The Limitations of Stranger-Interviewers in Rural Kenya Alexander A. Weinreb Hebrew University, Jerusalem Virtually all survey data are collected by “strangers,” that is, individuals with no prior social relationship with respondents. Although it has long been recognized that attitudes toward strangers vary cross-culturally, there has been no systematic discussion of how this variation might affect survey data. This article attempts such a discussion, using data from a longitudinal research study in rural Kenya. It reviews qualitative impressions of insider- and stranger-related issues within the specific Kenyan field setting, drawing primarily on field notes. Relevant areas of social theory and the data collection literature are reviewed briefly. Finally, using the project’s longitudinal survey data, empirical tests are presented which allow for an evaluation of differential data quality across insider- and stranger-interviewers on three dimensions: differential response rates, differential reliability of responses, and differential response validity. The results suggest that insider-interviewers increase response rates and collect more consistent data across survey waves. They also suggest that data collected by female insiders in particular appear to be superior for most questions and of equal quality for others.

Delivered by Ingenta to : Since there may be marked differences toward the strangers who interview Hebrewamong Universityrespondents of Jerusalem Mon, 31 Dec 09:13:05 people in different countries in conceptions of 2007 them, and that attitudes toward strangers vary what it means to give information to strangers cross-culturally. Yet a century of social research .|.|. careful comparative studies are imperative has yielded no systematic review of relevant before we can know how generally true speciftheoretical or methodological issues concerning ic findings may be .|.|. the collection of survey data by strangers. Nor —Sudman and Bradburn (1974:147) has it yielded any estimates of the size and scope of “stranger-interviewer” effects.1 Rather, where the issue has cropped up at all, it has t has long been recognized that the validity of been in either tangential asides in early methodsurvey data is affected by the attitudes of

I

Direct correspondence to Alexander A.Weinreb, Department of Sociology and Anthropology, Hebrew University, Jerusalem, Israel 91905 ([email protected]). The data used in this article were funded by grants from the Rockefeller Foundation and National Institutes of Health to Susan Watkins and Jere Behrman. The author owes an incredible debt to Susan, who encouraged this research project at every opportunity, and gave freely of her field notes and time. The author also thanks Jerry A. Jacobs and Amelia Weinreb for their extremely helpful substantive comments on a number of versions, and Rafe Stolzenberg, Herb Smith, Randall Collins, and four anonymous reviewers for their comments on the earliest versions.

1 There has been no systematic experimental, empirical, or conceptual treatment of this issue in any single article in the American Journal of Sociology, volumes 1–107; American Sociological Review, volumes 1–69; Public Opinion Quarterly, volumes 1–66; Sociological Methodology, volumes 1–32; Sociological Methods and Research, volumes 1–31; Social Science Research, volumes 1–31; Population Studies, volumes 1–56; or Demography, volumes 1–39. Nor have I found any such treatments in scholarly monographs that in all other ways effectively review and summarize the core literature on response effects (e.g., Fowler and Mangione 1990; Hagenaars and Heinen 1982; Tourangeau, Rips, and Rasinski 2000).

AMERICAN SOCIOLOGICAL REVIEW, 2006, VOL. 71 (December:1014–1039)

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1015

ological reviews, wherein the possibility of such the area, unless they also had some personal, effects has been acknowledged but then ignored usually familial, connection to someone in the sampled villages, were called “outsiders” or empirically (Mauldin 1965; Stycos 1960; “foreign” (Kawadhgone 2000:112).2 Sudman and Bradburn 1974), or in more speSurvey researchers use stranger-interviewcific criticisms of survey research as practiced ers (i.e., interviewers who have no prior social in the developing world (Heisel 1968; Stone relationship with a respondent or a respondent’s and Campbell 1984). family) for two main sets of reasons. The first The key aim of this article is to address this is purely practical. An insider is restricted to gap in the literature by presenting baseline estihouseholds that she or he knows. In contrast, a mates of insider- and stranger-interviewer effects stranger-interviewer can knock on any door or in a single rural area of Kenya. Its structure call any number. Stranger-interviewers are therefollows from this. In the first of four main secfore more flexible data collectors than insidertions, I define “insider” and “stranger” interinterviewers. This practical difference also has viewers, briefly describing the perceived implications for sampling and research design. advantages of stranger-interviewers over insidOver and above these practical concerns, er-interviewers. In the second section, I briefly however, more traditional epistemological reareview related issues in the extant sociological sons exist for using stranger-interviewers. These and data collection literature. In the third secbuild on one of two ideas about micro level tion, I describe initial qualitative impressions of social interaction: either people are “more likeinsider- and stranger-related issues within the ly to be reticent with those who are already specific Kenyan setting in which this research known to them” (Bulmer 1993:215) as a genwas conducted, drawing primarily on field notes eral rule, or an interviewer–respondent interthat describe respondents’ reactions to interaction is a distinct type of social interaction in viewers and to the research project. Then in the which to this: is the case. final section, I present empirical tests, noting Delivered by Ingenta The origins for each of these ideas also can signs of differential data qualityHebrew across insidUniversity of Jerusalem 31 Dec 2007 09:13:05in Simmel (1950a). Thus, for exambe located er- and stranger-interviewers in theMon, associated ple, the argument that strangers in general, but survey data on three levels: (1) differential especially “the stranger who moves on .|.|. often response rates, (2) differential reliability of receives the most surprising openness—confiresponses, and (3) differential response validdences which sometimes have the character of ity. a confessional and which would be carefully withheld from a more closely related person” BACKGROUND (Simmel 1950a:404). The reason for this openThe single most important characteristic that ness, Simmel suggests, is that a stranger typidistinguishes insider-interviewers from strangercally is deemed to have more objectivity, which interviewers is whether they know or are known arises from both “a particular structure comby the respondent or the respondent’s family. posed of distance and nearness, indifference, “Know” refers to a social relationship unrelatand involvement” (p. 404) and “freedom .|.|. ed to and predating specific interactional roles from commitments which could prejudice his associated with interviewers and respondents. perception, understanding, and evaluation of This definition of “insider” and “stranger” is the given” (p. 405). derived from two main sources. The first source Over decades of research activity, the perdraws on a classical sociological theme, formance of stranger-interviewers, as opposed Simmel’s reference to “entanglement in family and party interests” (Simmel 1950a:404) as the key characteristic of an insider. The second 2 Kawadhgone (2000:112) refers to page 112 of source draws on emic perspectives specific to field notes from the field site of Kawadhgone writthe Kenyan research settings. As described later, ten during the 2000 wave of data collection. Other site a perpetual source of tension during data colnames with associated field notes on which I draw lection was the extent to which the research in what follows are Gwassi (1994, 1995, 2000), project used individuals affiliated with the samOyugis (1994, 2000), and Wakula South (1994, pled communities. Those from other villages in 1996).

1016—–AMERICAN SOCIOLOGICAL REVIEW

1978), and (3) the literature on deception to the more general types of strangers described by Simmel, has tended to support these hypothe(Barnes 1994; Simmel 1950b). It also is inconses. In the early days of survey research (e.g., sistent with (4) elements of macro-oriented thesee Hyman et al.’s (1954) discussion of sex surory, particularly ideas about how trust veys in the United States and the United conventions (the propensity to trust or mistrust Kingdom), researchers were surprised at how a stranger) are constituted (Douglas 1970; readily respondents seemed to discuss deeply Durkheim [1926] 1964; Giddens 1990; Levy personal issues with strangers. Since then, 1972; Molm, Takahashi, and Peterson 2000). researchers have tended to forget that early surSecond, a number of methodological experprise. They also have developed techniques that, iments make it clear that survey respondents do by framing either specific questions or the envinot neatly distinguish between interpersonal ronment in which the question is asked, increase norms and task-oriented interview-related reporting of sensitive, offensive, or threatening norms, even in highly differentiated societies for behaviors, giving epistemological cover to which it can be reasonably assumed that indistrangers. The result has been the proven effects viduals are used to telling quite personal things on data quality of “confidentiality assurances,” to professionals of one sort or another (e.g., promises of “response anonymity,” telling the doctors, lawyers, therapists). Rather, there is respondent that his or her views are important, considerable overlap between these two types of and attempts to minimize interactional biases by interpersonal communication, suggesting that ensuring privacy during the interview or by communication between interviewers and increasingly substituting noninteractional data respondents alternates between formal rule folcollection modes such as self-administered lowing and tacit practices (e.g., Maynard and questionnaires and audio- and computer-assistSchaeffer 2002; Schaeffer and Presser 2003). ed interview techniques (Audio-CASI & CASI) Indeed, the fact that mainstream survey method(Aquilino 1994; Krysan et al 1994; Mangione, ologiststoare: increasingly adopting more flexible Delivered Hingson, and Barrett 1982; Mensch, Hewett,by Ingenta of interviewing over fully standardized— Hebrew Universitystyles of Jerusalem and Erulkar 2003; Singer, von Thurn, and Miller Mon, 31 Dec 2007 and09:13:05 impersonal—interviewing techniques, 1995; Taietz 1962). shows that this overlap can be leveraged to increase the accuracy of survey responses GENERAL LITERATURE (Cicourel 1974; Dijkstra 1987; Schober and Conrad 1997; Suchman and Jordan 1990). As noted earlier, no extant methodological litAs yet, no systematic methodological erature exists that deals with the relative effecresearch of this type has been conducted in tiveness of insider- and stranger-interviewers. developing societies. But field researchers’ However, the more general literature on accounts cast considerable suspicion on the strangers is, frankly, enormous. I have reviewed assumption that respondents distinguish relevant themes of this literature in some detail between interpersonal norms and task-oriented elsewhere (Weinreb 2000). But it is useful to at interview-related norms. For example, there are least summarize two key issues here. numerous accounts of culturally sanctioned, First, the basic proposition that respondents and even culturally prescribed, deception of are more honest with strangers than with insidstranger-researchers in developing societies, ers, the old Simmelian claim, draws on assumpsome with comical and others with quite seritions about microsocial relations that at best ous results (e.g., Agar 1980:61; Anderson are inconsistent with and at worst antithetical to 1986:342–3; Barley 1983:89; Dentan 1968:92; a large body of social theory at the micro level. Stone and Campbell 1984:32). Bulmer (1993) This body of theory includes (1) historical and refers to these as the “sucker bias,” a phenomcontemporary equation of truth-telling with enon specific to strangers, because insiders, intimacy (Collins 2004; Garf inkel 1967; having mastered local cultural codes, know how Goffman 1971; Tönnies 1961), (2) the experito avoid being suckered. From a methodologimental and semiotic literature on microsocial cal perspective, Agar (1980) and Stone and interaction (Breault, Barker, and Lemle 1987; Campbell (1984) are particularly significant Catania et al 1996; Christenfeld et al. 1997; Dindia, Fitzpatrick, and Kenny 1997; Givens because they match prior survey reports from

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1017

given respondents in rural India and Nepal, respectively, with postenumeration reports gathered from informal interviews in communities where they were embedded. In both cases, respondents reported telling untruths to the survey interviewers because the latter were strangers whose motives were unknown and therefore suspect. Finally, although the substantive focus of this article is on insider- and stranger-interviewer effects, analyses of response patterns also consider the interaction between gender of the interviewer and insider–stranger issues. This builds on a long of tradition of research investigating different types of “role independent” interviewer effects in general (much of it reviewed in Fowler and Mangione [1990], Hyman et al. [1954], Sudman and Bradburn [1974]), and on a small literature comprising interviewer gender effects in developing-country settings (Axinn 1991; Becker, Feyistan, and Makinwa-Adebusoye, 1995; Bignami-van Assche, Reniers, and Weinreb 2003; Blanc and Croft 1992).

