Food Deserts in British Cities, 2000-2001

The aims of the project were:<br> <li>to provide an evaluation of the nature of 'food deserts' in British cities as a contribution to the social exclusion and health inequalities debates and their policy implications;</li><br> <li>to design and conduct, within one strategically chosen area of poor retail access in Leeds, a major 'before/after' (baseline and follow-up) study of the impact of the opening of a large new food store on a group of low-income households.</li><br> <br> The study required a major and potentially extremely difficult exercise in social survey research, involving a two-wave food diary/household questionnaire survey, focused on the kind of deprived urban area known to pose enormous problems regarding response and attrition rates for social survey research. The survey design<br> consisted of two waves: <br> <li>pre-intervention (June/July 2000), approximately five months 'before' the opening of a new supermarket food store in November 2000, and</li><br> <li>post-intervention (June/July 2001), seven/eight months 'after' the opening of the new store.</li><br> Each wave consisted of:<br> <li>a seven-day food consumption diary/check list - the respondent completed this but interviewer placed and collected it;</li><br> <li>a wide-ranging, interviewer-administered household questionnaire. <br> <br> The diary and questionnaire were completed, as in the National Food Survey, by the person primarily responsible for the domestic food arrangements of the household.<br> <br> The survey fieldwork was contracted to and completed by Taylor Nelson Sofres (TNS). The survey was designed by the Southampton research team with technical advice from TNS and specialists at the team's industrial partner (J. Sainsbury plc). Survey instruments and methods were piloted in February 2000, and subsequently modified and refined.<br> <br> Targets of 1000 respondents in Wave 1 and 600 in Wave 2 were set, in an attempt to ensure sufficient statistical power for the assessment of dietary change across the survey waves. Intense efforts were made by the research team to minimize sample attrition between waves of survey and hence reduce attrition bias problems in the subsequent analysis; monetary incentives (in the form of non-food retail shopping vouchers) were structured to maximise recruitment in Wave 1 of the survey, and sustain retention of respondents into Wave 2.<br>

本项目的研究目标如下：
1. 评估英国城市中“食品荒漠（food deserts）”的本质，为社会排斥与健康不平等相关讨论及其政策启示提供研究支撑；
2. 在利兹一处经战略选定的零售可达性较差区域，设计并开展一项大型“前后对照（before/after）”研究（即基线与追踪研究），探究大型新建食品店开业对低收入家庭群体的影响。

本研究需在社会调查研究领域开展一项规模庞大且难度极高的工作，包含两波次的食品日记/家庭问卷调查，研究对象为已知的贫困城区——这类区域向来在社会调查中面临极低的应答率与极高的失访率问题。调查设计分为两个波次：
- 干预前波次（2000年6月/7月）：即2000年11月大型超市食品店开业前约5个月；
- 干预后波次（2001年6月/7月）：即新店开业后7至8个月。

每个波次的调查内容包括：
1. 为期7天的食品消费日记/核查清单：由受访者自行填写，但由调查员上门放置并回收；
2. 内容全面的、由调查员实施的家庭调查问卷。

本研究的日记与问卷填写规则参照《全国食品调查》，由家庭中主要负责日常食品采购安排的成员完成。

本次调查的实地工作由泰勒纳尔逊索福斯公司（Taylor Nelson Sofres, TNS）承接并完成。调查方案由南安普敦研究团队设计，并获得了TNS以及该团队产业合作伙伴——森宝利股份有限公司（J. Sainsbury plc）的专家提供的技术支持。调查工具与方法已于2000年2月开展预试验，并随后进行了修改与优化。

研究设定了第一波次1000名受访者、第二波次600名受访者的目标样本量，以确保获得足够的统计效力来评估各波次间的膳食变化。研究团队投入了大量精力以降低波次间的样本失访率，从而减少后续分析中的失访偏倚问题：研究设置了货币奖励机制（以非食品类零售购物券形式发放），以最大化第一波次的招募率，并维持受访者参与至第二波次的留存率。