| Chapter : | Introduction |
|
|
Senator Tom Eagleton’s papers included a weekly tally of constituent mail received by the office, broken out by subject. One of the more intractable questions in political science is how constituent views affect member behavior. These data allow for a fine-grained analysis of that very question, and it is very likely that similar data exist in other congressional archives. David Parker, who worked with the Eagleton papers for his project on Senate campaigns, put a note in his research file to possibly explore this project in the future. The possibilities are truly endless.
Potential Weaknesses in the Archival Approach
Over the years, we have heard a number of potential objections to the use of archival data. We would be remiss if we did not present these objections and respond to them.
Reason One: Political Papers Are Sanitized
The most commonly expressed concern is that collections are systematically incomplete—that is, that critical information is missing due to conscious “sanitizing” by donors. Thus, the argument goes, there will be substantial gaps in the archival record that are created before they arrive at a repository. Deliberate removal and destruction of documents is particularly problematic because it introduces the statistical problem known as bias into the collections. Because such sanitation is systematic and not random by definition, it is especially important for users of archival data to consider the data generation process and how such bias may affect attempts at inference. To address the threat of this bias, it is worthwhile to understand how likely the systematic weeding of papers is generally and where bias is most likely to intrude.
Archivists work closely with potential donors before papers arrive at a repository; they negotiate the terms of a gift and, of course, play a critical role in the transfer of these collections to their repositories. As many archivists have noted in their correspondence with us, there are logistical