Victoria University of Wellington - John Haywood
Wellington Statistics Group - John Haywood

John Haywood

Wellington Statistics Group

The Wellington Statistics Group (WSG) was formed with the intention of developing and furthering the professional (and social) links between the numerous statisticians who are scattered throughout the Wellington region. The group is recognised as a local branch of the New Zealand Statistical Association and WSG were very grateful to receive some initial sponsorship from NZSA. A common goal of an increased appreciation of the relevance and applicability of statistics is certainly shared by both bodies. WSG membership is currently free and is not limited to NZSA members, although we hope NZSA membership will increase as a result of WSG activities. So far, such activities have consisted of reasonably regular early-evening meetings, consisting of an "interesting talk", preceded by light refreshments and followed by a meal. There have typically been about 30 attendees at each talk.

Organisation of WSG is supposedly by a committee, although we don't commit to much! Committee members are:

John Haywood (Convenor)
Alistair Gray
Ian West

As well as receiving some financial support initially from NZSA, WSG is also supported by Victoria University of Wellington. The use of lecture theatres or meeting rooms in VUW's "down town" campus are provided at no direct charge to the group, following an agreement reached between JH and the VUW Pro-Vice-Chancellor (Research).

Dissemination of information to WSG members is by an email distribution list, with automated subscription (subject to approval, by JH currently) and automated removal. The list is hosted by the School of Mathematical and Computing Sciences at Victoria University of Wellington. Currently, there are about 245 subscribers.

The WSG was "formed" at a meeting held on 8 August, 2001, organised by David Vere-Jones. The meeting was also addressed by the Government Statistician, Brian Pink, on the topic, "The Impact of IT on Official Statistics".

In addition to Brian Pink we have heard five talks so far, with a sixth one scheduled for October 17, which is included in the listing here:

Dean Hyslop (Principal Advisor, NZ Treasury) "Understanding Changes in the Distribution of Household Incomes in New Zealand Between 1983-86 and 1995-98" (joint research with Dave Mare)

Brian Bull (National Institute of Water and Atmospheric Research) "Catch-at-age data from New Zealand fisheries" (joint research with Dave Gilbert, also NIWA)

Richard Arnold (Victoria University of Wellington) "Bayesian spectral analysis of white dwarf light curves" (joint research with Tony Vignaux and Denis Sullivan, also VUW)

Jean Thompson (JAD Associates, Wellington) "'Real world' statistical data and the role of the consultant statistician"

Paul Dyer (Head of Investment Strategy, AMP Henderson) "Are financial markets predictable? Some implications for investors"

Frances Krsinich and Mike Camden (Statistics New Zealand) "The use of noise to confidentialise tables"

Anyone who does not presently receive WSG announcements, and wishes to be informed of future events on a regular basis, is very welcome to contact John Haywood directly, via email:John.Haywood@vuw.ac.nz.  So if you're in the Wellington region, we'd like to see you at one of the WSG events soon; you'd be very welcome!

John Haywood

We have recently welcomed our new lecturer, Mik Black. Mik grew up in Nelson and studied at Canterbury before going to Purdue University in Indiana to do a PhD in statistical genomics. Another new arrival is Yuichi Hitose, who has recently begun a PhD with Chris Wild after working for ten years for a pharmaceutical company in Tokyo. And once again, we welcome back James Reilly. James just can’t kick his degree habit — this time he’s going for the big one and is swinging in to a PhD with Alastair Scott and Chris Wild. Farewell to Samuel Manda, who has just completed a two-year postdoc with Renate Meyer and will shortly take up a lectureship in Leeds, UK.

The Statistics Department is a joint player in one of the country’s five new Centres of Research Excellence, ‘The NZ Institute of Mathematics and its Applications’. Alastair Scott is on the board of directors.

[Note that: The New Zealand Institute of Mathematics & its Applications (NZIMA) has recently closed its call for proposals for programmes, applications for Maclaurin Fellowships, postgraduate scholarships and funding for other research activity in the mathematical sciences in New Zealand, but you may find it useful to bookmark their pagehttp://www.nzima.auckland.ac.nz/. - Ed.]

