$FL2@(#) SPSS DATA FILE MS Windows Release 13.0 spssio32.dll #########################Y@14 Apr 1021:

Problem 1: Operationalizing variables can be a challenge. One measure of a country’s human rights record is a measure known as the Political Terror Scale. This variable was developed by Mark Gibney at UNC-Asheville, and is coded for individual states according to the following criteria: Countries receive a score of 1 if people are not imprisoned for their views, and torture and political murder are extremely rare. Countries receive a score of 2 if there is some political imprisonment for nonviolent political activity, and torture and political murder are rare. Countries receive a score of 3 if there is extensive political imprisonment, and torture and political murder are common. Countries receive a score of 4 if Level 3 practices are extended to larger numbers of the population. Countries receive a score of 5 if Level 3 practices are extended to the whole population. To obtain these data, scholars used both Amnesty International Country Reports as well as Annual Reports from the US State Department. An astute observer might suggest that the use of State Department data may not be very reliable, because the US underreports abuses in countries it likes and overreports them in countries it does not. Is this the case? Using the human rights data for Assignment 4 (global_pts_comparison.sav), generate a correlation coefficient between the two measures (pts05am and pts05us). Using these findings, what can we preliminarily say about whether the State Department data is biased?