Wildlife consumption and recall accuracy – but is it recall of hunting, of cooking or of eating?


  • H. Newing,

    Corresponding author
    1. Durrell Institute of Conservation and Ecology (DICE), School of Anthropology and Conservation, University of Kent, Canterbury, Kent, UK
    • Correspondence

      Helen Newing, Durrell Institute of Conservation and Ecology (DICE), School of Anthropology and Conservation, University of Kent, Canterbury, Kent CT2 7NS, UK.

      Email: H.S.Newing@kent.ac.uk

    Search for more papers by this author
  • F. A. V. St. John

    1. Durrell Institute of Conservation and Ecology (DICE), School of Anthropology and Conservation, University of Kent, Canterbury, Kent, UK
    Search for more papers by this author


Read the Feature Paper: Assessing the accuracy of interviewed recall for rare, highly seasonal events: the case of wildlife consumption in Madagascar

Other Commentaries on this paper: Unusual data in conservation science: searching for validation

Response from the authors: Practical directions for the use of recall data in conservation science

The paper by Golden, Wrangham & Brashares (2013) on accuracy of recall on wildlife consumption is a welcome addition to the sparse literature on an important subject. They reach two main conclusions: first, counterintuitively, that studies of wildlife consumption that depend upon recall are best carried out in the low hunting season, and second, that annual recall may be more accurate than monthly recall, at least for taxa that are harvested rarely or seasonally. In this commentary, we highlight several underlying assumptions in their paper that demand further scrutiny and some points of methodology that need further consideration in order to work towards a robust methodology for studies of this kind. However, our main aim in doing so is not simply to critique this one paper but to make a wider point: if we are to avoid reinventing the wheel, then those of us working in conservation social science need to make full use of the wealth of knowledge held in other, more established disciplines (St. John, Edwards-Jones & Jones, 2010).

Some of the assumptions that call for further scrutiny are as follows. For each of these there is already an established body of knowledge that offers insights:

  • (1) The assumption that data gathered through daily diet calendars represents ‘true’ wildlife events. A more realistic assumption would be that daily diet calendars tend to give more accurate data than recall, although even this assumption should be treated with caution. There is a substantial literature on inaccuracy in daily calendars, particularly within health studies (e.g. Wiseman et al., 2005), development studies (Beegle et al., 2012) and ethnobiology (Shanley, 1999; Menton et al., 2010). Sources of inaccuracy include reporting fatigue; recall error connected to gaps or delays in completing the calendar; incomplete knowledge of the diary-keeper of the behaviour of all members of the household; and of course intentional errors related to social and cultural norms or the illegality of consuming certain species.
  • (2) The assumption that ‘deliberate misleading was absent or trivial’ because the research team had worked there for 5 years. Directly studying taboo (Jones, Andriamarovololona & Hockley, 2008) or illegal behaviours is challenging as those involved may not wish to reveal themselves for fear of the consequences (Keane et al., 2008; Gavin, Solomon & Blank, 2010). While there is evidence from Madagascar that asking questions about hunting legally protected species may not be problematic in areas where the level of law enforcement is low (Razafimanahaka et al., 2012), it is naïve to expect people to provide truthful answers to incriminating questions when asked directly (St. John et al., 2012). Evidence comparing direct questioning to specialized methods for asking sensitive questions provide evidence that people underreport their involvement in illicit, or otherwise sensitive activities (Solomon et al., 2007; Silva & Vieira, 2009).
  • (3) The assumption that ‘wildlife consumption can be viewed as nearly all of harvest’. Assuming that ‘harvest’ means what is hunted (by men) by each household, there are several reasons why harvest and diet may not map perfectly onto one another. These include gifts of meat or cooked food, and non-cash barter and informal sale of meat (or other wildlife products), both within and beyond the community. Greater information on the study community would be needed to evaluate these points, including on informal and formal economic interactions, levels of mobility, fluidity of community membership and patterns of interaction with outsiders. There is a very extensive literature within anthropology that could inform development of methodologies to deal with these points (for a brief overview see Russell & Harshbarger, 2003).
  • (4) Related to the above, the implicit assumption that any differences found were due to the difference in methodologies rather than gender differences. This is naïve, especially since the research design was based on the fact that men and women have different domains of activity – hunting and the preparation and cooking of food, respectively – and may therefore have different knowledge sets pertinent to wildlife consumption. More contextual information would be needed to assess this, but it raises the issue of what exactly men and women were reporting on: hunting, cooking or eating? There is a well-established literature in anthropology and development studies that could inform consideration of these issues.

In summary, although Golden et al. (2013) make an important contribution to an as yet poorly developed methodological field in conservation social science, and although they cite publications from a range of disciplines, there are several areas where additional disciplinary perspectives could strengthen both the methodology and the conceptual framework greatly. This is a common problem: there is a long list of disciplines that are relevant to studies in conservation social science, but there is a limit to how far any one researcher can develop expertise in multiple disciplines, and therefore, all too frequently we do waste time reinventing the wheel. In order to improve this situation, it is crucial that measures are taken to facilitate greater cross-disciplinary understanding. This point is not new, and yet progress remains slow (Mascia et al., 2003; Balmford & Cowling, 2006; Newing, 2010; 2011). We propose three steps that could facilitate cross-disciplinary exchange: (1) to report studies in a way that is accessible to researchers from a range of disciplines. For example Golden et al.'s analysis, using mean squared errors, is a useful approach for exploring the relative utility of different methods, but a simple chart displaying the quantity of each species consumed as estimated by each method and inclusive of a measure of variability would provide a more transparent and complete picture of their findings in a format readily understood by a wider audience; (2) to extend the practice of providing methodological detail as appendices to published papers or supplementary online material. This would greatly increase the potential for researchers to develop new cross-disciplinary methodologies; and (3) to make a targeted effort to develop the resources and institutional structures for interdisciplinary training, in order to build a new generation of innovative cross-disciplinary researchers in the field of conservation social science.