Renate Meyer has had two pieces of good news recently, and is now expecting not only two PhD students through a successful Marsden Grant, but also a new baby! The Marsden Grant funds a three-year project in Bayesian analysis of astrophysical data, but Renate suspects that Project Motherhood will take a lot more work and somewhat more than three years to complete!

Between June and August, the conference season saw people jetting all over the world: Alan Lee attended the 16th Australian Statistical Conference in Canberra; Ilze Ziedins, Geoff Pritchard, Wiremu Solomon, and Ru-Shuo Sheu attended the 28th Conference on Stochastic Processes and their Applications in Melbourne; Ilze also went to the Stochastic Networks conference at Stanford; Alastair Scott and James Reilly attended an Ottawa conference on Recent Advances in Survey Statistics; Arden Miller went to a conference on Design and Analysis of Experiments in Vancouver, and Paul Murrell to the Joint Statistical Meetings in New York. Brian McArdle travelled to Griffith and Otago universities to deliver his popular course on Multivariate Statistics.

On the political front, we have our very own parliamentary candidate in Leila Boyle, who stood as the Labour candidate for the Tamaki electorate in the General Election. After reducing the sitting National MP’s majority from about 5000 to 1200, Leila is pausing only to do a quick significance test before gearing up for her next campaign...

And finally we congratulate our PhD students Carl Donovan and Monique Mackenzie, upon their imminent wedding to take place on 30th November. By late afternoon they are sure to be in no state to remember who they actually invited, so for their own protection the venue is to remain a closely guarded secret!

During the study break, members of the department also took the opportunity to attend some overseas conferences. Dave Whitaker, Nye John, Murray Jorgensen and Judi McWhirter, together with PhD students, Carole Wright and Khangelani Zuma, attended the 16th Australian Statistical Association Conference in Canberra. Bill Bolstad was an invited speaker at ICOTS-6 in Cape Town. Also attending ICOTS was Bruce Millar, who is a part-time tutor in the department. Lyn Hunt travelled to Knoxville, Tennessee and presented a paper at the Statistical Data Mining and Knowledge Discovery Conference.

Nye John is currently on leave. He has just returned from visiting Europe where he attended IBC2002 in Frieberg, Germany. His term as Chairperson of the department was originally to end at the end of the year and so he stepped down from this position to coincide with the start of his Sabbatical. Dave Whitaker has been appointed as the new Chairperson.

Murray has returned from his sabbatical. He spent most of his time away in Canada, visiting with Prof Bill Reed, in the department of Mathematics and Statistics at the University of Victoria. He also attended several conferences including the International Conference of Robust Statistics (ICORS2002) in Vancouver, the 13th Annual Meeting of the Statistical Society of Canada in Hamilton, Ontario, and also the Mixture Models and Bump-Hunting and Measurements Error Workshop in Cleveland Ohio.

And last but not least, James Curran visited Venice, Italy, early September, where he was an invited speaker at the International Conference on Forensic Statistics.

Although Biometrics Matters Limited has existed only since last year I have been working as a singleton statistical consultant since 1997, or, to stretch a point, since 1990. A brief history is that while I was finishing an MSc in statistics and computing at the University of Waikato I started work for MAF as a biometrician at Ruakura in 1980. The big OE started in 1986, and lasted until 1999. I worked first at the Applied Statistics Research Unit (ASRU) at the University of Kent at Canterbury (the other one) for three years until becoming a biometric advisor in Cameroon working for the UK’s Overseas Development Administration (ODA) on an institutional strengthening project for 4½ years. A similar two-year spell followed in Nepal, both courtesy of having a British passport.

With no further ODA contracts available I approached the English research site of Pfizer, a US multinational pharmaceutical company, in 1997 as I had worked for them on-site in Kent through ASRU previously. My first work was on Phase 3 full clinical trials for Viagra and since then I have analysed trials in anti-arrhythmic, anti-fungal, stroke and overactive bladder drug programs. Two of these drugs have required Advisory Committee approval from the FDA (both gained) which gives an extra dimension to the usual drug development process, and one did not progress beyond Phase 2.

