Unobtrusive Research

Often researchers want to study social behavior without affecting it. In these cases, they often rely on data that is already exiting. They may analyze such written records as: government documents, reports, newspaper articles, photographic collections, letters, etc. For example, I wanted to know how the discourse on immigration had been presented by local media. In particular, I was interested in how immigrants had been depicted in photographs in the local newspaper. Using visual content analysis of images of immigrants found in the Greensboro News and Record I developed an online database of images by systematically reviewing daily newspapers from 1965 to 2006. These images were then analyzed, using a combination of inductive coding strategies (akin to Grounded Theory Method). Patterns, trends, and associations emerged from the constant comparison between images. Images were also tied to the major social and political happenings of the times to see how they are used to illustrate current events (an idea borrowed from Ethnographic Content Analysis. Comparisons were then made between the intended messages of the images (their connotation) and the unintended or hiding messages contained within them (denotation).

Obviously, unobtrusive methods, such as content analysis, are useful for looking at historical evidence. But, it is also a term applied to the statistical analysis of numerical data collected in the course of other research (secondary data analysis) or the combining of several datasets from disparate sources to study a social phenomenon (meta-analysis). Secondary analysis, like content analysis, makes use of already existing sources of data. However, secondary analysis typically refers to the re-analysis of quantitative data using often very sophisticated statistical techniques like structural equation modeling, hierarchical linear modeling, path analysis, etc. Likewise meta-analysis uses statistical techniques to merge and study a number of primary studies into a larger data-set providing a greater number of cases, historical periods, independent variables, contextual influences, etc. In both of these cases, unobtrusive methods allow researchers to look at an issue at a more macro-level then perhaps resources would allow them otherwise.

Example of Structural Equation Model for Job Dissatisfaction


I would like for you now to return to your Babbie textbook and read about some of the other methods of unobtrusive research in Chapter 11 Unobtrusive Research. Then take the chapter quiz on Blackboard.

%d bloggers like this: