024) & (0.023)\\[.75em]
Endorsement x & -0.047* & -0.046* & -0.050* & -0.045* & -0.054* & -0.048* \\
Poverty & (0.028) & (0.028) & (0.028) & (0.028) & (0.029) & (0.028)\\[.75em]
Level of & 0.102*** & 0.065** & --- & --- & --- & ---\\
Education & (0.024) & (0.027) & & & & \\[.75em]
Endorsement x & -0.014 & -0.016 & --- & --- & --- & ---\\
Education & (0.031) & (0.030) & & & & \\[.75em]
Literacy & --- & --- & 0.050*** & -0.006 & --- & --- \\
 & & & (0.015) & (0.017) & & \\[.75em]
Endorsement x & --- & --- & -0.024 & -0.019 & --- & --- \\
Literacy & & & (0.019) & (0.018) & & \\[.75em]
Owns a TV & --- & --- & --- &  --- & 0.047** & 0.051**\\
 & & & & & (0.021) & (0.021)\\[.75em]
Endorsement x & --- & --- & --- & --- & -0.047* & -0.049**\\
Owns a TV & & & & & (0.025) & (0.024)\\[.75em]
Constant & 0.597*** & 0.529*** & 0.598*** & 0.524*** & 0.588*** & 0.504***\\
& (0.018) & (0.026) & (0.019) & (0.026) & (0.024) & (0.028) \\[.75em]
N & 10485 & 10485 & 10485 & 10485 & 10485 & 10485 \\
R$^2$ & 0.01 & 0.04 & 0.01 & 0.04 & 0.00 & 0.04 \\
Dem. Controls & N & Y  & N & Y & N & Y \\[.75em]
\hline\\
\multicolumn{7}{p{13cm}}{Notes: OLS regressions predicting policy support. Two-tailed tests conducted for this robustness check. Robust standard errors clustered by PSU in parentheses.}\\
\multicolumn{7}{l}{*** $p<0.01$, ** $p<0.05$, * $p<0.1$}
\end{tabular}
\end{center}
\end{table}

\end{document}