Edited by Nick Johnstone
Annex 2: Survey Design and Protocol
The survey design and protocol drew inspiration from the principles laid out in Dillman’s (1978) ‘Total Design Method’. QUESTIONNAIRE DESIGN ● ● ● ● ● ● ● ● ● ● designed in collaboration between research teams (approximately 14 researchers) and advisory group members (single representative from each participating country); email correspondence and two meetings held at the OECD in 2003 and 2004 – 18 diﬀerent versions discussed; inputs on survey design obtained from representatives of the OECD’s Business and Industry Advisory Committee; two-way translation from English into French, Japanese, Norwegian, German and Hungarian; pre-tested amongst a selection of representative manufacturing facilities in Japan, Germany and Canada; subsequent modiﬁcations to ease completion, ensuring that survey did not exceed 12 pages in length and remained easily legible; sampling; population of manufacturing facilities with 50 or more employees in seven participating countries; sample derived from universal population databases (except for United States – database of TRI facilities); stratiﬁed sampling by industrial sector (2-digit level) and by facility size (50–99; 100–249; 250–499; Ͼ500). DATA COLLECTION ● postal surveys mailed out to almost 17 000 manufacturing facilities on or around 7 January 2003 (see schedule below); 268 Appendix 2 ● ● 269 ● additional possibility to ﬁll in questionnaire online for United States survey (give website address); accompanying letter (OECD and departmental/university letterheads) addressed to chief executive oﬃcers and/or ‘environmental managers’; two postal reminders (in some cases telephone) to a selection of nonrespondents within one and two months of initial mail-out to increase response rate. Schedule for data collection Indicative timeframe 16 November 2002...
You are not authenticated to view the full text of this chapter or article.