lected three rounds of survey data between 1994 and 2000.3 The main focus of the project was on the role of social networks in contraceptive and AIDS-related behavior. It sampled from a population of ever-married women and the husbands of those currently married. Although the specific survey instruments changed somewhat across time, each wave of the survey asked a series of questions about relatively sensitive topics including household economics, gender relations and attitudes, behaviors and conversational networks related to AIDS, and family planning. More generally, as a setting for fieldwork, South Nyanza, as well as the other predominantly Luo districts in Nyanza province (all districts except Kisii and the new Nyamira and Suba districts) is one of the poorest areas in Kenya.4 (Weinreb 2001a). Some of this poverty is attributable to the area’s relative distance from the hubs of economic and nongovernmental organization networks in Nairobi. However, according to both the popular imagination and the popular political discourse IMPRESSIONS FROM THE FIELD amongtothe: Luo, the main cause of Nyanza’s Delivered by Ingenta Hebrew Universityimpoverishment of Jerusalem is the fact that, as an ethnic Following methodological norms, officially Mon, 31 Dec 2007 09:13:05 group, Luo have en masse been associated with sanctioned data collection practice in subthe political opposition since at least 1966, Saharan African countries, as elsewhere, which has led to reduced flows of government includes choosing interviewers from a cohort of spending, either directly or through parastatals professionals who have appropriate language (Weinreb 2001b). Either way, residents of South skills, then bringing them into the sampled area Nyanza District are aware of their relative lack for the duration of data collection. In the subof access to various types of capital and public sequently discussed research project, this option was too expensive for us, forcing us to use local interviewers. This gave rise, albeit organically 3 All these data, the survey, and the field notes, are rather than through official design, to an internow in the public domain (available at http:// est in insider–stranger issues. kenya.pop.upenn.edu). Also, more specific inforIn this discussion, it is useful to introduce the mation on sample characteristics, data quality, and so research project from which the data in this on is presented at the beginning of the empirical secarticle are drawn, and to describe two important tion. characteristics of the field setting. I then sum4 Since the sample was initially drawn, South marize key impressions about insider–stranger Nyanza has been subdivided into three districts. differences that emerged from the fieldwork Consequently, two sites (Gwassi and Wakula South) and use these to anchor ideas that I subsequently currently are in Suba District; one (Oyugis) is in test in the analysis section. Karachuonyo District; and the fourth (Kawadhgone)

THE SETTING I use data from the Kenya Diffusion and Ideational Change Project (KDICP). Based in four sites in South Nyanza District, Nyanza Province, in southwest Kenya, the KDICP col-

is in Homa-Bay District. Irrespective of this redistricting, this article uses the slightly outdated district nomenclature as the summary name, preferring it over both the three new names and less geographically specific alternatives such as “Nyanza Province,” which includes other districts, including the nonLuo Kisii District.

1018—–AMERICAN SOCIOLOGICAL REVIEW

goods relative to parts of the country that historically have been more favored (particularly areas of the Central and Rift Valley provinces). In each of the four sites and across all three rounds of the survey, we selected interviewers from the large local pool of under- or unemployed Form 4 graduates, that is, high school graduates who lived in or around one of the four research sites, or who happened to be visiting the area when interviewer positions were advertised. Inevitably, this meant that a proportion of the interviews conducted during the project were performed by interviewers who claimed to know the respondent or someone in the respondent’s household well. Because our approach deviated from the widely accepted standard, there was discussion from the outset about the bias that might ensue from using local interviewers.

ideas. In relatively traditional patriarchal settings such as Nyanza, these alone were likely to raise at least some hackles. Finally, suspicions also fed on more local political perceptions. Particularly in 10 semistructured interviews with older men (transcripts available at the project Web site) and countless informal conversations, local residents displayed a standardized “political opposition discourse” centered on the idea that Luos had borne a disproportionate share of the costs of independence, both before and after Kenya’s transition to multiparty democracy in 1992. Among the core ingredients of this discourse were conspiratorial theories about the murder of their most able leaders, the relative absence of effective Luo representation in the central government, and systematic discrimination against them in terms of public services provided. Although the specific implications of these factors varied somewhat across informants, all RESPONDENTS’ VIEWS OF INTERVIEWERS respondents in some way expressed a lack of As KDICP data collection proceeded, we conf idence in the cur rent authorities. became increasingly comfortable with our Unfortunately, as with large field projects in choice of local interviewers. At the root of this many other Delivered to : settings, it was these same authorcomfort was our growing awareness of two fac-by Ingenta ities from which we—the KDICP—needed to Hebrew University of Jerusalem tors. First, tension between local suspicion of our receive official permits to conduct our research. Mon, 31 Dec 2007 09:13:05 motives, on the one hand, and appreciation for Whether because of a single one of these our economic contributions and the opportunifactors or a combination of them, there was ties they afforded some local residents, on the some suspicion of the KDICP among individother hand, appeared to generate data collection ual respondents. This manifested itself in an problems. Second, insider-interviewers appeared unwillingness to believe our mission as laid out to be better at minimizing these problems than in the introductory statement, which, following stranger-interviewers. I discuss each of these in standard survey protocol, every interviewer turn. made to a respondent before the formal start of the interview. Instead of that stated mission, or perhaps in addition to it, respondents appeared LOCAL SUSPICION. Local suspicion of the to assign us motives that more neatly fit local KDICP stemmed from three factors. First, the discourses, their own ends, or both. relatively low levels of social differentiation in The most extreme of these motives were a these communities meant that respondents typsmall number of unusual but influential cases ically had quite limited experience with strangerthat signaled respondents’ quite serious misoriented interactions in general, or certainly trust of us and our intentions. During a number with anonymous interrogators pushing them to of interviews, KDICP interviewers were told to assign motives to our presence in lieu of this “shake the dust from their feet” (Gwassi experience. As one suspicious mother-in-law 1995:24) (i.e., to get out quickly). In a handful of a respondent put it: “You people .|.|. you must of cases, they were chased away by panga be after something” (Gwassi 1995:24). (machete)-wielding male respondents who Second, because of its partial focus on conaccused them of being agents of the government traception, the KDICP often was publicly assoor devil worshippers. On one occasion, a responciated with the family planning movement, dent wielding a panga in one hand and a Bible which hurt it in some quarters, while probably in the other arrived at our doorstep early one helping it in others. Equally damaging was its association, in local eyes, with overtly feminist morning, the day after he had been interviewed,

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1019

projects in many developing countries (Belcher demanding that we return the hard copies of his et al. 1976; Bulmer 1993; Choldin, Kahn, and and his wife’s questionnaires. He claimed not to Ara 1967; Stone and Campbell 1984). In the have slept at all the previous night because of KDICP, it manifested itself in a number of ways. concerns that he and his wife would die if we First, it drew on the fact that, as noted earlidid not give them to him. We did so (Gwassi er, we were using people, most of them from the 1995:35). More frequently, this devil-worshipimmediate area, and that the general lack of ping theme played itself out in an outright employment opportunities in the area generatrefusal to participate in the survey, although ed intense competition for jobs as interviewers, unaccompanied by even an implicit threat of also pushing chiefs and others into requesting, violence. Most commonly, however, it simply and sometimes demanding, that we employ cerpeppered other local conversation about the tain people (Gwassi 1995:19; Gwassi research project.5 2000:58–59; Oyugis 2000:54). Indeed, in some All in all, these responses—and other angry cases, respondents refused to participate in the ones such as “those whites, they want to know survey if no one from their own immediate or what we think and they go back and laugh” extended family was hired in any capacity by the (Gwassi 1995:12)—resemble colorful examproject (e.g., Gwassi 1995:22). In the most seriples drawn from early fertility and household ous of these cases, a whole clan declared that surveys in Jamaica (Back and Stycos 1959), its members would refuse to participate in the Bangladesh (Choldin, Kahn, and Ara 1967), survey after it became known that no prospecIndia (Agar 1980), and elsewhere in Kenya tive candidate associated with the clan had (Kearl 1976). In these cases, too, stranger-interpassed the interviewer selection tests. Only after viewers were alternately thought to be agents of we threatened to pull the project from the whole the secret service, army conscription agencies, area—a threat that speedily engaged the interor a compulsory population control program, or ventiontoof: the local chief, assistant chief, and simply “godless women heralding theDelivered end of theby Ingenta of other clans (whose children or vilworld” (Choldin, Kahn, and AraHebrew 1967:251). Universitymembers of Jerusalem lagers’ children were directly benefiting from Mon, 31 Dec 2007 09:13:05 These issues are directly relevant to debates our presence)—were the leaders of the recalciabout how effective stranger-interviewers could trant clan persuaded to lift their embargo be in our setting, for the simple reason that the (Wakula South 1996:70–74). ulterior motives respondents ascribed to us Second, in a similar vein, there was also at appeared to be mediated through the social least one case in which a chief was accused of identity of the interviewer. Moreover, in the taking bribes to change the names on our intercase of the KDICP, it was the insider-interviewer selection list (Gwassi 1995:19), although viewers’ mediation that appeared to work more he had no access to the lists. In another case, a effectively than that of stranger-interviewers in local leader was accused of minimizing advercalming fears or reducing suspicions. Examples tising of the positions outside his clan (Oyugis of this mediation are given in the following dis1994:18). cussion. The KDICP also was considered a potential source of economic assistance or development INTERVIEWS AS AN OPPORTUNITY FOR GAIN. If beyond its role as an employer. Friendly inforsuspicion caused only a small minority of indimal conversations with locals, whether they viduals to refuse outright to participate in the were in the sample or not, often were followed survey (actual nonresponse rates are presented by “invitations” to assist them with school fees, in Section 4), it was largely because the project college fees, money for setting up a business or was also seen as a potential source of economfor local development infrastructure, and so on. ic assistance or as an agent of development. Less frequent, although not uncommon, were Again, this general problem affects research offers of marriage or even land (on which either one of the foreign researchers or Kenyan university-educated field supervisors would settle), 5 See Gwassi (1995:21, 25; 2000:25, 81, 92) for each of which would entail the establishment of patron–client ties, and therefore transfer oblimore details on why some of the respondents came gations. In short, we repeatedly received the to associate our project with devil worship.

