At Royal Statistical Society Ordinary Meetings, a peer-reviewed paper on a topical subject is presented for live discussion. Following the presentation, two invited contributors initially propose a 'vote of thanks' to the authors before discussion is opened to the audience, members of which each have no more than 5 minutes' speaking time (equivalent to 400 words when subsequently the paper and discussion are published in the journal). Written discussion is also encouraged, provided that it is received within 2 weeks after the meeting. If received in advance, such pieces may also be read out by proxy at the meeting if the contributor cannot be present and time permits.

Advance preprints of papers are available from http://www.rss.org.uk/preprints and questions may be sent to journal@rss.org.uk.

This paper explores the use of anchoring vignettes as a means to adjust survey reports of health system performance for differential reporting behaviour using data contained within the World Health Survey (WHS). Survey respondents are asked to rate their experiences of health systems across a number of domains on a five-point categorical scale. Using data provided through a set of vignettes we investigate variations in reporting of interactions with health services across both socio-demographic groups and countries. We show how the method of anchoring vignettes can be used to enhance cross-country comparability of performance. Our results show large changes in the rankings of country performance once adjustment for systematic country-level reporting behaviour has been undertaken compared to a ranking based on raw unadjusted data.

Keywords: Anchoring vignettes; Cross-country comparison; Health care responsiveness; Hierarchical Ordered Probit