The empirical groundedness of the "social capital" concept: the case of disoriented answers

Respondents often answer questions other than asked. In such instances, accounts are stimulated by questions but are not oriented to the tasks posed by the questions. This fact is rather unknown to positivistically-minded sociologists and tends to be overlooked by those trained in qualitative methods. Actually, how the latter deal with this problem cannot be said because their work with answers is rarely shown and is often reduced to gathering illustrations for ideas they seem to support. Although qualitative methodologies claim holistic approach, the attention sociologists pay to collected data is selective. While some accounts are taken into consideration, others are omitted as bearing no relevant evidence or as inadequate to the question. While much thought and paper is devoted to the art of asking questions, little is given to the problem of understanding answers. «Disoriented answers» occur when a respondent produces an answer that, on its face value, is not among the expected. Disoriented answers can be discounted by reference to contextual effects or discarded altogether. Still, this is a persistent fact that begs proper explanation. This paper proposes to assume that any answers are relevant as they exhibit respondents’ understanding of questions and manifest their reference to some models that are broader than the supposed semantic scope of the question posed. The problem of disoriented answers is a challenge to qualitative and quantitative methodologies alike. The argument of this paper proceeds as follows: (1) «social capital», while being generally referred to effects of social networks, lacks any clearly defined referent, or «phenomenon», (2) 2 cases of sponta- neously achieved accounts of social network’ effects are contrasted to failed attempts directed at discovery of such accounts during an inquiry into the possible contexts of the social capital manifestations, (3) this contrast is explained by the fact that questions inquired into different contexts, that of ordinary production activities and that of innovative activities with the failed elicitation in the former context and successful in the latter, (4) question-answer pairs in the failed elicitation are examined, and it is shown that disoriented answers therein exhibit narrative structures, (5) these structures arise as respondents’ attempt to accommodate the model of the question task within their model of their own workplace situations, (6) hence the failure to elicit references to social capital manifestations is due not to the deficiency of our tool but to the actual absence of social capital in contexts other than innovations. The capacity of the instrument not to find a phenomenon where there is none, is an advantage as compared to the survey instruments that ask suggestive questions and assume that respondents share the researcher’s knowledge of the phenomenon under study, that they match it to their actual in situ phenomena, and that they (dis)confirm the match. On the contrary, we assume that concepts have various lay meanings and usages and that respondents can negotiate these meanings and apply concepts of questionnaire/guide items to practices other than those sought by the researcher. They do this not to flatter researchers’ suggestions but in order to interpret questions about supposed experiences in terms of actual experiences. While this accommodation is purely conceptual, researchers might be led to treat it as empirical. The guide we employed did not use suggestive questions (of the «tell me about trust» type), but inquired into possible contexts of social capital (the «tell me about your work» type). The answers were expected to produce spontaneous accounts of «social capital» manifestations taken broadly as networks effects. The analysis proposed here did not rely on the usage of concepts; instead, what it did rely on was the discovery of narrative structures as products of respondents’ conceptualizations of their workplace experiences.