JRSM >
Volume 100, Number 12 >
Pp. 571-572
School of Nursing and Midwifery, Trinity College Dublin, 24 D'Olier Street, Dublin 2, Ireland
Correspondence to: Professor Mike Clarke Email: mclarke{at}cochrane.co.uk
 | SUMMARY |
|---|
 TOP SUMMARY INTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Design A random sequence was generated by flipping a coin to determine whether a letter would have ‘Sir Iain Chalmers’ or ‘Iain Chalmers’ typed under the signature. This sequence was prepared independently of the preparation of an alphabetic list of the intended recipients and was then used to determine which letter would be sent. The allocation could not be influenced by prior knowledge of the recipient. Two batches of letters were sent.
Setting Medical royal colleges and associated faculties, and postgraduate medical and dental schools in the UK.
Participants Presidents and deans of medical royal colleges and deans of postgraduate medical and dental schools in the UK.
Main outcome measures Rates of response and the mean number of days between the posting of the original letter and the date on the response.
Results No differences between the groups were detected in the response rates: 91% and 90% for ‘Sir Iain Chalmers’ and ‘Iain Chalmers’ (relative rate for response of 1.01, 95%CI 0.83-1.23, P=0.92), or in the mean number of days to response (‘Sir’: 32 days; ‘no Sir’: 33 days).
Conclusions This finding is consistent with a systematic review of responses to postal surveys, in which the effect of the status of the signatory was investigated. Combining our result with the two trials that are most comparable to our study, in which letters from professors were compared with letters from students, gives a relative response rate of 1.00 (95% CI 0.91-1.10, P=0.99). There is, therefore, no evidence from the existing randomized experiments that the status of the signatory has any impact on the likelihood or promptness of response—even if the status was conferred by the British monarch.
| INTRODUCTION |
|---|
|
TOPSUMMARYINTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
A large systematic review that was ongoing at that time had identified hundreds of randomized trials of strategies to improve the response rates to postal surveys.2,3 This review, which identified trials evaluating 98 different ways of increasing response rates to postal questionnaires, found that monetary incentives, the use of recorded delivery, a teaser on the envelope (such as a comment suggesting that the addressee may benefit if they open it) and a more interesting questionnaire topic could more than double the odds of a response.3 The review included some trials of the influence of the status of the signatory on response rates but it had not identified any trials assessing the value of a knighthood (see Discussion). As IC had recently been knighted, he proposed doing a randomized trial to assess whether making explicit in his letters that he was a knight would have any detectable effect on the likelihood of a response or the time to response.
| METHODS |
|---|
|
TOPSUMMARYINTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
The first batch of letters was posted in Oxford to the presidents or deans of the medical royal colleges and associated faculties on Thursday 18 January 2001 by first class mail. The second batch was posted in Oxford to the deans responsible for postgraduate degrees at the medical and dental schools on Thursday 10 May 2001 by first class mail. A reminder letter was sent in the name of Ms Lizi Holmes to people in this second batch who had not responded within four weeks. These reminder letters included a freshly printed version of the original letter (repeating ‘Sir Iain Chalmers’ or ‘Iain Chalmers’ in accordance with the randomization) but without a signature. No further reminder letters were sent and all replies received were counted in assessing response rate and time to response.
In 2007, two of the authors (LC and TC) extracted and cross-checked data from copies of the original letters and their responses. These data included whether the original letter was in the ‘Sir’ or ‘not Sir’ group, the date on the response letter and the salutation used in the response letter. Analyses were done of the proportion of responses, the mean number of days between the posting of the original letter and the date on the response letter, and the salutation used in the response letter.
| RESULTS |
|---|
|
TOPSUMMARYINTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Among the 20 replies in the ‘Sir’ group, 14 were sent back to ‘Sir’. This was followed by ‘Iain’ or ‘Ian’ in 12 cases, and by ‘Chalmers’ in the other two. Two of 18 replies in the ‘no Sir’ group were to ‘Sir Iain’. The other salutations used were ‘Iain’ or ‘Ian’ in eight cases (‘Sir’ 4, ‘not Sir’ 4), ‘Dr Chalmers’ in five cases (‘Sir’ 0, ‘no Sir’ 5), ‘Professor Chalmers’ in three cases (‘Sir’ 0, ‘no Sir’ 3), ‘Mr Chalmers’ in two cases (‘Sir’ 0, ‘no Sir’ 2), and ‘Ms Holmes’ (in reply to the reminder letter) in four cases (‘Sir’ 2, ‘no Sir’ 2).
| CONCLUSIONS |
|---|
|
TOPSUMMARYINTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Among those trials, the two that are most comparable to our study are those that assessed the effect of the status of the sender of the letter. In these, letters from professors were compared with letters from students.4,5 We have not found any other similar randomized trials. A fixed effects meta-analysis combining the response rates for those two trials with the response rate in our trial gives a relative rate of 1.00 (95% CI 0.91-1.10, P=0.99).
In conclusion, there is no evidence from the existing randomized experiments that the status of the signatory of letters that seek responses has any impact on the likelihood or promptness of response—even if the status has been conferred by the British monarch. In discussing this paper, our colleague Sir Iain assured us that he was grateful for the high rate of response to his letters to 42 senior members of the British medical establishment; was confirmed in his prediction that his recently acquired (in 2001) status would not have much impact on the likelihood of his letters being answered; but remains very grateful for the honour conferred on him and his colleagues (including us) by his knighthood (Iain Chalmers, personal communication).
| Footnotes |
|---|
Contributorship MC refined an idea conceived by Iain Chalmers, conducted the analyses, drafted the manuscript and agreed the final version. LC extracted data, revised the manuscript and agreed the final version. TC extracted data, revised the manuscript and agreed the final version.
Acknowledgments We are grateful to Iain Chalmers, Phil Alderson and Lizi Holmes (Gliddon) for their assistance with the conduct of the randomized trial, and to the respondents to the original letters. We are also grateful to Iain for his suggestions, which improved an earlier version of this manuscript.
| REFERENCES |
|---|
|
TOPSUMMARYINTRODUCTIONMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Related articles in JRSM:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
