adding additional references

paper.bib

@@ -14,7 +14,7 @@ @article{murray2018
 
 @article{vanbuuren2011,
   title={mice: Multivariate imputation by chained equations in R},
-  author={van Buuren, Stef and Groothuis-Oudshoorn, Karin},
+  author={{van Buuren}, Stef and Groothuis-Oudshoorn, Karin},
   journal={Journal of statistical software},
   volume={45},
   pages={1--67},
@@ -68,4 +68,38 @@ @report{amado2022
     month={Aug}
 }
 
+@article{price2015,
+  title={The limits of observation for understanding mass violence},
+  author={Price, Megan and Ball, Patrick},
+  journal={Canadian Journal of Law and Society/La Revue Canadienne Droit et Soci{\'e}t{\'e}},
+  volume={30},
+  number={2},
+  pages={237--257},
+  year={2015},
+  publisher={Cambridge University Press},
+  doi={10.1017/cls.2015.24}
+}
+
+@article{price2014,
+  title={Big data, selection bias, and the statistical patterns of mortality in conflict},
+  author={Price, Megan and Ball, Patrick},
+  journal={The SAIS Review of International Affairs},
+  volume={34},
+  number={1},
+  pages={9--20},
+  year={2014},
+  publisher={JSTOR},
+  url={https://www.jstor.org/stable/27000935}
+}
 
+@article{gargiulo2022,
+  title={Statistical Biases, Measurement Challenges, and Recommendations for Studying Patterns of Femicide in Conflict},
+  author={Gargiulo, Maria},
+  journal={Peace Review},
+  volume={34},
+  number={2},
+  pages={163--176},
+  year={2022},
+  publisher={Taylor \& Francis},
+  doi={10.1080/10402659.2022.2049002}
+}

paper.md

@@ -43,9 +43,9 @@ The data compiled by the joint JEP-CEV-HRDAG project are publicly available from
 
 Collecting data on human rights abuses in conflict settings is a difficult and often dangerous task. Organizations collecting such data may have constrained resources, lack physical access to areas where violence is occurring, or may be unable to collect data due to security concerns or community mistrust, among other challenges. As a result, the data produced by these data collection efforts is seldom a complete enumeration of all violence that occurred nor a statistically representative sample. Furthermore, instances of violence that are documented may be missing key information about victims, perpetrators, or contextual details about the violent events. The incompleteness of the data is not a critique of the data itself nor the organizations that courageously document human rights violations, but rather it is an empirical reality that must be addressed in quantitative work analyzing conflict-related violence.
 
-The data analyzed in the joint JEP-CEV-HRDAG project was no exception to this empirical reality and was subject to two types of missing data: missing fields and underreporting. Related to missing fields, some records were missing socio-demographic information about victims such as their age, sex, or ethnicity, information identifying armed groups thought to be responsible for the violence, or precise information about the date and location of a particular violent event. These gaps in the data pose challenges for analyses seeking to stratify the data based on any fields containing missing values. With respect to underreporting, some instances of violence were not documented by any of the databases we received, leaving some victims' stories untold. Moreover, this missingness is unlikely to be randomly distributed among members of the victim population, meaning that inferences drawn from samples of documented victims alone could result in erroneous conclusions about patterns of violence.
+The data analyzed in the joint JEP-CEV-HRDAG project was no exception to this empirical reality and was subject to two types of missing data: missing fields and underreporting. Related to missing fields, some records were missing socio-demographic information about victims such as their age, sex, or ethnicity, information identifying armed groups thought to be responsible for the violence, or precise information about the date and location of a particular violent event. These gaps in the data pose challenges for analyses seeking to stratify the data based on any fields containing missing values. With respect to underreporting, some instances of violence were not documented by any of the databases we received, leaving some victims' stories untold (see estimates of underreporting in @amado2022, as well as related discussions in other contexts @price2015, ). Moreover, this missingness is unlikely to be randomly distributed among members of the victim population, meaning that inferences drawn from samples of documented victims alone could result in erroneous conclusions about patterns of violence.
 
-The joint JEP-CEV-HRDAG project employed two statistical methods to address the two types of missingness. To address missing fields within records of documented victims, the project used the `R` package `mice` [@vanbuuren2011] to perform multiple imputation [e.g., @murray2018], probabilistically filling in missing values at the record level multiple times. Multiple systems estimation [e.g., @bird2018; @chao2001], performed on the imputed replicate files, was then used to estimate the number of missing observations, that is, the number of the victims never documented by any of the data sources used in the project.[^mse] To do this, we used a Bayesian latent class multiple-capture model [@manriquevallier2016] implemented in the `R` package `LCMCR`. The analyses presented in the technical appendix of the joint project combine these two methods to examine patterns of enforced disappearance, homicide, kidnapping, and forced recruitment of minors in the armed conflict.[^displacement]
+The joint JEP-CEV-HRDAG project employed two statistical methods to address the two types of missingness. To address missing fields within records of documented victims, the project used the `R` package `mice` [@vanbuuren2011] to perform multiple imputation [e.g., @murray2018], probabilistically filling in missing values at the record level multiple times. Multiple systems estimation [e.g., @bird2018; @chao2001], performed on the imputed replicate files, was then used to estimate the number of missing observations, that is, the number of the victims never documented by any of the data sources used in the project.[^mse] To estimate the number of missing observations, we used a Bayesian latent class multiple-capture model [@manriquevallier2016] implemented in the `R` package `LCMCR`. The analyses presented in the technical appendix of the joint project combine these two methods to examine patterns of enforced disappearance, homicide, kidnapping, and forced recruitment of minors in the armed conflict.[^displacement]
 
 [^mse]: Multiple systems estimation is also called capture-recapture in some disciplines.