sciences (see American Heritage Dictionary Editors 1992; Fagan 1985, 1991a, 1991b; Oxford Dictionary Editors 1995; Podolefsky and Brown 1993; Rathje 1996; Thomas 1979; Turnbaugh, Jurmain, Nelson, and Kilgore 1996). Perhaps the defining characteristic of all garbology “digs” is that they combine traditional concerns of archaeological method and theory to produce results that are immediately relevant to understanding and mitigating current social dilemmas (Rathje 1996). The highly publicized “garbage crisis” of the 1980s literally had the Garbage Project’s name on it and made it relatively easy to convince the public at large that the study of contemporary refuse provided a significant contribution to society. The “crisis,” however, was not given great media coverage until the Mobro 4000 garbage barge, which took garbage out to sea, sailed in 1987. During its first fourteen years, the Garbage Project studied problem areas that were less literally “archaeological” in nature.

The Garbage Project’s first data collection format, called “the regular sort,” was designed to sample and record household pickups of fresh refuse (a “pickup” is all of the materials placed out by a single household on one regular refuse collection day). From the beginning, project procedures have rigorously protected the anonymity of the households that discarded the refuse that was sampled.

Solid waste managers have been characterizing wastes by material composition (paper, plastic, glass, etc.) and weight since the 1880s. To these traditional measures, the Garbage Project added a series of innovations, including records from package labels (brand, cost, solid weight or fluid volume of original contents, specific type of contents, packaging materials) and more detailed breakdowns of refuse categories, such as “food waste”—separated into “once-edible food” versus “food preparation debris” and both identified by specific food item—(Hughes 1984). Because of their exacting level of detail, the regular-sort data files that document residential refuse are ideal for analyzing the role of specific household behaviors in generating wastes. Today, the Garbage Project’s fresh refuse records, compiled from the long-term ongoing study in Tucson, Arizona, and short-term studies in five other cities, form a one-of-a-kind database that in the year 2000 encompassed twenty-seven years of time depth.

Garbage Project studies of fresh refuse have consistently documented a few basic patterns in the way we interact with the material world around us: First, what people say they do and what they actually do are often different. For example, while respondents rarely report to interviewers that they waste any food at home, two decades of Garbage Project studies have documented that households generally waste about 15 percent of the solid food they buy (Fung and Rathje 1982; Rathje 1976, 1986). Such misreports characterize a broad range of household behaviors. In other words, people who are interviewed or fill out surveys do not accurately report how much food they waste, what they eat and drink, what they recycle, or the household hazardous wastes they throw away (Rathje and Murphy 1992a, 1992b).

This discovery, of course, is not a great surprise. It is common knowledge among behavioral scientists that any methodology that depends upon the accuracy of answers given to interviewers or in surveys suffers from problems of informant bias (Webb, Campbell, Schwarts, and Sechrest 1966). Respondents may not be able to accurately and quantitatively recall specific behaviors, such as how many ounces of green beans they ate the day before or how often they discard a half-full container of pesticide, and even if respondents can accurately recall behaviors, such as beer drinking or changing the oil in their cars, they may not want to admit to the specifics.

At this point it should be noted that systematic sorts of garbage avoid informant biases. Refuse data, like virtually all archaeological data, are quantitative: packaging and commodity wastes can be weighed, measured for volume, and chemically analyzed, and their labels can be read for further information, all without relying upon the memory or honesty of respondents. When refuse is identified by specific household (versus recording only the generating household’s census tract), the Garbage Project obtains permission from the discarders.