1020—–AMERICAN SOCIOLOGICAL REVIEW

you should be straightforward about that. impression that locals’ attempts to cultivate (Kawadhgone 2000:125) relationships with the presumably wealthy and well-connected researchers and senior field staff Finally, this desire to appear worthy of our prewere, in part at least, instrumental. Properly sumed development role also emerged in local harnessed, we were their pathway to “barely leaders’ advice to respondents and in their inforimaginable opportunities” (Kawadhgone mal comments to us. For example, respondents 2000:100). reported being told by one local health educaMost commonly of all, however, our pretor “to keep their houses very clean in case we sumed development role generated daily demonvisited” (Wakula South 1994:26). And in inforstrations of respondents’ desire to give us what mal conversations with field staff, local informwe wanted so that we might effectively help ants sometimes distanced themselves from the them as individuals, or else help their commudevil-worshipping rumors that they thought nities. At the simplest level, this emerged in the might damage the project’s perceived developfield staff’s daily reports of respondents asking ment potential, often painting the accusers as “what is this interview going to do for me?” But more general troublemakers (Gwassi 1995:35). this motivation to benefit also emerged in other Alternatively, they exaggerated local attitudes ways. For example, in a handful of cases, a toward core project interests to make the area respondent was impersonated by someone else sound more deserving of our attentions. For (e.g., Oyugis 1994:12) or, because of a logging example, during the first wave of the survey, error on our part, we mistakenly interviewed the when the substantive focus was on contracepsame respondent twice within the same few tive use, one local leader reported that “everydays without the respondent acknowledging the one” likes family planning (Gwassi preceding interview (Gwassi 1995:47). In each 1995:14–15)—a complete falsification accordof these cases, the respondents later reported ing to data on contraceptive prevalence rates for thinking that they might derive someDelivered extra ben-by Ingenta to : the immediate area (e.g., from both KDICP efit from having played these (extra) roles as Hebrew University of Jerusalem data and the South Nyanza subsample of the respondents. However, we also saw signs31ofDec a 2007 09:13:05 Mon, 1993 Kenya Demographic and Health Survey desire to benefit, individually or collectively, in data). Moreover, by the final wave, in which the respondents’ sometimes frustrated response to project’s core substantive interest had shifted to certain questions, including some seemingly AIDS, a local leader reported that 80 percent of innocuous ones. They could not figure out why the local population had died of AIDS and that we were asking a specific question, and conse90 percent of the children younger than 5 years quently could not figure out the best way to were orphans (Gwassi 2000:86–7). This too respond. In one case, for example, a female was absolutely groundless empirically. But like supervisor conducted an interview with a difthe family planning comment 5 years earlier, it ficult male respondent: sent us a message. Local leaders thought this is When she got to the “when were you and your first what we wanted to hear, or that it was the best wife married” question, [the respondent] said way to engage us in solving local development “That’s the sort of question I won’t tolerate. What problems. I return to some of these issues in the will this help me with?” He answered a few more empirical Section 4 because they indicate how [questions], but at the radio question [Do you own social desirability biases were constituted, in a radio?] blew up, and said: “Why are you asking predictable ways, thereby allowing for some about radios? Are you going to give me one?” analytic leverage. (Gwassi 1995:11–12)

In another case, a male respondent physically took the questionnaire from the interviewer and told her to leave. When the supervisor visited him to ask for an explanation, and for the questionnaire, he told her: This research is useless; you can’t be asking what we own; you should offer a solution like Bamako and CARE. If you want family planning you should be straightforward about that; if you want AIDS

THE LIMITS OF PRIVACY. Our comfort with local interviewers also grew as we realized that the effectiveness of the standard methods used by the KDICP to minimize interactional biases in the data also appeared to be affected by the tension between local suspicion of the KDICP and the opportunities represented by the project. For example, we instructed interviewers to

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1021

try to minimize interruptions from the outside Similarly, and perhaps more important, the and to stop any third parties from listening to, KDICP also used confidentiality assurances or otherwise interfering in, the interview. The and promises of response anonymity, following problem was that in addition to the layout of the what has long been standard protocol for develhouseholds and their age structure—respecoping societies (e.g., Demographic and Health tively, dwellings often were small and clustered Surveys 1987:46; United Nations 1961). Here in extended family (multigenerational) comtoo, however, we began to suspect that the credpounds, and high fertility meant that individual ibility of these assurances was only as strong as families within these compounds also tended to the respondent’s confidence in the promise giver, be large—our attempts to restrict the flow of in this case, the interviewer, the organization she people in and out of earshot were partly underor he represented, or both. Without that confimined by the fact that the very act of particidence, no matter how vociferous the promises pating in the survey as a respondent signaled the were, it seemed they did relatively little to negate ownership of knowledge. Thus, we came across people’s fear of telling the “wrong people” about mothers-in-law or husbands who, because they an undesirable activity or attitude, where the felt threatened by their daughters-in-law or wrong people increasingly appeared to be wives being interviewed in private, would try, strangers over whom respondents had no leversometimes successfully, to interrupt or terminate age.8 the interview;6 village headmen who directed us to more educated members of the community INSIDERS AS PROBLEM SOLVERS or failed to list some of the more problematic Over and above repeated reports that by hiring households during our initial construction of locals as interviewers we were providing some sample lists (they were found later during fieldtangible “benefits” to the community—conwork); headmen or other higher-status individsideredtoa :positive thing locally, even with the uals who demanded to be interviewed, even ifby Ingenta Delivered of Jerusalemit generated—we emerged from competition they did not fall into the sample Hebrew frame, toUniversity sig31 for Dec 2007 the09:13:05 field with considerable anecdotal evidence nal their status and, implicitly, theirMon, support that insider-interviewers were able to assuage our “development” project (see Rudolph and respondents’ fears much more effectively than Rudolph [1958] for similar status-related themes stranger-interviewers. In some cases, insiders in India).7 did “damage control” after respondents had demonstrated their unhappiness with strangers. These stories fall into three main categories. 6 More generally, female field supervisors mentioned university men, complaining about “modern women” being those who “were always talking back” (Wakula South 1994:25). It is reasonable to assume that parts of this attitude also existed at the village level, with implications for the behavior of women respondents. 7 Recently, a small number of researchers who work in developing country settings have sought to minimize interactional biases by experimentally substituting self-administered questionnaires or CASI/Audio-CASI modes of data collection for faceto-face interviewing. The results to date are mixed, generating somewhat higher reports of sensitive behavior, but with some unexpected error (Mensch 2006, personal communication; Mensch, Hewett, and Erulkar 2003; Plummer et al. 2004). This switch to alternative modes does not in itself resolve the insider versus stranger issue. It is possible, for example, that CASI/Audio-CASI would yield even higher reports of sensitive behavior if they were introduced to respondents by insider-interviewers.

COMPLETING INTERVIEWS. Frequent reports suggested that insiders could more successful-

8

No extant literature demonstrates the utility of these types of assurances in developing country settings. For example, all 113 of the studies analyzed in the Singer, Von Thurn, and Miller’s (1995) metaanalysis of the utility of confidentiality assurances were conducted in developed societies. In developing societies, in contrast, it has been asserted that “[t]he anonymity of survey research is generally incomprehensible” (Kroeger 1983:467), and that, therefore, emphasizing the confidentiality of responses in these contexts “may not be necessary or even desirable” (Hershfield et al. 1993:244). There also are recorded cases showing extremely undesirable effects of confidentiality assurances (e.g., Back and Stycos 1959).

1022—–AMERICAN SOCIOLOGICAL REVIEW

ly leverage the fact that they knew the respondents into completed interviews, either through finding missing respondents or by neutralizing their unwillingness or suspicions. For example, one respondent claimed to be too injured to answer the questionnaire until the supervisor sent an interviewer who was a clan member, and who discovered no injury at all (Oyugis 1994:18). Another respondent told her interviewer, whose family she knew: “If it wasn’t you, the son of so and so, we would demand 50 bob for the interview” (Gwassi 1994:37; see also Kawadhgone 2000:115–116). In another case, the supervisor claimed that the respondents were always friendly to a particular interviewer because they all knew that he was “someone’s son” (Gwassi 1994:49). Moreover, when the questionnaire stealer, described earlier, was eventually persuaded to complete the questionnaire, he justified his change of heart with the statement that he knew both the family of the interviewer, saying “at least its not an outsider” (Kawadhgone 2000:125), and the supervisor’s father.

prescient observation: “One would think that women talking to women .|.|. would get more responsive respondents, but that doesn’t seem to be the case—is it because the interviewers are strange, i.e. not their confidantes?” RESPONSE ACCURACY. In addition to diminishing nonresponse, insider-interviewers’ familiarity with certain individuals or families—over and above the more general information about, say, local politics that all local interviewers had—also made them a fount of information about certain respondents. In particular, the realization that “they [knew] individual histories” (Oyugis 1994:24) meant that somewhat unusual stories reported in interviews could be cross-checked, or that when being interviewed by an insider, the respondents themselves had much less leeway to lie about things that were publicly known, even if they wanted to lie. I return to this point in the analysis. ANALYSIS

Field impressions and a review of relevant issues ITEM-RESPONSE PATTERNS. Two types of anec-by Ingenta Delivered to : in the extant sociological and data collection litUniversity of Jerusalem dotal evidence also suggested thatHebrew insider-interDec 2007 09:13:05 erature push us in the same direction. viewers were generating better data.Mon, First,31 there Specifically, in a setting such as rural Nyanza, were reports from the field. On a verbal autopthey provide very weak support for the classisy section, for example, a supervisor described cal position in which stranger-interviewers are respondents’ sour expressions in response to seen as outperforming their insider-interviewquestions about the death of a family member: er counterparts. On the contrary, both field “You are just somebody coming out of the blue, impressions and the general sociological literthey know you aren’t related” (Oyugis 2000:55). ature suggest that insider-interviewers generate The implicit message was that without a prior (1) higher response rates, (2) greater response social relationship, stranger-interviewers’ symreliability, and (3) greater response validity than pathies must be fake; they are not asking for the their stranger-interviewer counterparts. In addiright reasons. Similarly, in a focus group before tion, field impressions suggest that a given the initial survey wave, informants “said they interviewer’s insider versus stranger status vistrust the CBD workers [Community Based à-vis a given respondent is salient to both senDistribution agents used to distribute contrasitive and nonsensitive questions, suggesting ception, and in this area almost always from that (4) the scope of stranger-interviewer effects one’s own village], even the men .|.|. said they extends across an array of variables. talk freely because they’ve known them a long I now test these hypotheses empirically, time’ (Wakula South 1994:20). although I switch the order of 3 and 4. These Second, there were emerging patterns in the final two sections of the analysis also explore data. During the first wave of the survey, for the interaction between interviewers’ insiderness example, simple tabulations of the entered and sex, building on three discrete literatures. data—data entry was conducted during fieldFirst, there is a small but suggestive literature work in all three waves—showed that a higher on interviewer gender effects in other subproportion of women interviewed by women Saharan African settings (Becker et al. 1995; reported no network partners than women interBlanc and Croft 1992), particularly in the viewed by men. In response, the principal investigator’s field notes contain the following quite KDICP data (Bignami-van Assche et al. 2003).

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1023

Second, the fact that the sampled population is patrilineal, patrilocal, and patriarchal, with relatively fixed gender roles across a range of everyday tasks, implies that an interviewer’s sex would be more salient to a larger number of questions than other role-independent characteristics such as marriage, educational level, or fertility.9 Third, there is some anecdotal evidence from informal conversations with supervisors that respondents preferred talking to a male interviewer they knew than to an anonymous female interviewer (e.g., Wakula South 1994:20). I begin, however, by introducing relevant aspects of the KDICP data not yet described.

