18
MAR’s economic discrimination variable is coded for most individual years in the 1990s
(1990-1996) (Gurr 2002).
For every case, the maximum level of economic discrimination
exercised by a particular state against all of the minorities within its boundaries of the country is
recorded. Since the impact of economic discrimination on the likelihood of receiving intervention
for each case is the key concern, the value one year before the date of intervention is taken to
represent economic discrimination for that minority. If a value for the economic discrimination
level is not available for that specific year, the preceding value closest to the year in question is
used. For states that did not receive intervention, the mean value of economic discrimination for
the relevant period (1990-1996) is used.
Religious discrimination is coded using information on religious restrictions placed on
pursuit or expression of the group’s religious rights. The categories are as follows: (0) no
restrictions; (1) activity informally restricted; (2) activity somewhat restricted; and (3) activity
sharply restricted (MAR 2002 Codebook: 80). The variable is coded for the period between 1990
and 2000, based on values for 1990-91, 1992-93, 1994-95 and 1996. Recoding of this variable is
exactly the same as for the economic and political discrimination variables explained above. The
MAR-based religious discrimination variable is adapted to the host state level by taking the
maximum level of religious discrimination against any of the minorities within the state in the
same way as for the economic and political discrimination variables. For states that had received
intervention, the value of religious discrimination one year before that occurred is taken to
represent political discrimination for that state. Missing values are handled in the same way as for
economic discrimination. For the states that did not receive intervention, the mean values for
political, economic and religious discrimination are used for the period from 1990 to 1996.