culturally and geographically distinct from the mainland. This methods section is taken from Stanford et al. 2013.
4.2.1 Collation of Secondary Data
Secondary data were collated from three sources. Firstly, from fisheries statistics annually collected at the provincial and district level by the agency for
Fisheries and Oceans DKP. Secondly, through a once a decade census, most recently conducted in 2010, and thirdly through a social-economic survey
Pendataan Sosial Ekonomi Penduduk conducted by the Centre for Statistics BPS that began in 2005 and was formalized into the social protection program in
2008 and repeated in 2011 BPS, 2011. Fisheries statistics were used from 2010 and combined with poverty statistics from 2011, the closest year available to the
decadal census of 2010.
4.2.2 Verification
Both through human failings see for example Heazle and Butcher, 2007, distrust between government agencies and the reality that much of the catch of
West Sumatra is landed at widely dispersed sites and quickly sold on the beach, making monitoring very difficult, there are concerns over the reliability of some
fishing statistics. In the face of this uncertainty this research has taken a slightly different approach. Census data, rather than fisheries statistics, were used for four
of the five measures in the two indices in order to overcome the common obstacle in measuring fishing dependency of inadequate data Phillipson, 2000. These
were routine censuses, using a standardized methodology that covered all economic sectors of society. No interview based data is impervious to mistruths
yet the census interviews were conducted by local members of the community and because these interviews were not tied to financial help from a specific sector
there was little incentive to misreport. Besides data accuracy the advantage of using census data is that it is available at an extremely detailed spatial scale,
valuable for GIS mapping. Data from the 2010 census data were triangulated with fisheries statistics collated from both the BPS and the DKP to identify
discrepancies. Where these existed, data from adjacent years were compared to
identify if the anomaly was an input error for 2010. Where repeated discrepancies occurred these were brought to the attention of the relevant local staff in the DKP
for clarification. The only measure of fishing dependency that originated from fisheries
statistics was production data. A sensitivity analysis was used to assess the effect that inaccuracies in these data would have on the fishing dependency index. The
fishing dependence index was run with and without the production data to identify the changes that resulted, and the results from the fishing dependency index using
2010 data were compared with 2008 data. This sensitivity analysis demonstrated both robustness in the analysis and consistency between results from 2008 and
2010 see section 4.3.2.
4.2.3 Analysis Generating the fisheries dependence index