INTERVIEWERS’ SELECTION, CHARACTERISTICS, ASSIGNMENTS

AND

At all four sites and in all rounds of the survey, male and female interviewers were selected from a pool of Form-4 graduates, described earlier. Altogether, 97 interviewers (61 men and 36 women) were employed (in K3). In general, the mean age of the interviewers was 24 years. Approximately one third of the interviewers were married. One third had at least one child, and three fifths had worked as interviewers before (the vast majority in prior waves of this project). In addition, consistent with the norms of patrilocal exogamy in this population, none of the married women interviewers and only 23 percent of the unmarried ones were born in the THE DATA sampled sublocations. The effects of these social characteristics on responses in these data—“roleThree rounds of data collection were conductindependent” interviewer effects (Sudman and ed on the KDICP: in 1994–1995, 1996–1997, Bradburn 1974)—have been assessed elsewhere and 2000. The data used in the following analy(Bignami-van Assche, et al. 2003). sis are from the second and third waves, henceTo assess their level of acquaintance with a forth referred to as K2 and K3, because only in given respondent, interviewers were instructed these waves were data on interviewers’ characto answer the following question after comteristics collected. Delivered by Ingenta toeach : interview: pleting The sample design used a simple clusterHebrew University of Jerusalem sampling strategy. That is, four rural sublocaHow well do you know the respondent’s family? Mon, 31 Dec 2007 09:13:05 1. Not at all tions (administrative areas that include multiple 2. By name only villages) in South Nyanza were selected. Within 3. Quite well each of these sublocations, villages were ran4. Very well domly selected, and all ever-married women of reproductive age (younger than 50 years) known In the following analysis, stranger-interviewers to be de jure residents during the initial survey are those who claimed not to know the responwave (K1) were interviewed. The villages then dent’s family at all (Code 1), and insider-interwere resampled before K3 to introduce a new viewers are those who reported knowing the respondent panel. This yielded interviews with respondent’s family quite well or very well 924 women, who constitute the main sample for (Codes 3 and 4). These two are combined most of the analyses. The exception is a series because of the relatively small cell size. There of models that explore the effects of selectivialso is an intermediate category of “acquainty between K3 and the prior 1996–1997 round tance-interviewers,” which comprises those who (K2) using merged K2/K3 data files. In K2, claimed to know respondents families “by name 779 women were interviewed, 611 of whom only.”10 were successfully interviewed in K3, for a folTable 1 shows the distribution of completed low-up rate of 78 percent. K3 women’s interviews by level of interview-

9 Exploratory analyses showed this to be the case.

Parallel models using other observed interviewer characteristics, some of which are allegedly important in developing-country research (e.g., Ware [1977] recommends using married interviewers in fertility surveys) found much weaker effects (results available from the author).

10

These data may therefore underestimate the total stranger-interviewer effect because all interviewers were from the area, and thus fluent in local dialects and idioms, whereas normal practice in developing-country research, as noted, is to bring in professional stranger-interviewers from the outside, usually from urban areas.

1024—–AMERICAN SOCIOLOGICAL REVIEW

Table 1. Distribution of Female Respondents, by Type of Interviewer Interviewer Knows the Respondent N (Row %a) Stranger

Acquaintance

Insider

Totalb

Obisa Owich Kawadhgone Wakula South

115 (41.8) 091 (50.0) 095 (36.0) 059 (29.1)

100 (36.4) 064 (35.1) 121 (45.8) 063 (31.0)

060 (21.8) 027 (14.9) 048 (18.2) 081 (39.9)

275 182 264 203

Female Interviewer Male Interviewer

169 (41.4) 191 (37.0)

140 (34.3) 208 (40.3)

099 (24.3) 117 (22.7)

408 516

Total

360 (39.0)

348 (37.7)

216 (23.4)

924

a

Note: Row % may not sum to 100.0 because of rounding. b Totals refer to completed interviews.

er’s insiderness, and by interviewer’s sex and ly random with respect to interviewers’ insidersublocation. In all, nearly 40 percent of the ness. In particular, interviewer’s insiderness in K3 interviews were conducted by stranger-interis totally independent of the respondent’s viewers, and slightly more than 20 percent by observed socioeconomic and demographic charinsider-interviewers. The remainder fall into acteristics (educational attainment, wealth, the acquaintance category. Notwithstanding income, family size), as well as AIDS-related local norms of patrilocal exogamy, little of this attitudes and behavior and gender attitudes, as insider–stranger variation is associated with the measured in K2, even at a liberal 10 percent siginterviewer’s sex. There is, however,Delivered more sig-by Ingenta to :level. In fact, the only K2 variables nificance nificant variation by research site.Hebrew For example, Universityrelated of Jerusalem to K3 insiderness are four indicators of Mon, Dec 2007 09:13:05 the percentage of interviews conducted by 31 insidmobility and attitudes toward family planning.11 ers ranges from 14.9 percent in Owich to 39.9 percent in Wakula South. This distribution fits INSIDER- AND STRANGER-INTERVIEWERS’ with general characteristics of these areas EFFECTS ON RESPONSE RATES because Wakula South is on the small island in Lake Nyanza-Victoria. It makes more sense, Our impression that an insider-interviewer was therefore, that the social networks in which more likely than a stranger-interviewer to return these interviewer–respondent relations are from the field with a completed questionnaire embedded would tend to be different. after any of the up to three reported visits to a Interviewer training followed standard procegiven respondent is reflected in actual response dures. In addition to working through the nuts and rates. This was because of differences in both the bolts of the survey instrument over several days, overt and the numerically much more significant trainers repeatedly emphasized the importance of covert nonresponse. By overt nonresponse, I neutral responses and nondirective probing. The refer to an explicit refusal to participate in the surquestionnaire also included a confidentiality vey, amounting to 2.9 percent of all household assurance at the beginning of the interview, and three reminder assurances at later points. It should be noted that although these are not 11 The K3 interviewers were less likely to know experimental data, it is reasonable to have conrespondents who previously reported having been fidence in the magnitude of observed response to Nairobi (chi2[3df] = 10.6; p = 0.014), or being able variability. Not only have methodological studto speak Kiswahili, the main African language in ies frequently used nonexperimental data to Kenya (chi2[3df] = 15.7; p = 0.001). They also were explore broad patterns of response effects (e.g., more likely to know respondents who had previousBerk and Bernstein 1988; Schaeffer 1980; ly reported larger conversational networks (chi2[57df] Sudman et al. 1977; van Tilburg 1998), but post= 141.2; p < 0.001) and, related to this, respondents fieldwork selectivity checks in this case also who had talked to their husbands about family planconfirm that interviewer assignments were largening (chi2[3df] = 6.35; p = 0.012).

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1025

Table 2. Response Consistency Across Survey Waves, by Type of Interview Stranger

Acquaintance

Insider

Speaks Kiswahili Speaks English Has Ever Been to Nairobi Has Ever Used a Method of Family Planning

.79 .89 .86 .89

.86 .98 .89 .90

.89 1.00 .89 .94

Number of Children Ever Born Number of Deceased Children Size of Family-Planning Related Conversational Network Size of AIDS-Related Conversational Network

.74 .73 .50 .62

.74 .72 .42 .51

.85 .90 .65 .65

Mean Consistency Score

.75

.75

.85

Note: A consistent answer is defined in one of two ways. Either the respondent gave the same response in both waves. Else the respondent, in K2, (a) claims not to be able to speak Kiswahili or English, or to have never been to Nairobi or ever used contraception, but reports the opposite in K3; or (b) reports a lower (lifetime) number of children ever born, children deceased, or network size in K2 than K3.

visits for strangers and 1.3 percent for insiders. of responses across insider- and stranger-interviewers across two survey waves. This measure By covert nonresponse, I refer to respondents of test–retest reliability, also known as the “simclaiming to be too busy to be interviewed, or to ple response variance,” is an important compohaving a family member reporting that the nent of the mean square error of the estimated respondent is temporarily away. The covert nonmean (Bailar 1976) and a standard indicator of response rate hovered around 45.4 percent for valid measurement (Zeller and Carmines 1980). Delivered to : strangers and around 36.1 percent for insidersby Ingenta 12 The key problem in estimating this type of Hebrew University of Jerusalem across all three visits. Mon, 31 Dec 2007 09:13:05 test–retest reliability is that many characterisTogether, these differences in overt and covert tics, behaviors, and attitudes can legitimately nonresponses across levels of interviewers’insidchange between two waves, especially where, as erness meant that stranger-interviewers returned in this case, there is a lengthy gap between the with a completed questionnaire after 52 percent waves (Litwin 2003). To avoid this, I restricted of their visits to respondents’households, as comthe analysis to a small subset of questions about pared with 63 percent for insider-interviewers. relatively fixed characteristics, the only such questions asked in either wave. For each of INSIDER- AND STRANGER-INTERVIEWERS’ these questions, I defined a logically consistent EFFECTS ON RESPONSE RELIABILITY answer in one of two ways. Either the respondent gave the same answer in both waves, or she The collection of interviewer-related informareported a transition that might reasonably have tion in the K2 and K3 waves of the survey occurred. For example, between the two waves, allows us to explore the differential reliability she might well have visited Nairobi, learned more Kiswahili or English, used contraception for the first time, and so on. Or she might well 12 Low overt refusal rates are typical in develophave given birth to more children, had more ing-country surveys, most likely because projects children die, or developed a larger conversatypically require government permits and the blesstional network size for ever having talked to ings of all local headmen (as did the KDICP). This people about AIDS or family planning. In conmeans that only those with a history of run-ins with trast, logically inconsistent answers were the local headman or those who otherwise care little defined as those in which the woman dropped about being considered a “troublemaker” overtly her earlier admission that she had ever visited refuse. In contrast, covert refusal is a much more legitNairobi, could speak some Kiswahili or English, imate avoidance mechanism. People slip out the back had ever used contraception, and so on. door as interviewers approach the front, claim to be Table 2 presents the results from this someone else, say that they are about to go to a funertest–retest analysis, with the reliability expressed al, or the like.

1026—–AMERICAN SOCIOLOGICAL REVIEW

as the proportion of respondents who gave a consistent answer, as defined earlier, in both rounds. The comparison is made across the three categories of respondents (Columns 1 to 3): those who in both K2 and K3 were interviewed by stranger-interviewers, by acquaintance-interviewers, or by insider-interviewers. The sample sizes are too small to allow much leverage (only 20 individuals were interviewed by insiders in both rounds), but the trend is unambiguous. Across all eight questions, there is a clear and consistent pattern in which women interviewed during both rounds by insider-interviewers gave more reliable responses than those interviewed during both rounds by strangerinterviewers. The respective mean consistency across the eight questions is 0.85 for insiderinterviewers and 0.75 for stranger-interviewers. In contrast, those interviewed by acquaintance-interviewers appear no more nor less reliable than those interviewed by strangerinterviewers.

wealth, earnings, use of and attitudes toward family planning, family planning–related conversational networks, gender attitudes, AIDSrelated behavior and attitudes, and AIDS-related conversational networks. Among the 127 variables were both the main dependent and independent variables of the project and the core sociodemographic control variables. Standard estimation procedures were used: logit regression on dichotomous variables, tobit on continuous earnings data, and Ordinary Least Squares (OLS) on other continuous variables. Two types of models were estimated. The Model 1 group explored variation across three types of interviewers: strangers, acquaintances, and insiders. Stranger-interviewers, representing the standard type of interviewer in survey research, were defined as the reference group. The Model 2 group introduced an interaction term between each of the insider and stranger classifications and the interviewer’s sex, yielding six distinct types of interviewers: female stranger, female acquaintance, female insider, male stranger, male acquaintance, and male THE SCOPE OF INSIDER- AND STRANGERinsider.toHere, Delivered by Ingenta : female stranger-interviewers, the INTERVIEWER EFFECTS Hebrew Universitystandard of Jerusalem type of interviewer in demographic Mon, 31 Dec 2007 To assess the scope of stranger-interviewer and09:13:05 health surveys, were defined as the refereffects, a series of multivariate models were ence group. estimated in which an assortment of response Summary results are presented in Table 3 variables were regressed on stranger-interviewer (full results for the Model 2 group of estimates characteristics. Specifically, 127 variables were are available at Web site Appendix A). With selected from eight distinct sections of the quesrespect to the scope of the stranger-interviewer effects, the results are unambiguous. In the tionnaire: general characteristics, household

Table 3. Response Variation by Type of Interviewera Variable Category

Number of Variables

Model 1

Model 2

General Characteristics Household Wealth Individual Earnings Use of and Attitudes to Family Planning Family Planning Related Conversational Networks Gender Attitudes AIDS Related Behavior and Attitudes AIDS Related Conversational Networks

017 021 015 007 017 020 013 015

11 07 03 02 05 10 06 07

15 11 07 03 05 12 07 07

Total

127

51

67

Note: All models were run using logit regression on dichotomous variables, Tobit on reported earnings data, and OLS regression on other continuous variables. The ‘Model 1’ group explores response variation across 4 levels of insiderness; the ‘Model 2’ group explores variation across 6 categories, representing the interaction between 3 levels of insiderness and the interviewer’s sex. a Full model results available at ASR website (www2.asanet.org/asr/journals/2006/toc054.html).

