The Limitations of the Slave Narrative Collection: Race and Representativeness

Attendants at Old Slave Day, Southern Pines

It is probable that the interviewer's race affected an informant's response. As noted earlier, the staffs of the Writers' Projects in the states in which former slaves were interviewed were overwhelmingly white. The relative absence of black interviewers introduced an important source of bias, for the interviewer's race was a significant factor in eliciting responses from the former slaves. The etiquette of Southern race relations influenced the definition of the interview situation for these aged African Americans, and some of their interviewers were even members of the former slaveholding families. As a result, informants may frequently have told their white interviewers "what they wanted to hear." For similar reasons many were undoubtedly less than fully candid or refused to tell a complete story, resulting in a kind of self-censorship.

Fortunately, some black interviewers were involved in obtaining the narratives, and it is therefore possible to assess the influence of race by comparing the responses of those interviewed by whites with those of people interviewed by blacks. For example, in a systematic analysis of the Slave Narrative Collection and similar interviews obtained earlier by Fisk University interviewers, Paul D. Escott found that 72 percent of the ex-slaves interviewed by whites rated the quality of their food as good, while only 46 percent of those interviewed by blacks did. Similarly, 26 percent of those responding to white interviewers expressed unfavorable attitudes toward their former masters compared to 39 percent of those who responded to black interviewers.28

However, it is important to recognize that distortion is not something inherent in the interview situation, but is relative to the specific questions asked by the researcher. There are questions and issues that were not affectively charged by the race of the interviewer. For example, in a systematic analysis of family patterns in the ex-slave interviews, Herman R. Lantz found no evidence that the race of the interviewer affected the overall reporting of family relationships.29 Therefore, depending upon how unobtrusive or subtle the measures used by the researcher, how much the race of the interviewer affected responses is an open question.

Another problem surrounding the Slave Narrative Collection concerns the representativeness of the informants. Despite the large number of interviews obtained by the Writers' Project, less than two percent of the available ex-slaves were interviewed. This in itself is not an insurmountable problem, except that it has been impossible to determine the processes by which informants were selected. There appears to have been little concern in the Writers' Project for systematic sampling procedures or for obtaining a representative sample of the former slaves, since the problem is nowhere mentioned in the project's extensive correspondence. The skew of the sample can be seen simply in the following figures: while blacks over eighty-five years of age lived primarily in rural areas in the 1930s, those whose accounts are found in the collection were overwhelmingly urban residents. Apparently the primary basis for selection was availability; those in closest proximity to the cities in which the Federal Writers were based were most likely to be interviewed. Finally, the number of interviews obtained by each of the participating states varied considerably, ranging from three in Kansas to nearly seven hundred in Arkansas.

NEXT: Should the Slave Narrative Collection Be Used?

Slave Narratives:

An Introduction to the WPA Slave Narratives