*****************************************************************************************************
* INTRODUCTION
* This .do file contains the commands to replicate my article's empirical analysis
* using STATA. Also please note that these data were updated to include a handful of observations
* (approx. 10) since the article was submitted to Harvard Negotiation Law Review. These data
* allow you to replicate the results reported, although not the exact statistics.
*
* The data include some variables broadly discussed in the paper but not analyzed in final models.
* I include this additional information to allow others to examine alternate specifications of
* the statistical models.
*
*****************************************************************************************************
* SET-UP AND LOAD DATA
* Please note that you may need to download the data and load it from your computer hard drive.
use "http://www.poliscidata.com/replication/caseObservationsReplication.dta"
*****************************************************************************************************
* DESCRIPTIVE STATISTICS
* Table 1: DEMOGRAPHIC COMPARISON OF COBB COUNTY AND UNITED STATES
* Table 1 is based on sources cited in text, not sample of mediation sessions
* Table 2: DESCRIPTIVE SUMMARY OF MEDIATION SESSION OUTCOMES
summ result
* Table 3: DESCRIPTIVE SUMMARY OF CASE TYPE AND COMPLEXITY VARIABLES
summ divorce complaint modification contempt visitation custody damages legitimate paternity sepMaintenance ///
childSupport galappointed pleadLessNotices
* Table 4: DESCRIPTIVE SUMMARY OF MONTHS WAIT VARIABLE
summ monthswait
* TABLE 5: DESCRIPTIVE SUMMARY OF JUDICIAL ASSIGNMENT VARIABLES
summ ingram bodiford staley robinson schuster grubbs kreeger nix flournoy
* TABLE 6: DESCRIPTIVE SUMMARY OF ATTORNEY-RELATED VARIABLES
summ numberlawyers smoothAttRate attOffset smoothAttSessions
* TABLE 7: DESCRIPTIVE SUMMARY OF ADMINISTRATIVE EFFORT VARIABLE
summ notices
* TABLE 8: DESCRIPTIVE SUMMARY OF MEDIATOR-RELATED VARIABLES
summ smoothMedRate medothercases medIsFemale medIsAttorney medHourlyRate
*****************************************************************************************************
* REGRESSION ANALYSIS MODEL
* TABLE 9: MULTIPLE-VARIABLE REGRESSION ANALYSIS OF MEDIATION SESSION OUTCOMES
logit result complaint modification contempt damages galappointed pleadLessNotices monthswait ///
grubbs smoothAttRate smoothAttSessions notices smoothMedRate medIsFemale medIsAttorney medHourlyRate medothercases
* TABLE 10: MARGINAL IMPACT OF SIGNIFICANT VARIABLES ON SETTLEMENT PROBABILITY
* Please note: you may need to install Scott Long's post-estimation package for STATA for "prchange" to work.
prchange
* TABLE 11: MULTIPLE-VARIABLE REGRESSION ANALYSIS OF MEDIATOR SETTLEMENT RATES
* Please see separate .do file (this table is based on another data file).
* TABLE 12: CROSS-TABULATION OF REPEATED MEDIATION SESSIONS OUTCOMES
table result repeatmed
* Note, if the result is partial settlement, I consider models to incorrectly guess outcome.
table result guessResult
* TABLES 14, 18
* Tables 14 and 18 are based on state-agency documents cited in text.
* TABLES 15 - 16 illustrate a theoretical concept.
* TABLE 17: PROCESS ACTUALLY USED IN CASES NOTIFIED TO MEDIATE BY CASE TYPE
* column 4 based on sample of mediated cases
table casetype
* column 5 based on sample of mediated cases, limited to settlements
table casetype if result ==1
* column 3 based on sample of cases referred to mediation during study period but not mediated
use "http://www.poliscidata.com/replication/resolvedPriorReplication.dta"
table casetype
* column 2 is the sum of case count in columns 3 and 4