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1027

Model 1 group, different responses are given to Becker et al. 1995; Blanc and Croft 1992; stranger- and insider-interviewers in 51 (40 Krysan et al. 1994).13 percent) of the variables, with significance levThe key to inferring response validity is thereels set at the 5 percent level. This includes fore to identify the direction of the social desirabout two thirds of the selected general charability bias for any given behavior or attitude, acteristics variables, notwithstanding the fact, and then to assess the relative proximity of any as discussed earlier, that an interviewer’s insidgroup’s aggregated responses to questions about erness in K3 is almost completely independent that behavior or attitude. To this end, 11 quesof a respondent’s observed socioeconomic and tions were selected from those used to condemographic characteristics, as measured in struct Table 3. Three are related to sources of K2. It also includes about half of the gender attiincome, five to gender attitudes, and three to tudes and AIDS-related variables, and about AIDS. Some of the questions are behavioral, one third of the wealth and family planning whereas others are attitudinal. Most important, variables. for all 11 questions, the trajectory of the social Allowing for the interaction with sex of the desirability bias can be identified with some cerinterviewer in Model 2 increases the observed tainty. scope of stranger-interviewer effects. Specifically, different responses are given to SOURCES OF SOCIAL DESIRABILITY. Daily stranger- and insider-interviewers in 67 (53 meetings with supervisors, who reported local percent) of the 127 selected variables, with attitudes toward the project and frequent conparticularly strong effects on reported general versations with elders, and others, who characteristics and gender attitudes. approached the research team for conversation In summary, there is a generalized strangerduring fieldwork clarified views of the research interviewer effect. Respondents tended to report project and helped define the factors that creDelivered by Ingenta to : different things to stranger- and insider-interated socially desirable responses to given quesHebrew University of Jerusalem viewers on about half of the selected variables. Mon, 31 Dec 2007 09:13:05 tions in the Nyanza setting. Here I list three of Moreover, whereas some of these were core these factors. variables of interest, and some could be thought The first and dominant factor was poverty, of as sensitive variables, many others were neispecifically, the lack of labor opportunities, ther core variables of interest nor overtly senlack of access to capital, general hunger for sitive variables. development, and respondents’ difficulty distinguishing the KDICP research goals from VALIDATING RESPONSES more development-oriented projects. These led respondents to tend to exaggerate their poverTo identify which of the six interviewer types ty as a strategy to trigger material assistance. I used in the Model 2 group of Table 3 appear to refer to it as the “poverty response motive.” generate the most accurate type of data, we must be able to validate survey responses. The absence of external data against which given 13 The external data that do exist cannot help here. responses can be compared means that we must do this inferentially: (1) by drawing on standard For example, the field notes on which I drew in the models of the response process that describe first part of the article can sometimes be matched with specific survey respondents, but tend to focus on how respondents evaluate a series of possible more general background characteristics of the indiresponses to any given question (Cannel, Miller, viduals than on their responses to individual quesand Oksenberg 1981; Tourangeau, Rips, and tions. Similarly, other surveys have been administered Rasinski 2000), and (2) by conforming to a in the area (e.g., the South Nyanza subsample of the standard assumption in the response effects lit1993, 1998, and 2003 Kenyan Demographic and erature, that is, that better methods of data colHealth Surveys), but the sampling frame was differlection generate higher reports about undesirable ent, and all interviewers were the equivalent of female phenomena or lower reports about desirable strangers, undermining comparisons between aggrephenomena (e.g., Aquilino 1994; Axinn 1991; gate response patterns.

1028—–AMERICAN SOCIOLOGICAL REVIEW

ed to their own infection or chances of infection. The second factor drew on public health proI refer to this as the “stigma response motive.” motions, ubiquitous in Kenya throughout the 1990s, including messages and programming on the radio, posters for clinics and schools, billI TEM -S PECIFIC D IRECTION OF S OCIAL boards, and other public advertisements. These DESIRABILITY BIAS. Given the strength of these messages often emphasized fidelity, the imporattitudes toward poverty, public health, and tance of open discussions about sexual practice HIV/AIDS, I assumed that these three deterbetween spouses, and the use of condoms.14 minants of response affected all KDICP responThey also built on approximately 20 years of dents irrespective of whether they were family planning campaigns and a more recent interviewed by an insider- or a stranger-interfocus on women-specific programs in general, viewer, or by a male or female interviewer. including efforts to change gender attitudes Thus, insider- and stranger-interviewer effects (Watkins 2000). The prevalence of these pubwould emerge as intermediate effects, mediatlic health discourses meant that respondents ing the extent to which the desire to benefit would tend to assume that attitudes promoted by would affect an actual response. I now describe modern public health campaigns should be northe expected trajectory of social desirability on mative, and therefore that foreign and/or urban each item as well as the expected mediating development workers wanted modern responseffects of insiders and strangers. es if they were to continue assisting local popI assumed that the expected social desirabilulations. I refer to this as the “modernity ity bias on sources of income questions would response motive.” be driven by the poverty response motive, The third factor, whose effects were very depressing reports about these sources (e.g., “No, I don’t get any income from selling things powerful but limited to a small subset of varifrom my shamba [agricultural plot], from a ables, was HIV/AIDS-related stigma. In the small business, or from membership in a credKDICP research setting, there wereDelivered numerousby Ingenta to : Hebrew University Jerusalem itofmerry-go-round”). However, because these signs of AIDS-related stigma. Frequently, refMon, 31 Dec 2007 09:13:05 questions address a type of information that is erences to AIDS used local code words (e.g., partly public knowledge—the types of business chira, a condition resulting from the breaking activity listed here can be observed directly, of traditional taboos). Respondents systematiand the cyclical beneficiaries of credit merrycally underreported family members’ mortaligo-rounds are likely to be known—I also ty from AIDS or HIV infection, even while they assumed that respondents interviewed by insidacknowledged its high prevalence in the comers would have less leeway to lie comfortably munity, or excessive mortality. Moreover, wives than those interviewed by strangers. As noted in known to be HIV+ frequently were kicked out the prior qualitative section, supervisors idenof their homes by their husband or in-laws and tified this quite early in fieldwork as an importold to return to their natal families. In short, the tant advantage of insider-interviewers (see also strength of this stigma suggested that responBarnes 1994). dents would tend to underreport anything relatGender attitude questions also seemed to be affected by the dominant poverty motive, although in this case it related to a poverty of 14 The intensity of AIDS-related social marketing perceived rights vis-à-vis men rather than finanwas particularly pertinent in Nyanza because the latcial resources. Thus, socially desirable responster has the highest HIV prevalence in Kenya (Office es would emphasize the relative lack of rights of the President et al. 2004). In a 1999 seroprevalence because that might trigger programs (and hopesurvey in neighboring Kisumu district, for example, fully an inflow of resources) to establish womenalso dominated by Luos, 30.1 percent of the sampled specific business initiatives, or “women’s adults had positive test results for HIV (Buvé, Caraël, groups” in general. Consistent with the popuHayes et al. 2001). At the research sites represented lar “women in development” paradigm in these data, adult mortality appears to have (Jacquette 1990), this was an important part of increased threefold during the 1990s (Doctor and development on the ground in Kenya during in Weinreb 2005), as reflected during data collection in the 1980s and 1990s, such that by 2000, there the constant cycle of postponed interviews attributwere 25,000 registered women’s organizations, able to respondents being “away at a funeral.”

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1029

Table 4. Frequency with which Observed Responses were Furthest from, or Closest to, the Expected Socially Desirable Response, by Type of Interviewer. Distance from Socially Desirable Response Furthesta

Closest

Female-strangers 00 1b Female-acquaintances 00 0b Female-insiders 11 0 Male-strangers 00 3 Male-acquaintances 00 2 Male-insiders 00 4b Notes: a Source for furthest and closest are models in Appendix A. b These three types of interviewers were joint closest to the expected socially desirable response on one of the eleven items.

almost all of which had been established in the RESULTS16 preceding two decades (Morel-Seytoux 2000). For expository ease, the predicted responses In addition, because Luo society is traditionalacross all six categories of interviewer’s insidly patriarchal, patrilineal, and patrilocal, emphaerness and sex are presented in two ways. First, sizing the relative lack of women’s rights also they are graphed in Figures 1a to 1c (tabulated 15 made cultural sense. Finally, as with the results with associated levels of statistical sigsources of income questions, I assumed that nificance are presented in Appendix A). Second, the extent to which a respondent would be able the overall frequency of each type of interto exaggerate her poverty of rights would be limviewer’s to the most and least socialited in interviews with insiders. Delivered by Ingenta toproximity : ly desirable responses is summarized in Table Hebrew University of Jerusalem The expected social desirability of respons4. In each case, it should be noted that these estiMon, 31 Dec 2007 09:13:05 es across the three AIDS questions appeared to mated coefficients are net of controls for other draw on a different combination of motives. I observed interviewer characteristics, the most assumed that responses to the first two quessignificant of which (despite varied effects tions, which asked how frequently the responacross variables) tended to be previous experident had conversations about AIDS with her ence working as an interviewer and marital stahusband and about her comfort suggesting that tus. he use condoms, drew primarily on the moderIn the absence of stranger-interviewer effects, nity response motive, pushing any given women we expect horizontal lines in Figures 1a to 1c toward answering “yes” to both questions. The either across all six interviewer categories or, other AIDS-related question addressed how where there is an interviewer gender effect, likely the respondent thought she was to become with a step between interviewer categories 1, 2, infected in the future. Here I assumed that the and 3 and categories 4, 5, and 6. That is not the stigma response motive would depress reportcase here. None of the response categories are ed risk, implying that higher reports of perceived risk would tend to be more accurate. horizontal across the different types of inter-

15

In theory, the modernity response motive may have pushed a relatively small group of respondents in the opposite direction, because by exaggerating their rights vis-à-vis men, they could signal their internalization of egalitarian discourses, thereby winning points from potential aid “investors.” On the aggregate level, this is not a convincing argument. It would have had to overcome both the dominant poverty response motive and culturally entrenched gender attitudes.

16

Steps taken to assess the possible effects of selectivity (e.g., the extent to which nonrandom interviewer assignment affected estimated response variability by interviewer’s insiderness) are discussed on Web site Appendix B. Models also were specified that explored variation in insider–stranger effects across the four research areas. Because there were very few differences between sites, and nothing systematic, the results presented in this section are “main effects” of interviewers’ insiderness.

