Analyzing Messy Data
There are several text books and articles with suggestions for analyzing data generated from questionnaires and other sources of codifiable data. They offer sophisticated statistical techniques and procedures for illustrating relationships of large data sets and large numbers of variables. Analysis has been described as a process of "cutting the problem down to size" or reducing it into smaller more comprehensible parts. This is a process of solving the component parts of the problem and assembling them into a solution for the larger problem. 1
William James, a nineteenth century psychologist, made the observation that people do not start a day's work with problems to be solved. The start of a day resembles a "great big buzzing confusion" or what his student John Dewey called "an indeterminate situation." Russell Ackoff called this situation a "mess," and indicated that what we usually confront are messes and not problems. 2
If information and organizational data is so conceptless and messy, how do we take steps to analyze it? This chapter first provides a perspective on the systematic errors that are likely to occur in field research. It then outlines methods for improving the value of the information collected and analyzed.
Researchers can expect to find three types of systematic errors when they attempt to use qualitative data collection and analysis procedures in field settings. These errors occur during a study's composition and conceptualization, implemention and data collection and recording. These biases emerge from the mismatch between research requirements and field setting dynamics.
The composition and conceptualization of the research can produce systematic errors which may "add," "distort," or "delete" unique information. 3 For example, certain relevant variables might not be included and irrelevant ones