Still Going: Recent Debates on the Goldschmidt Hypothesis 265Looking Backward
Barnes and Blevins (1992) neglect much of what has been accomplished in the
Goldschmidt debate. As a consequence, dated arguments are repeated and the
analysis is subject to conceptual and methodological problems found in older stud-
ies. A brief review of the early Goldschmidt literature, its documented limitations,
and the resulting new generation of studies provides a context to understand our
critique of Barnes and Blevins.
The 1970s to early-1980s period
The 1970s through the mid-1980s represents an essentially knowledge-building
and largely empirical phase in the Goldschmidt literature. In general, studies focus
on replicating and testing Goldschmidt’s hypothesis (Leistritz and Ekstrom, 1986).
Different methodologies, regional contexts, and dependent variables or indicators
of well-being are used to test essentially the same relationship: the effects of large-
scale, hired-labour-dependent farming on communities and rural people. Specific
studies and salient characteristics of this are summarized elsewhere and noted
briefly here (Lobao, 1990; Swanson, 1990).
The purpose of most studies is to examine indicators of one or both farm con-
cepts – scale (sales or acreage) and organization. The latter taps off-farm or non-
family dependence on production factors and includes concepts such as land
tenure, extent of hired labour use, managerial control and capital use. The studies
conceptualize relationships in a generally linear manner, contrasting the effects of
large as opposed to small farms. For example, they hypothesize that the larger the
farm and/or the greater use of hired labour, the more negative the impact on well-
being. Most research is limited to particular regional contexts. Findings are not
directly comparable because of the different geographic settings, time periods and
methodologies. Studies are mainly empirically driven, aimed at testing the effects
of large-scale farming but not at drawing out conceptually why such effects may be
expected. Conclusions are framed in simple defence or support of Goldschmidt
rather than in recognizing the complexity of relationships that determine commu-
nity well-being.
On balance, studies of this period tend to support Goldschmidt but few do so
unequivocally. One of the first deviations recognized was that results depended
upon the indicator of farm structure: larger-scale farming was not found to be
related to poorer conditions in several studies (Flora et al, 1977; Harris and Gil-
bert, 1982; Swanson, 1982). Findings also diverged for other reasons, including
the time period, whether well-being outcomes were measured by economic or
non-economic indicators, regional location, unit of analysis and the modelling or
shape of relationships (Lobao, 1990).