1030—–AMERICAN SOCIOLOGICAL REVIEW

Delivered by Ingenta to : Hebrew University of Jerusalem Mon, 31 Dec 2007 09:13:05

Figure 1a–1c.—Predicted Response Values by Category of Questions and Type of Interviewer.2 Notes: 1 FS = female stranger; FA = female acquaintance; FI = female insider; MS = male stranger; MA = male acquaintance; MI = male insider. 2 Predicted values are net of other interviewer characteristics.

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1031

viewers. Moreover, as shown in Appendix A, and consistent with the high prevalence of effects noted in Table 3, in 9 of the 11 cases, the differences between female stranger-interviewers (the reference category in the regressions from which these values are derived) and either female or male insider-interviewers are statistically different, at the 5 percent level. More specific to the main aim of this section, which is to compare responses across interviewers’ levels of insiderness and their sex, three broad explanatory patterns can be identified in the predicted values.

(between female stranger and male stranger) in Appendix A. Rather, virtually all gender-ofinterviewer effects are found in the differential pattern of responses to insider-interviewers, and only in one case is there a difference across acquaintance-interviewers. This appears visually across the figures in the tendency of the male and female interviewer slopes to move in different directions, a suspicion largely confirmed in Table 4. The latter shows that in contrast to female insiders’ success in avoiding socially desirable responses, described earlier, a plurality of the most socially desirable responses—5 out of 11—were given to male insider interviewers.

MORE VALID DATA REPORTED TO FEMALE INSIDER-INTERVIEWERS. This is the dominant ABSENCE OF AN INSIDER PREFERENCE OR A pattern among the three broad explanatory patPREFERENCE FOR STRANGERS. The is the most terns.. As summarized in Table 4, for all 11 interpretively conservative explanatory pattern. variables, the responses to female insider-interThere is no statistically significant response viewers appear least prone to the social desirvariation of any type for 2 of the 11 selected ability bias. In seven of these, differences in variables, and none related to female insiderreports to female insiders and in the reference interviewers for 4 of the 11. Therefore, even if category are statistically significant. the general shape of the distribution mirrors It is also useful to note the scale of these statistically Delivered by Ingenta to : significant variation in similar varireporting differences. Among the respondents Hebrew University of Jerusalem ables, stranger-interviewers are frequently no interviewed by female insider-interviewers, for Mon, 31 has Dec 2007 09:13:05 better and no worse than insider-interviewers. example, 31 percent agree that a husband the right to leave a wife who fights him frequently. Among the respondents interviewed DISCUSSION by all other types of interviewers, 52 to 60 perThe communicative rituals from which these cent agree with this statement. response effects emerge, or not, as described in The responses to sources of income and the last few paragraphs, are the unknown in AIDS-related questions, presented in Figures 1a this analysis. Without data to explore the linand 1c, respectively, tell a similar story. The guistic and general semiotic architecture of the responses of women interviewed by female survey interaction at the most micro level, we insiders were furthest from the most socially have no way of knowing how what Maynard and desirable response for all six questions. Thus, Schaeffer (2002:4) call “situational oscillations” women interviewed by female insider-interand “analytic alternations” vary by insider- and viewers appear to have the most sources of stranger-interview across any of the variables in income, but also appear to be least comfortthese Kenyan data. But the product of those able suggesting condom use to their spouses, to oscillations and alternations is clear. At least in report talking with him about AIDS the least frethis setting, there appears to be a considerable quently, and to have the highest perceived risk insider–stranger response bias for many variof HIV infection. In a poor setting with high ables, and a systematic bias across specific HIV prevalence and repeated public calls for types of questions, especially if we examine marital fidelity, these differences are concepthe aggregated patterns of slope effects pretually important. sented in Figures 1a, 1b, and 1c, instead of restricting ourselves to independent items sigSHIFTING DIFFERENCES BETWEEN MALE AND nificant at the 5 percent level. FEMALE INSIDERS. None of the 11 variables More specifically, in accordance with the show a statistically significant gender-of-intersocial desirability criteria, the results suggest viewer effect across stranger-interviewers that data collected by female insider-interview-

1032—–AMERICAN SOCIOLOGICAL REVIEW

ers are superior for many questions and of equal have been shown to increase both the content quality for several. Conversely, as described and accuracy of responses (Dijkstra 1987; from the other side of the coin, data collected Schober and Conrad 1997; Suchman and Jordan by female strangers and, to a lesser degree, data 1990), and therefore have been welcomed, albeit collected by male strangers often is inferior to, with some residual suspicion, in the methodbut never better than, data collected by female ological mainstream (e.g., Beatty 1995; insiders. In addition, and more generally, data Schaeffer 1991; Tourangeau 1990). collected by insiders in general appear to be The second strategy, which has relied less more consistent across survey waves. Moreover, on formal methodological tests than on reviews insiders also are more successful at actually of field experiences, has largely been a by-prodfinding respondents and returning with a comuct of research in developing countries and has pleted questionnaire. In this setting, therefore, focused on building social relationships in a we can reasonably conclude that insider-interpreliminary interview (e.g., Axinn, Fricke, and viewers, especially female insiders, appear to be Thornton 1991) or on using interviewers from generally better data collectors than strangerthe same setting (e.g., Ross and Vaughan 1986). interviewers. Mapped onto the framework described in this The key question, of course, is whether we analysis, the effectiveness afforded by each of can generalize from this single Kenyan setting these methodological strategies derives from to other settings. In combination with the estabthe fact that they aim to more closely approxilished theoretical literature and other accounts mate interactional patterns among insiders than of fieldwork, I think it is reasonable to do so, among strangers, an approach already modeled although conservatively. For example, I would by a caste of qualitative interviewers (Jansen expect that parallel effects could be identified 1980; Holstein and Gubrium 1995; Mayoni in other areas of the developing world in which 1983). The possibility remains, however, that the there are similar structural characteristics to marginal Delivered by Ingenta to :gains from using an interpersonal those found in South Nyanza. These characterinterviewing Hebrew University of Jerusalemstyle will be limited in certain istics would include a general mistrust of 2007 Mon, 31 Dec 09:13:05 research settings. Training interviewers to strangers and groups associated with the govassume the stance of a friend by being simulernment. To the extent that this is true, it sugtaneously professional and informal will not be gests that the results presented here may also enough. It will not negate the respondent’s susinform data collection practices in developed picion of strangers and, consequently, facilitate countries with relatively large immigrant popthe collection of higher quality data. Only the ulations or other minorities in which, for a num“real thing,” an insider known by the respondent, ber of reasons, there is continued suspicion of or one whose family is known by the respondent, outsiders. For until or unless members of these will be able to do this. groups develop the same tolerance for pesky strangers, and more specifically for the strangerCONCLUSION interviewers who, albeit politely, demand to As I have argued, the fact that we almost always know personal things, response patterns may be employ stranger-interviewers to collect data affected by insider–stranger concerns of the does not mean that we should do so uncriticaltype illustrated in this analysis. ly. In certain areas of the world, insider-interThis argument is in line with trends in recent viewers may collect more data, as well as more methodological literature. In fact, it may be reliable and accurate data, on a surprisingly possible to coopt some parts of that literature large range of topics. Foundational social theointo a more general insider–stranger paradigm. ry suggests that this relative success of insiderFor example, recent research on ways to facilinterviewers over stranger-interviewer itate emotional comfort between interviewers counterparts is related to variation in the visiand respondents have tended to emphasize one bility of and tolerance for strangers, and that the of two strategies, both recommended in some very presence of strangers asking questions form by Cicourel (1974) three decades ago. indexes a certain level of social differentiation. The first, focused on research in developed This begs full empirical verification. However, country settings, has emphasized the development of interpersonal interviewing styles. These this single Kenyan case, in which I have pre-

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1033

sented baseline estimates of insider- and sides of the debate. The response distribution on stranger-interviewer effects, is consistent with most of the variables suggests that, at least in this the underlying hypothesis. setting, women respondents do not favor female Important conceptual gaps remain. It is usestranger-interviewers over male interviewers, ful to at least mention a few of them. First, as but that they do favor female insider-internoted earlier, recent research suggests that comviewers. Thus, there are no gender-of-intermunication between interviewers and responviewer effects per se in this setting. Rather, dents alternates between formal rule following these function through insider–stranger effects. and tacit practices (Schaeffer and Presser 2003). Again, further research is necessary to explore To the extent that this is the case, it seems likethe extent to which this type of interaction in ly that these alternations are somewhat differinterviewer effects exists in other research setent where interviewers are strangers as opposed tings. At the very least, however, it raises a more to insiders. If this represents the microsocial general question about the extent to which roleorigins of these response effects, it would be useindependent interviewer effects, long observed ful to observe the enactment of these alternain social sciences, sometimes index other unobtions, seeing precisely where they differ across served characteristics. these two types of interviewers. Fourth, we also can expect stranger-interSecond, the 11 questions on which the most viewer effects to vary across time in any given detailed analysis was undertaken deal with relsetting because one of the most important strucatively sensitive or threatening topics: sources tural determinants of openness to strangers in of income, women’s autonomy, and AIDS-relatgeneral may be the level of social density and ed questions. But it is important to recall the diversity (Douglas 1970; Durkheim [1926] generalized insider–stranger effects noted in 1964), which is not fixed at either the general Table 3: statistically significant variation in macro level of community norms or the indiresponse distributions across levels Delivered of insider-by Ingenta vidual to level. : Indeed, we can expect increasing ness for 67 of 127 variables. InHebrew other words, interactions University of Jerusalemwith strangers in general and there is reason to believe that insider–stranger increasing Mon, 31 Dec 2007 09:13:05familiarity with stranger-intervieweffects equally influence both sensitive and noners and their scripts in particular to reduce the sensitive questions, or at least what we think are scope of stranger-interviewer effects, perhaps sensitive or nonsensitive questions in given coneven reversing them in those conversational texts. This is in contrast to interviewer effects domains in which people become accustomed more generally, which tend to be limited to sento interacting with other types of professional sitive variables in relation to which interviewstrangers. On the other hand, the increasing ers’ social identity is the most “salient” (e.g., realization within the response effects literaSchaeffer and Presser 2003; Sudman and ture that interviewer–respondent interaction Bradburn 1974). draws heavily on normal types of interactional Third, another question raised by the analyrules (e.g., Maynard and Schaeffer 2002; sis relates to the differential quality of data colSchaeffer and Presser 2003) suggests that intilected by female and male insiders. Although the mate insiders, in this case female insiders, may matching of interviewers and respondents always have a marginal advantage over strangers according to sex has been criticized in develin certain conversational domains, no matter oping-country research (Becker et al. 1995; how skilled the latter. Blanc and Croft 1992), it remains standard pracFinally, and this may be the most important tice for survey research in general, and for point in terms of the practical implications of developing countries in particular, mainly this research, failure to address the potential because of the widely held belief in what Warren for insider–stranger effects has analytic impli(1988) calls the “focal gender myth of field cations where such effects are significant and research,” that is, the belief that women interunobserved. Besides the effects on response viewers are less threatening and have better rapreliability described earlier, two such effects port, and that this leads to higher quality data, are likely. The first includes the standard effects especially in interviews with women (e.g., of response bias: undermining of univariate staOppenheim 1992:84). The presented analysis tistics accuracy (and hence comparability) and supports a middle position between the two distorting of analytic relationships (e.g., Blalock