As a ‘contractor’ there is the flexibility to work from home and in fact in 1999 I moved permanently back to New Zealand, working by email from here. Electronic communications have improved recently by using Virtual Private Network software which ‘tunnels’ securely through any Internet connection, such as with ADSL. Up to a quarter of each year is also spent overseas, particularly when a new drug project for me starts and Statistical Analysis Plans need to be written (and non-pharmaceutical follow-up contract work in Nepal has just ceased). The time difference with the UK has advantages and disadvantages, with telephone contact possible only in late evenings. However often work requested in the UK at 5 pm is in the Inbox first thing the next morning, as can often occur during the ‘rapid response’ stage of drug approval whereby US or European regulators raise queries, often statistical, to the sponsoring company to help their decision making after the drug submission. It is exciting to be part of the ‘knowledge economy’ doing work not greatly typical of New Zealand’s statistical scene, but not unrelated either to previous experiences.

Our numbers have been reduced still further by the departure of Graham Wood to McQuarrie and John Giffin to Canterbury. Both will be greatly missed. Chin Diew Lai has taken on the additional burden of Subject Leader until a new Professor is appointed.

Mark Bebbington returned from a 3 month visit to Minneapolis, somewhat bemused by temperatures of 33C being followed three days later by snow. Besides investigating airline security by always forgetting and trying to carry on nail-clippers, Mark was participating in the IMA program on Mathematics in the Geosciences. Careful scheduling resulted in the sabbatical being followed by 6 weeks parental leave. Always the careful statistician, Mark and wife Jill were careful to provide for paired data. Craig and Anna are doing well, although Mark looks awful.

Steve Haslett has been less here than there over the last six months with projects overseas at the US Bureau of Labor Statistics as part of an NSF/ASA Senior Fellowship on small area estimation for employment statistics, at the Office of National Statistics (UK) on interviewer effects in small area estimation, and National Statistics Office (Thailand) as part of a World Bank project on improving establishment surveys. He reports that Len Cook has been sighted at Drummond Gate (very near to Random House), is still his usual enthusiastic self, and is very much enjoying heading up ONS.

Ganesalingam returned briefly in July, having been given a flight home from Sultan Qaboos University in Saudi Arabia. He appeared suddenly and unannounced in the Statistics corridor, his head wrapped in scarves in case of sudden sandstorms. He seems to be enjoying his experience of the Middle East.

We are now starting to grapple with the intricacies of organizing next year’s NZSA conference (see elsewhere for details). We are also hosting a short course on Multi-level Modelling in December run by visitors Bill Browne and David Draper. For details seehttp://www-ist.massey.ac.nz/msm/multi.asp.

I’ll start with the new (and some not so new) faces in SMASH Wellington. Mike Camden has joined us from Massey University. Frances Krsinich and Keith McLeod returned from MSD and OE respectively. Other newcomers since the last Newsletter are Katrina Daish (temporarily based in Christchurch), Guido Stark and Jamas Enright. As well Melanie Gin got left off last Newsletter’s list of arrivals.

Frank Nolan has left SNZ, and in fact NZ, to work for the UK ONS under SNZ’s old boss Len Cook. He recently paid a flying visit back here to tell us what the ONS was like. We also got some information on how the Brits are coping with Len, and vice versa. Those of you who know Len will see why we were interested. A large gap has been created by the departure of Alistair Gray to work full-time for SRA. SNZ though has not seen the last of him as Alistair remains on the ABS Methodology Advisory Committee and SNZ does the occasional presentation to that group. Frances Krsinich presented a paper to the November MAC meeting on her work on adding noise to business microdata for confidentiality purposes. At present we use cell suppression which means values in the cells that are likely to allow a user to deduce a business’s response do not get released. With the new methodology you’ll get an approximate number which should be good enough for research. It looks very promising and we are currently trialling it for one of our business datasets to see if what works in theory, works in practice.

Victoria Wilcox has left (again) to go to MSD (for the first time). We are all taking bets as to when she’ll return, though buying a house in Khandallah (not the very posh end I’m told) may mean a long-tail in the distribution. Also leaving are Vicky Barlow (BNZ) and Gareth Minshall (OE).

