1
INTRODUCTION
Over the past twenty years, Mexico has undertaken dramatic economic and political reforms, the results
of which have greatly impacted fundamental relationships in Mexican society. When I set out, in 1998, to
begin a research program on how these reforms – and concurrent social changes – have shaped the nature
of collective action and protest in Mexico, I quickly discovered that, while there is a long tradition of
social movements research on Mexico, there existed no broad collections of data that recorded protest as a
general and nationwide phenomenon. In recent years, I have begun to construct the Mexico Protest Event
Database (MPED). This project is still in its nascent stages, consisting of 3200+ coded events compiled
from the state news agency, Noticias de Mexico (Notimex), for the two-year period of 1999 through 2000.
Over the next two years, coding will extend this coverage through 2006. It is hoped that, in the years to
come, this can be extended backward in time as well.
The nature of this data collection project is explained in greater detail elsewhere (see Strawn 2005). In
this paper, I outline how a protocol was developed to search electronic archives containing Notimex
articles. The discussion that follows proceeds in three steps. First, I introduce Notimex as the data source
used for the broader MPED project, addressing the source itself and why this was chosen over possible
alternatives. Second, I examine methodological concerns related to the procedure of sampling news
articles, and explain how the search protocol used here was developed iteratively. Third, I evaluate the
overall effectiveness and efficiency of the specific search terms in the protocol using data on search hits
and word counts for the more than 18,500 articles that were produced by the sampling procedure and
coded for 1999-2000 period currently coded.
Page 1 of 20 
Convention