Sneddon, Soutar and Mazzarol (2011) fitted the Bass mixed-influence diffusion model to aggregate annual series of pre-sale wool testing adoption in Australia and mapped the estimated parameters onto Abrahamson's fad/fashion/efficient-choice typology, reading the 1988-1992 boom in Additional Measurements testing as a faddish/fashion-driven episode that collapsed with the 1991 abandonment of the Wool Reserve Price Scheme. The authors themselves identified the central limitation: aggregate annual data identify only population-level parameters, so the actual adopter-level mechanisms (efficiency-driven choice, grower-to-grower imitation, fashion-setting by external actors such as brokers, and institutional enforcement) and the validity of the aggregate diagnostic remain unidentified. This problem asks for adopter-level evidence that decomposes the observed adoption and abandonment of Additional Measurements and Clean Colour testing into these mechanisms, tests whether episodes classified as faddish/fashionable or efficient at the aggregate level correspond to those same social processes at the grower level, and links the documented temporal shifts in the diffusion parameters to dated institutional and market events.

Background

The spread of a new technology through a population of potential adopters is classically described by the Bass mixed-influence model (Bass 1969), in which the instantaneous adoption hazard is p+q F(t)p+q\,F(t): F(t)F(t) is the fraction of the population that has already adopted, pp (the coefficient of innovation) measures external influence from sources outside the adopter group such as mass media and change agents, and qq (the coefficient of imitation) measures internal, word-of-mouth influence transmitted by previous adopters. Abrahamson (1991) developed a typology of diffusion and rejection processes distinguishing efficient choice (adoption because the innovation is objectively superior for adopters), fads (contagious, transient enthusiasm), fashions (diffusion driven by outside actors who set what is considered legitimate), and forced selection (adoption compelled by administrative or institutional pressure), warning that observationally similar adoption curves can be generated by very different mechanisms. Sneddon, Soutar and Mazzarol (2011) combined these two frameworks for a naturally occurring agricultural case: the adoption of pre-sale wool testing in Australia, where raw wool lots are objectively measured in a laboratory before auction (Additional Measurements, AM, covering laboratory measurement of sale lots, and Clean Colour, CC, testing of wool colour). Fitting the Bass model to annual aggregate tested weights, they reported three episodes with strikingly different parameters: AM I (1988-1992, p=0.11p=0.11, q=1.01q=1.01), dominated by imitation and interpreted as a fashion-setting boom driven by external influence — fixed-price incentives under the Wool Reserve Price Scheme (WRPS), the Australian government floor-price mechanism that had stabilised wool returns, and promotion of AM by powerful industry participants — that collapsed with the 1991 abandonment of the WRPS; AM II (1993-2006, p=0.07p=0.07, q=0.12q=0.12), read as faddish behaviour whose slower diffusion included an efficient-choice adoption decision from 1994 and possibly a forced selection or compliance process after 1994; and CC (1993-2006, p=0.06p=0.06, q=0.78q=0.78), an imitation-driven fashion-setting diffusion whose external influence came from wool-selling brokers who received incentives for testing their clients' wool and from industry supporters who promoted the tests. Auction price comparisons between tested and untested lots gave only partial support to purely efficiency-based explanations. A companion qualitative study by the same authors (Sneddon, Soutar and Mazzarol 2009), using a sensemaking perspective (an organizational-theory lens that treats growers' adoption decisions as meanings constructed retrospectively through social interaction) on woolgrowers' adoption, abandonment and use of testing technologies, showed that on-farm innovation is a dynamic social process shaped by growers' identity frames and social context, but it did not quantify mechanism contributions. The open issue left by the 2011 paper is that its conclusions rest on aggregate annual data: the estimated pp and qq are population-level summaries that cannot reveal which subgroups, regions or relationships with wool brokers (the intermediaries who prepare and sell growers' wool at auction) generated imitation or external influence, do not indicate changes in the level of imitation and innovation at specific points in time, and labelling an episode 'faddish' or 'fashionable' from aggregate parameters leaves the adopter-level processes unidentified.

