Over the past 15 months, I’ve made repeated requests to the University of Western Australia for a complete copy of Lewandowsky’s Hoax data in order to analyse it for fraudulent and/or scammed responses. Up to now, none of my previous requests were even acknowledged.
I was recently prompted to re-iterate my longstanding request by the retraction of Lewandowsky’s Fury. This time, my request was flatly and permanently denied by the Vice Chancellor of the University himself, Paul Johnson, who grounded his refusal not on principles set out in university or national policy, but because the University administration’s feelings were hurt by my recent blogpost describing the “investigation” by the University administration into the amendment of Lewandowsky’s ethics application .
In September 2012, I carried out several preliminary analyses of Lewandowsky’s data using a grey version then in circulation. Like Tom Curtis of SKS, I concluded that some of the responses were fraudulent. In response, Lewandowsky argued that I had not “proved” that the responses were fraudulent. The grey version of the data lacked important metadata for the individual responses, all of which was necessary for a forensic examination. In addition, Lewandowsky had removed several questions (including CYIraq) from the grey version and had removed numerous responses for various reasons, including duplicate IP addresses, incomplete data or implausible consensus or age responses.
In order to carry out a thorough analysis, I particularly wanted to see metadata that included the questionnaire used by each respondent and the date of each response.
In February 2013, I sent a polite request to Lewandowsky, who did not acknowledge my request.
Subsequent to this, Roman Mureika obtained from coauthor Oberauer a version of the dataset that included the CYIraq and life satisfaction questions, but still without metadata on questionnaires and dates as well as the several hundred responses that Lewandowsky had excluded.
After waiting a couple of months, I sent a polite request to Caixing Li of the UWA Human Resources Ethics Office. Again no response.
Reminded of these past refusals by the recent retraction of Fury and Barry Woods’ efforts to obtain Lewandowsky data, I once again requested data, this time writing Murray Mayberry, Head of the School of Psychology, copying the Human Resources Ethics Office, the Vice Chancellor and the Australian Research Council, as follows:
Last year, the editor of Psychological Science suggested that I submit a comment to the journal regarding statistical errors in Lewandowsky et al (Moon Hoax).
Since then, I have unsuccessfully been trying for over a year to obtain comprehensive data from the University of Western Australia pertaining to the Lewandowsky “Hoax” study. In the last year, I have received no acknowledgement whatever.
Let me recap the request.
1. After my initial failure, Roman Mureika has received a subset of the original data, from which several hundred responses had been removed. I request a copy of the dataset including the removed responses, with a denotation of the removed responses.
2. I request that each response (row) show the version of the questionnaire. There are two reasons for this: first, Lewandowsky said that the versions had different question orders for “counterbalancing”. Second, the questionnaire version provides some information on the originating blog. This information would be retained in any competent design.
3. I request that each response (row) show the date of each response. This is important because the responses are not homogeneous to order number. In addition, Lewandowsky made a preliminary presentation of results while the survey was still open and I wish to check if this had any effect. Again this information would be retained in any competent design.
4. The survey was also filled out by respondents at the UWA using a different questionnaire number. Although this form of distribution was listed in Lewandowsky’s ethics amendment, Lewandowsky excluded this data from the original analysis. Lewandowsky has said that this exclusion didn’t matter, but I wish to verify this.
Previous requests that were not acknowledged include a request to Lewandowsky on February 6, 2013, to Caixia Li on April 4, 2013.
For your information, the former employee of the University of Western Australia, who has thus far withheld the data, also criticized me in articles, published insulting commentary on a blog then sponsored by the University and purported to diagnose that I have psychological disorders in an article now retracted by a journal but defiantly re-published on a UWA website.
This time, I received a response from Vice Chancellor Johnson himself, flatly and categorically refusing to provide me the data. But Johnson purported to justify the refusal not in accordance with university or national policy, but because of my blog post on the ethics “investigation” of the amendment of Lewandowsky’s ethics application:
Dear Mr McIntyre,
I refer to your series of emails to University officers including Professor Maybery and myself (which you have copied to other recipients including the Australian Research Council) in which you request access to Professor Lewandowsky’s data.
