Predicting Response Rate: A Natural Experiment
Predicting Response Rate: A Natural Experiment
复制标题
预测响应率：自然实验
DOI：
--
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
C. Weis
中科院分区：
文献类型：
--
作者：
K. Axhausen;C. Weis
虽然关于调查方法的文献（相关教科书参见 Richardson et al, 1995 或 Dillman, 2000）讨论了答复负担，但似乎没有关于其事前预测的文献，也没有关于由此产生的答复率的文献。尽管如此，市场研究公司必须提前估计采访者的时间要求，以便能够计算研究预算。苏黎世社会研究协会的乌苏拉·雷曼 (Ursula Raymann) 利用其评分系统对一系列自我管理的调查进行了评分（表 1），这些调查是由阿克斯豪森及其在交通规划与系统研究所 (IVT) 的合作者进行的。表 1 回答负担：按问题类型和行动划分的分数 项目 分数 问题或过渡（最多 3 行） 2 每附加行 1 封闭式是/否答案 1 简单数字答案（例如出生年份） 1 最多 5 种可能性的评级 2 超过 5 种可能性的评级 3 左、中、右评级 2 3 个及以上等级的量表 2 用卡片进行的最佳排名 4 第二个和每个附加最佳排名 3 回答最多至5 个字 1 对最多 2 行子问题的回答 2 a) 对 ≤8 种可能性的半开放式问题的回答 2 每增加一个 2 b) 对 ≥8 种可能的半开放式问题的回答 4 每增加一个 3 对“请具体说明”的回答 2 对开放式问题的第一个回答 6 对开放式问题的每一个附加答案 3 混合展示卡 6 向受访者赠送/展示一张卡片 1 展示卡上的每个回答类别 1 过滤器0.5 分支 0.5 © Gesellschaft Fur Sozialforschung，苏黎世，2006
While the literature on survey methods (see Richardson et al, 1995 or Dillman, 2000 for rele-vant textbooks) discusses response burden, there seems to be no literature on its ex-ante pre-diction, nor on the resulting response rates. Still, market research firms have to estimate their interviewers’ time requirements in advance to be able to calculate a budget for a study. Using its point system, Ursula Raymann of the Zurich-based Gesellschaft fur Sozialforschung, rated a series of self-administered surveys (Table 1), which had been conducted by Axhausen and his collaborators at the Institute for Transport Planning and Systems (IVT). Table 1 Response burden: Points by question type and action Item Points Question or transition (up to 3 lines) 2 Each additional line 1 Closed yes/no answers 1 Simple numerical answer (e.g. year of birth) 1 Rating with up to 5 possibilities 2 Rating with more than 5 possibilities 3 Left, middle, right rating 2 Scales with 3 and more grades 2 Best of ranking with cards 4 Second and each additional best ranking 3 Answer to subquestions of up to 5 words 1 Answers to subquestion of up to 2 lines 2 a) Response to half-open question with ≤8 possibilities 2 Each additional one 2 b) Response to half-open question with ≥8 possibilities 4 Each additional one 3 Answer to “please specify” 2 First answer to an open question 6 Each additional answer to the open question 3 Mixing showcards 6 Giving/showing a card to the respondent 1 Per response category on a showcard 1 Filter 0.5 Branching 0.5 © Gesellschaft fur Sozialforschung, Zurich, 2006