1034—–AMERICAN SOCIOLOGICAL REVIEW

potentially serious analytic problem. At the very 1961; de Leeuw, Mellenbergh, and Hox 1996). Indeed, the potential for such implications is perleast, it suggests that researchers need to incorhaps all the greater in this case because strangerporate ways to measure the extent of the poteninterviewer effects appear to offer particularly tial insider–stranger bias in their research design, fertile ground for response bias. In these data, especially where they are conducting research for example, the respondents who emerge from among populations that are structurally more interviews with female insider-interviewers susceptible to such effects. Ultimately, it is appear different from those who emerge from hoped that careful comparative studies will interviews with female stranger-interviewers. allow researchers to identify such settings The former have fewer traditional gender attiempirically, and to identify how shifting structudes and more sources of income, but their tural boundaries, such as fluctuations in the levresponses to AIDS-related questions are less els of social density and diversity, affect the likely to match the messages of public health generation of these biases. Until then, our sole campaigns. In contrast, the latter (women interguide is theory and, perhaps less reliable, variviewed by strangers) appear more downtrodden able researcher instincts and experience. This (by their husbands) and poorer, but more in article suggests that these two sources lend tune with AIDS-related messages. The differthemselves to somewhat greater skepticism ence between these two indicates two different about the efficacy of stranger-interviewers than types of populations, each of which blends modhas been acknowledged to date, certainly by ern and traditional characteristics in a very difpractitioners of survey research. But it also sugferent way, and each of which, in preliminary gests that just as there are limits to what we can analysis using K2 data only, appears to generexpect of stranger-interviewers, there also are, ate different multivariate estimates (Weinreb at least in some settings, alternatives. 2000). However, and this is the important point, each of these populations appears to be a simDelivered by Ingenta to A.Weinreb : Alexander is a lecturer in the Department ple product of different types of interviewer Hebrew UniversityofofSociology Jerusalem and Anthropology, Hebrew University. stimuli. This in turn points to the most insidiMon, 31 Dec 2007 09:13:05 A social demographer by training (Ph.D. from the ous effects of insider–stranger biases. On the University of Pennsylvania; doctoral committee constructural level, foundational social theory stituted by Susan Watkins, Herbert Smith, and Randall implies that tolerance for stranger-interviewCollins), his primary substantive research focuses on ers is related to tolerance for other types of family and community networks and AIDS, and his behavioral innovations from the outside. This primary geographic focus is East Africa. Over sevmakes the choice of insider- or stranger-intereral years, Weinreb was Field Director of two NIHviewers endogenous to subsequent analyses of funded longitudinal research projects in Kenya and behavioral innovation. Given that the identifiMalawi. He currently teaches courses on the sociolcation of such causal mechanisms is the explicogy of Africa, the sociology and demography of AIDS, it aim of much social science research, this is a data collection, and microsociological theory.

.40 .34 .39

.44 .40 .60 .86 .64

.48 .40 .27

Delivered by Ingenta to : Hebrew University of Jerusalem Mon, 31 Dec 2007 09:13:05 .34 .26 .48

.46 .44 .60 .88 .62

.47 .39 .30

Female acquaintance

Note: Significantly different from predicted value for “female stranger” at: * p = .10; ** p = .05; *** p = .01.

Source of Income in the Last Month —Respondent claims that she: ——01. Sold things grown on her shamba (agricultural land) ——02. Received money from a small business ——03. Received money from a credit merry-go-round Gender Attitudes —Agrees that a husband has the right to leave his wife if: ——04. She disrespects his family ——05. She neglects household chores ——06. She fights him frequently ——07. She is sexually unfaithful ——08. She drinks too much AIDS —Respondent claims to: ——09. Talk to her husband about AIDS frequently ——10. Feel comfortable suggesting the use of condoms to her husband ——11. Be at moderate or great risk of catching AIDS

Model # and Dependent Variable

(Reference) Female stranger

.19* .11** .66*

.29 .24 .31** .53* .38

.60* .64*** .52

Female insider

.34 .31 .33

.50 .38 .52 .90 .68

.44 .42 .20

Male stranger

Type of Interviewer

Table A1.—Predicted Value of Selected Dependent Variables by Type of Interviewer, Net of Effect of Other Interviewer Characteristics

APPENDIX

.41 .27 .25**

.52 .46 .58 .87 .64

.45 .38 .25

Male acquaintance

.56** .31 .44

.62** .27* .60 .86 .62

.36** .32 .21

Male insider

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1035

1036—–AMERICAN SOCIOLOGICAL REVIEW

REFERENCES

Effects, Question Reliability and Sample Attrition.” Demographic Research SC1:31–75. Agar, Michael. 1980. The Professional Stranger: An Blalock, Hubert M. 1961. Causal Inferences in Informal Introduction to Ethnography. New York: Nonexperimental Research. Chapel Hill: Academic Press. University of North Carolina Press. Anderson, Myrdene. 1986. “Cultural Concatenation Blanc, Ann K. and Trevor N. Croft. 1992. “The Effect of Deceit and Secrecy.” Pp. 323–8 in Deception: of the Sex of Interviewer on Responses in Fertility Perspectives on Human and Nonhuman Deceit, Surveys: The Case of Ghana.” Presented at the edited by Robert W. Mitchell and Nicholas S. annual meeting of the Population Association of Thompson. New York: State University of New America, Apri1 30–May 1, Denver, CO. York Press. Breault, K.D., Chris Barker and Russell Lemle. 1987. Aquilino, William S. 1994. “Interview Mode Effects “Informal Helping in Partner and Stranger Dyads.” in Surveys of Drug and Alcohol Use.” Public Journal of Marriage and the Family 49(3):541-47. Opinion Quarterly 58:210–40. Bulmer, Martin. 1993. “Interviewing and Field Axinn, William G. 1991. “The Influence of Organization.” Pp. 205–17 in Social Research in Interviewer Sex on Responses to Sensitive Developing Countries: Surveys and Censuses in Questions in Nepal.” Social Science Research the Third World. 2d ed., edited by Martin Bulmer 20:303–18. and Donald P. Warwick. London, England: UCL Axinn, William G., Thomas E. Fricke, and Arland Press. Thornton. 1991. “The Microdemographic Buve, A., M. Carael, R.J. Hayes, et al. 2001a. Community-Study Approach: Improving Survey “Multicentre Study on Factors Determining Data by Integrating the Ethnographic Method.” Differences in Rate of Spread of HIV in subSociological Methods and Research 20:187–217. Saharan Africa: Methods and Prevalence of HIV Back, Kurt W. and J. Mayone Stycos. 1959. The Infection.” AIDS 15, (Suppl 4): S5–S14. Survey Under Unusual Conditions: Cannell, Charles F., Peter V. Miller, and Lois Methodological Facets of the Jamaica Human Oksenberg. 1981. “Research on Interviewing Fertility Investigation. Ithaca, NY: Society for Techniques.” Sociological Methodology Applied Anthropology, Cornell University. 12:389–437. Delivered by Ingenta to : Bailar, Barbara A. 1976. “Some Sources of University Error Catania, Joseph A., Diane Binson, Jesse Canchola, Hebrew of Jerusalem and Their Effect on Census Mon, Statistics.” M. Pollack, Walter Hauck, and Thomas J. 31 Dec 2007Lance 09:13:05 Demography 13(2):273-85. Coates. 1996. “Effects of Interviewer Gender Barley, Nigel. 1983. The Innocent Anthropologist: Interviewer Choice, and Item Wording on Notes From a Mud Hut. London, England: Penguin Responses to Questions Concerning Sexual Books. Behavior.” Public Opinion Quarterly 60:345–75. Barnes, John A. 1994. A Pack of Lies: Towards a Choldin, Harvey M., A. Majeed Kahn, and B. Hosne Sociology of Lying. Cambridge, England: Ara. 1967. “Cultural Complications in Fertility Cambridge University Press. Interviewing.” Demography 4:244–52. Beatty, Paul. 1995. “Understanding the Christenfeld, Nicholas, William Gerin, Wolfgang Standardized/Non-Standardized Interviewing Linden, Mara Sanders, Jennifer Mathur, James D. Controversy.” Journal of Official Statistics Deich, and Thomas G. Pickering. 1997. “Social 11:147–60. Support Effects on Cardiovascular Reactivity: Is Becker, Stan, Kale Feyistan, and Paulina Makinwaa Stranger as Effective as a Friend?” Adebusoye. 1995. “The Effect of Sex of Psychosomatic Medicine 59:388–98. Interviewers on the Quality of Data in a Nigerian Cicourel, Aaron V. 1974. Theory and Method in a Family Planning Questionnaire.” Studies in Family Study of Argentine Fertility. New York: John Wiley Planning 26:233–40. and Sons. Belcher, D. W., F. K. Wurapa, I. M. Lourie, K. Collins, Randall. 2004. Interaction Ritual Chains. Kwabia, and S. K. Avle. 1976. Experience in Princeton, NJ: Princeton University Press. Selecting, Training and Supervising Interviewers de Leeuw, Edith, Gideon J. Mellenbergh, and Joop in a Rural Health Project: Danfa Project, Ghana, J. Hox. 1996. “The Influence of Data Collection Danfa Project Monograph Series, No. 11. Los Method on Structural Models: A Comparison of Angeles, CA: UCLA School of Public Health. a Mail, a Telephone, and a Face-to-Face Survey.” Berk, Marc L. and Amy B. Bernstein. 1988. Sociological Methods and Research 24:443–72. “Interviewer Characteristics and Performance on Demographic and Health Surveys (DHS). 1987. a Complex Health Survey.” Social Science Interviewer’s Manual: For Use With Model “A” Research 17:239–51. Questionnaire For High Contraceptive Prevalence Bignami-van Assche, Simona, Georges Reniers, and Countries (Basic Documentation–5). Columbia, Alexander A. Weinreb. 2003. “An Assessment of MD: Institute for Resource Development. the KDICP and MDICP Data Quality: Interviewer Dentan, Robert Knox. 1968. The Semai: A Nonviolent