Problem Statement

For the Australian pre-sale wool-testing diffusion documented by Sneddon, Soutar and Mazzarol (2011) — the Additional Measurements series (1988-2006, split into the 1988-1992 boom-and-collapse episode and the 1993-2006 slow-growth episode) and the Clean Colour series (1993-2006) — determine, with adopter-level evidence, the mechanisms that generated the observed aggregate patterns, and test whether the aggregate Bass/Abrahamson classification of the episodes is valid at the micro level. Specifically, the question is to: (1) use farmer-level adoption and abandonment histories, together with broker and institutional records and auction price data — or, where such data are not already available, a defensible adopter-level research design that specifies the obtainable sources from which they would be assembled and linked and the identification strategy — to decompose adoption behaviour into efficiency-driven choice (responses to measured price premia and discounts and farm-level profitability), imitation among growers, fashion-setting by external actors (wool brokers, industry bodies, promotional campaigns), and institutional enforcement or forced selection, allowing for cross-sectional heterogeneity across growers, regions and broker relationships; (2) test directly whether the source's aggregate-parameter interpretations — AM I (the 1988-1992 boom) as a fashion-setting process driven by external influence and imitation, AM II (the 1993-2006 series) as faddish behaviour that included an efficient-choice adoption decision from 1994 and possible forced selection or compliance thereafter, and CC as a fashion-setting process driven by broker incentives and imitation — are in fact generated by those social processes at the adopter level; and (3) estimate how the external-influence and imitation parameters pp and qq shift over time and identify which dated, measurable institutional and market events (including the 1991 abandonment of the Wool Reserve Price Scheme and changes in broker incentives) account for the identified shifts. The single overarching objective is a causally defensible identification of the mechanism mix behind the observed diffusion, replacing interpretation-by-aggregate-parameters; refitting richer curves to the aggregate annual data alone, without adopter-level evidence or a specified adopter-level research design of the kind called for in (1), identifies only population-level coefficients, not that mechanism mix.

References

  1. Joanne Sneddon, Geoffrey Soutar, Tim Mazzarol (2011), Modelling the faddish, fashionable and efficient diffusion of agricultural technologies: A case study of the diffusion of wool testing technology in Australia, Technological Forecasting and Social Change 78(3), 468-480, DOI 10.1016/j.techfore.2010.06.005, https://doi.org/10.1016/j.techfore.2010.06.005
  2. Frank M. Bass (1969), A New Product Growth for Model Consumer Durables, Management Science 15(5), 215-227, DOI 10.1287/mnsc.15.5.215, https://doi.org/10.1287/mnsc.15.5.215
  3. Eric Abrahamson (1991), Managerial Fads and Fashions: The Diffusion and Rejection of Innovations, Academy of Management Review 16(3), 586-612, DOI 10.2307/258919, https://doi.org/10.2307/258919
  4. Joanne N. Sneddon, Geoffrey N. Soutar, Tim Mazzarol (2009), On-farm innovation in the Australian wool industry: A sensemaking perspective, Experimental Agriculture 45(3), 295-312, DOI 10.1017/S0014479709007881, https://doi.org/10.1017/S0014479709007881
  5. Daekook Kang (2021), Box-office forecasting in Korea using search trend data: a modified generalized Bass diffusion model, Electronic Commerce Research 21(1), 41-72, DOI 10.1007/s10660-020-09456-7, https://doi.org/10.1007/s10660-020-09456-7
  6. Yari Vecchio, Marcello De Rosa, Gregorio Pauselli, Margherita Masi, Felice Adinolfi (2022), The leading role of perception: the FACOPA model to comprehend innovation adoption, Agricultural and Food Economics 10(1), article 5, DOI 10.1186/s40100-022-00211-0, https://doi.org/10.1186/s40100-022-00211-0
  7. Rafael Mesa Manzano, Javier Esparcia Pérez (2023), Theoretical framework and methods for the analysis of the adoption-diffusion of innovations in agriculture: a bibliometric review, Boletín de la Asociación de Geógrafos Españoles 96, DOI 10.21138/bage.3336, https://doi.org/10.21138/bage.3336
  8. Ezequiel Bernardo Gonzalez, Marcos Horacio Easdale, Diego Mariano Sacchero (2021), Socio-technical networks modulate on-farm technological innovations in wool production of North Patagonia, Argentina, Journal of Rural Studies 83, 30-36, DOI 10.1016/j.jrurstud.2021.02.015, https://doi.org/10.1016/j.jrurstud.2021.02.015
  9. M. J. Hötzel, J. N. Sneddon (2013), The role of extensionists in Santa Catarina, Brazil, in the adoption and rejection of providing pain relief to calves for dehorning, Journal of Dairy Science 96(3), 1535-1548, DOI 10.3168/jds.2012-5780, https://doi.org/10.3168/jds.2012-5780