I am aware that you have made inflammatory statements on your weblog “Climate Audit” under the heading “Lewandowsky Ghost-wrote Conclusions of UWA Ethics Investigation into “Hoax”” including attacks on the character and professionalism of University staff. It is apparent that your antagonism towards Professor Lewandowsky’s research is so unbalanced that there is no useful purpose to be served in corresponding with you further. I regard your continued correspondence to be vexatious and there will be no further response to your requests for data.
Professor Paul Johnson,
First, Johnson’s objections to my blogpost of a few days ago do not justify their prior refusals.
Second, I do not accept Johnson’s accusations that the blogpost demonstrated that I had become “unbalanced” nor do I agree that it contained “inflammatory statements”.
I try to write with minimal editorializing and few adjectives in order to let the facts speak for themselves as much as possible. Thus, “inflammatory statements” are unusual in Climate Audit posts, as even my severest critics generally recognize. (Mann, for example, observed in his pleadings in Mann v Steyn that I had never publicly accused him of “fraud”).
In the post in question, I compiled a detailed chronology of events in which I quoted directly and copiously from the original correspondence with negligible editorial comment. In my review, I didn’t see any language that can fairly be described as “inflammatory” – if readers disagree, please tell me. Nor does it appear to me that I made any editorial statements about the “professionalism” of the University staff. Lewandowsky said that the correspondence showed that he and the university administration had meticulously dotted each i and crossed each t. While I believe otherwise, I did not directly express my opinion or editorialize otherwise in the blogpost. Instead I thoroughly presented the facts to readers, leaving them to draw their own conclusions.
I particularly take exception to Johnson’s claim that this blogpost demonstrates that I have become “unbalanced”. On the contrary, it seems to me that the blogpost shows my usual carefulness in avoiding needless editorializing. Even my severest critics have long recognized that Climate Audit posts avoid libelous claims and, when re-read, seldom offer targets. Oscar Wilde once observed that a true gentleman never hurts someone’s feelings unintentionally. If I have unintentionally hurt anyone’s feelings within the University of Western Australia administration, I apologize.
In addition, as is my policy, if there are any inaccuracies in the criticized blogpost, I will undertake to promptly correct them when brought to my notice.
In any event, even if my blogpost did contain “inflammatory language” about university administrators (which I deny), that is not grounds for refusing data.
Johnson seems to be unaware of how data obstruction played out in climate. Phil Jones famously said “Why should I make the data available to you, when your aim is to try and find something wrong with it.” This attitude has never been acceptable to the wider public that pays the salaries of Jones and other climate scientists. Much of the public distaste for Phil Jones, Michael Mann and the Climategate correspondents arose from their attempts to obstruct data access.
Post Climategate, it has become somewhat harder for climate scientists to obstruct data access, even to critics, though problems remain at many journals. One notable exception is Nature which has moved decisively to eliminate the charade under which obstructing authors used third parties an excuse for not providing data. Nature now requires that authors must obtain permission from third party authors to release any previously unarchived data, thus cutting off the daisy chain previously beloved by obstructing authors.
Now Vice Chancellor Johnson of the University of Western Australia has joined the ranks of data obstructionists. I believe that this was an unwise decision on Johnson’s part, one that I hope that he promptly reconsiders.
University data policies say that research data “must be available for discussion with other researchers” – a policy that I referred to in my initial request. Here are other relevant clauses:
3.2 Research Data must be held along with other records associated with the research project and retained in accordance with the Western Australian University Sector Disposal Authority and the UWA Records Management Policy.
3.3 Wherever possible, original research data must be retained in the school or research centre in which they were generated and retained in accordance with clause 3.2 of this policy
3.4 Individual researchers are able to hold copies of their research data for their own use. However, retention solely by the individual researcher provides little protection to the researcher or the University in the event of an allegation of falsification of data.
3.7 In all cases, prior to the publication of research findings a Location of Data Form must be completed.
3.8 Research data related to publications must be available for discussion with other researchers.