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1037

and Donald P. Warwick. London, England: UCL People of Malaya. New York: Holt, Rinehart, and Press. Winston. Holstein, James A. and Jaber F. Gubrium. 1995. The Dijkstra, Wil. 1987. “Interviewing Style and Active Interview. Thousand Oaks, CA: Sage Respondent Behavior: An Experimental Study of Publications. the Survey Interviewer.” Sociological Methods Hyman, Herbert H., William J. Cobb, Jacob J. and Research 16:309–34. Feldman, Clyde W. Hart, and Charles Herbert Dindia, Kathryn, Mary Anne Fitzpatrick, and David Stember. 1954. Interviewing in Social Research. A. Kenny. 1997. “Self-Disclosure in Spouse and Chicago, IL: University of Chicago Press. Stranger Interaction: A Social Relations Analysis.” Jacquette, J. 1990. “Gender and Justice in Economic Human Communications Research 23:388–412. Development.” Pp. 54-69 in Persistent Inequalities, Doctor, Henry V. and A. A. Weinreb. 2005. “Mortality edited by Irene Tinker. Oxford, England: Oxford Among Married Men in Rural Kenya and Malawi.” University Press. Etude de la Population Africaine 20:165–77. Jansen, Sue Curry. 1980. “The Stranger as Seer or Douglas, Mary. 1970. Natural Symbols: Explorations Voyeur: A Dilemma of the Peep-Show Theory of in Cosmology. New York: Pantheon Books. Knowledge.” Qualitative Sociology 2:22–55. Durkheim, Emile. [1926] 1964. The Division of Kawadhgone. 2000. Unpublished KDICP field notes Labor in Society. New York: The Free Press. by Susan Watkins (KDICPJournal-HH3-K3.pdf). Fowler, Floyd J. and Thomas W. Mangione. 1990. Available at http://kenya.pop.upenn.edu/ Standardized Survey Interviewing: Minimizing level%203/Kenya/level3_kenya_fieldnotes. Interviewer-Related Error. Newbury Park, CA: Kearl, B., ed. 1976. Field Data Collection in the Sage Publications. Social Sciences: Experiences in Africa and the Garf inkel, Harold. 1967. Studies in Middle East. New York: Agricultural Development Ethnomethodology. Englewood Cliffs, NJ: Council, Inc. Prentice-Hall. Kroeger, Axel. 1983. “Health Interview Surveys in Giddens, Anthony. 1990. The Consequences of Developing Countries: A Review of the Methods Modernity. Stanford, CA: Stanford University and Results.” International Journal of Press. Epidemiology 12:465–81. Delivered by Ingenta to : Givens, David. 1978. “Greeting a Stranger: Some Krysan, Maria, Howard Schuman, Lesli Jo Scott, University of Jerusalem Commonly Used NonverbalHebrew Signals of Paul Beatty. 1994. “Response Rates and Mon, 31 Dec 2007and 09:13:05 Aversiveness.” Semiotica 22:351–67. Response Content in Mail Versus Face-to-Face Goffman, Erving. 1971. Relations in Public: Surveys.” Public Opinion Quarterly 58:381–99. Microstudies of the Public Order. New York: Basic Levy, Marion J., Jr. 1972. Modernization: Latecomers Books. and Survivors. New York: Basic Books. Gwassi. 1994. Unpublished KDICP field notes by Litwin, M. S. 2003. How to Assess and Interpret Susan Watkins (KDICPJournal-HH1-A-DecSurvey Psychometrics: The Survey Kit. 2d ed. 1994.pdf). Available at http://kenya.pop.upenn.edu/ London, England: Sage Publications. level%203/Kenya/level3_kenya_fieldnotes. Mangione, Thomas, Ralph Hingson, and Jane Barrett. Gwassi. 1995. Unpublished KDICP field notes by 1982. “Collecting Sensitive Data: A Comparison Susan Watkins (KDICPJournal-HH1-B-Janof Three Survey Strategies.” Sociological Methods 1995.pdf). Available at http://kenya.pop.upenn.edu/ and Research 10:337–46. level%203/Kenya/level3_kenya_fieldnotes. Mauldin, W. P. 1965. “Application of Survey Gwassi. 2000. Unpublished KDICP field notes by Techniques to Fertility Studies.” Pp. 93–118 in Susan Watkins (KDICPJournal-HH3-K3.pdf). Public Health and Population Change: Current Available at http://kenya.pop.upenn.edu/ Research Issues, edited by Mindel C. Sheps and level%203/Kenya/level3_kenya_fieldnotes. Jeanne C. Ridley. Pittsburgh, PA: University of Hagenaars, Jacques A. and Ton G. Heinen. 1982. Pittsburgh Press. “Effects of Role-Independent Interviewer Maynard, Douglas W. and Nora Cate Schaeffer. 2002. Characteristics on Responses.” Pp. 91–130 in “Standardization and Its Discontents.” Pp. 3–46 in Response Behavior in the Survey-Interview, editStandardization and Tacit Knowledge: Interaction ed by W. Dijkstra and J. van der Zouwen. London, and Practice in the Survey Interview, edited by England: Academic Press. Douglas W. Maynard, Hanneke HoutkoopHeisel, Donald F. 1968. “Attitudes and Practice of Steenstra, Nora Cate Schaeffer, and Johannes van Contraception in Kenya.” Demography 5:632–41. der Zouwen. New York: John Wiley and Sons. Hershfield, Allan F., Niels G. Rohling, Graham B. Mayoni, Joseph R. 1983. “Eager Visitor, Reluctant Kerr., and Gerald Hursch César. 1993. “Fieldwork Host: The Anthropologist as Stranger.” in Rural Areas.” Pp. 241–52 in Social Research in Anthropologica 25:221–49. Developing Countries: Surveys and Censuses in Mensch, B. S., P. C. Hewett, and A. Erulkar. 2003. the Third World. 2d ed., edited by Martin Bulmer “The Reporting of Sensitive Behavior Among

1038—–AMERICAN SOCIOLOGICAL REVIEW

Adolescents: A Methodological Experiment in in Human Relations.” Pp. 307–16 in The Sociology Kenya.” Demography 40: 247–68. of Georg Simmel, translated and edited by Kurt H. Molm, L. D., N. Takahashi, and G. Peterson. 2000. Wolf. New York: Free Press. “Risk and Trust in Social Exchange: An Singer, Eleanor, Dawn R. Von Thurn, and Esther R. Experimental Test of a Classical Proposition.” Miller. 1995. “Confidentiality Assurances and American Journal of Sociology 105:1396–427. Response: A Quantitative Review of the Morel-Seytoux, Sylvie. 2000. Review of Gender Experimental Literature.” Public Opinion Issues in the USAID/Kenya Integrated Strategic Quarterly 59:66–77. Plan (ISP) 2001–2005. Washington, DC: Women Stone, Linda and J. Gabriel Campbell. 1984. “The in Development Technical Assistance Project. Use and Misuse of Surveys in International Oppenheim, A. N. 1992. Questionnaire Design, Development: An Experiment from Nepal.” Interviewing and Attitude Measurement. London, Human Organization 43:27–37. England: Pinter Publishers. Stycos, J. M. 1960. “Sample Surveys for Social Oyugis. 1994. Unpublished KDICP field notes by Science in Underdeveloped Areas.” Pp. 375–88 in Susan Watkins (KDICPJournal-HH1-A-DecHuman Organization Research, edited by R. N. 1994.pdf). Available at http://kenya.pop. Adams and J. Preiss. Homewood, IL: The Dorsey upenn.edu/level%203/Kenya/level3_kenya_ Press. fieldnotes. Suchman, Lucy and Brigitte Jordan. 1990. Oyugis. 2000. Unpublished KDICP field notes by “Interactional Troubles in Face-to-Face Survey Susan Watkins (KDICPJournal-HH3-K3.pdf). Interviews.” Journal of American Statistical Available at http://kenya.pop.upenn.edu/ Association 85:232–53. level%203/Kenya/level3_kenya_fieldnotes. Sudman, Seymour and Norman M. Bradburn. 1974. Plummer, Mary L., Daniel Wight, David A. Ross, Response Effects in Surveys: A Review and Rebecca Balira, Alessandra Anemona, Jim Todd, Synthesis. Chicago, IL: Aldine Publishing Zachayo Salamba, Angela I. N. Obasi, Heiner Company. Grosskurth, John Changalunga and Richard J. Sudman, Seymour, Norman M. Bradburn, Ed Blair, Hayes. 2004. “Asking Semi-Literate Adolescents and Carol Stocking. 1977. “Modest Expectations: About Sexual Behaviour: The ValidityDelivered of Assistedby Ingenta The Effects to : of Interviewers’ Prior Expectations Self-Completion Questionnaire (ASCQ) in Responses.” Sociological Methods and HebrewData University ofonJerusalem Rural Tanzania.” Tropical Medicine Mon, 31and Dec 2007Research 09:13:056:171–82. International Health 9:737–54. Taietz, Philip. 1962. “Conflicting Group Norms and Ross, David A. and J. Patrick Vaughan. 1986. “Health the “Third” Person in the Interview.” American Interview Surveys in Developing Countries: A Journal of Sociology 68:97–104. Methodological Review.” Studies in Family Tönnies, Ferdinand. 1961. Custom: An Essay on Planning 17:78–94. Social Codes. Chicago, IL: Henry Regnery Rudolph, L. and S. H. Rudolph. 1958. “Surveys in Company. India: Field Experience in Madras State.” Public Tourangeau, Roger. 1990. “Comment on Suchman Opinion Quarterly 33:235–44. and Jordan.” Journal of American Statistical Schaeffer, Nora C. 1980. “Evaluating Race-ofAssociation 85:250–51. Interviewer Effects in a National Survey.” Tourangeau, Roger, Lance J. Rips, and Kenneth Sociological Methods and Research 8:400–19. Rasinski. 2000. The Psychology of Survey ———. 1991. “Conversation With a Purpose—or Response. Cambridge, England, Cambridge Conversation? Interaction in the Standardized University Press. Interview.” Pp. 367–91 in Measurement Errors in United Nations, Department of Economic and Social Surveys, edited by Paul B. Biemer, Robert M. Affairs. 1961. The Mysore Population Study. New Groves, Lars E. Lyberg, Nancy A. Mathiowetz, York: United Nations. Seymour Sudman, and Paul B. Biemer. New York: van Tilberg, Theo. 1998. “Interviewer Effects in the Wiley and Sons. Measurement of Personal Network Size: A Schaeffer, Nora C. and Stanley Presser. 2003. “The Nonexperimental Study.” Sociological Methods Science of Asking Questions.” Annual Review of and Research 26(3):300-328. Sociology 29:65–88. Wakula South. 1994. Unpublished KDICP field notes Schober, Michael and Frederick G. Conrad. 1997. by Susan Watkins (KDICPJournal-HH1-B“Does Conversational Interviewing Reduce Survey Jan1995.pdf). Available at http://kenya. Measurement Error.” Public Opinion Quarterly pop.upenn.edu/level%203/Kenya/level3_kenya_ 61:576–602. fieldnotes. Simmel, Georg. 1950a. “The Stranger.” Pp. 402–8 in Wakula South. 1996. Unpublished KDICP field notes The Sociology of Georg Simmel, translated and by Susan Watkins (KJournal-HH2-Dec96.pdf) edited by Kurt H. Wolf. New York: Free Press. Available from author. ———. 1950b. “Knowledge, Truth, and Falsehood Ware, Helen. 1977. Language Problems in

THE LIMITATIONS OF STRANGER-INTERVIEWERS IN RURAL KENYA—–1039

Demographic Field Work in Africa: The Case of the Cameroon Fertility Survey. London, England: International Statistical Institute & World Fertility Survey, Scientific Report No. 2. Warren, Carol A. B. 1988. Gender Issues in Field Research. Qualitative Research Methods Series. Thousand Oaks, CA: Sage Publications. Watkins, Susan C. 2000. “Local and Foreign Models of Reproduction in Nyanza Province, Kenya, 1930–98.” Population and Development Review 26:725–59. Weinreb, Alexander A. 2000. Integrating Respondents, the Community, and the State in the Analysis of Contraceptive Use. Ph.D. dissertation, Population Studies Center, University of Pennsylvania, Philadelphia, PA.

———. 2001a. “A State-Centered, Political Capital Approach to the Explanation of Demographic Differences: With Special Reference to Regional/Ethnic Inequality in Sub-Saharan Africa.” In Proceedings of the General Conference of the International Union for the Scientific Study of Population (IUSSP), August 18–24, Salvador, Brazil. ———. 2001b. “First Politics, Then Culture: Accounting for Ethnic Differences in Demographic Behavior in Kenya.” Population and Development Review 27:437–67. Zeller, Richard A. and Edward G. Carmines. 1980. Measurement in the Social Sciences: The Link Between Theory and Data. New York: Cambridge University Press.

Delivered by Ingenta to : Hebrew University of Jerusalem Mon, 31 Dec 2007 09:13:05