Helen Stott has taken over as Chief Analyst, to my immense relief. She has shepherded the report on sustainable indicators to release, to her immense relief. The report, like all SNZ output, can be found on our website (http://www.stats.govt.nz) and it’s well worth a look (the report and the website).

Robert Templeton went to The Netherlands and Denmark recently to visit statistical offices and present a paper, "Using tax data and PRN’s to reduce accuracy decay in repeated business surveys" at the International Conference on Improving Surveys. Not his only overseas travel lately as he also went to ASC16 in Canberra where he presented two papers, "Analysis of longitudinal business data" and "Potential use of EFTPOS data in New Zealand for official statistics". John Crequer also was at ASC presenting a paper "Benchmarking with the Kalman filter". John stayed another week in Canberra visiting our time series colleagues at the ABS, as well as trying to get ABS information on various topics as requested by his colleagues. Hard work, so he took the following week off to explore a bit of Central Australia. We await photographic proof. By the time you read this Steve Johnston will have returned from his three weeks visiting various statistical agencies in North America and Europe discussing agricultural surveys.

UoC Stats is in the process of interviewing for a new lecturer, and those of us teaching in 2002, truly look forward to an expanded group in 2003. In 2003 we shall also, once again, try to fill the Statistics Chair.

So what is happening: Marco Reale continues his work on financial math/stat and is co-teaching a new Financial Math course at the 4th year level in 2002. Mike Steel is busy with NZIM Core (Condor et al.) and with research and helping to start a process whereby bioinformatics will have a place in UoC teaching . Irene and members of the Research Centre for Health Care Technology NZ are actively pursuing how bio-statistics, bio-engineering and medical and core physics could play a major role in moves on campus for a Health Sciences Research Centre and Health Research Organisation - all this is part of a Health Sciences push here. Irene is involved in a committee, as Co-Director and interim analysist, to qualify if not quantify "equity", or its lack, here. That may ruffle some feathers!

Jennifer Brown is on sabbatical this year, and apart from one month in Europe, is spending her time in NZ. She is trying to spend at least one week at each University in NZ. The month in Europe was spent first in Italy where she, Bryan Manly and Marco Reale co-hosted an environmental statistics workshop with the University of Naples. Then she was in France and met up with Malcolm Faddy to continue their work with likelihood based modelling. The highlight of the French visit was the fill-your-own 3 litre wine cask with wonderful-tasting red wine for $6. She won’t report how many times the 3 litre cask was refilled. Since then, she attended the NZSA conference in Hamilton, and hosted a visit from Mark Boyce, University of Alberta. Mark and Jennifer are working on an approach to combine adaptive management with population modelling.

Irene Hudson is presenting a paper at the 2002 Royal Statistical Society Conference in Plymouth UK (3-6 September) in the Signal processing and time series session. The paper deals with new statistical developments aimed to resolve whether it is natural habitat or a 7 fold permutation pattern of phyllotaxis that govern wood morphology spatial patterns.

Irene visited Otago University recently and spoke with people interested in global climate change, in which she is developing statistical methods using phenological indicators as proxies - very much enjoyed Otago University and found a great restaurant called "Thyme Out".

Easaw Chacko continues research in sampling and time series.

Chris Frampton left Lincoln University to take up a position with the Christchurch School of Medicine.

I’ve been at DOC for two and half years - with the newly created position of statistician being made permanent last year. There is a great mix of project work, consulting and facilitating training. A major project has been looking at the impact of a 1080 poison possum-control operation on tomtits, which I presented early results from at the December conference in Christchurch. Our results were summarised in the June North and South: "Extensive monitoring of tomtit populations in the Tongariro Forest before and after a 1080 operation in 2001 shows that no adverse impacts occurred. This is considered to be significant because tomtits ...are considered vulnerable to poisoning".

Another project showed the role of the statistician in conserving data. After a discussion with me at the December conference, Robert Davies came to light with an electronic version of 6,200 bird counts in the central North Island in 1978-81, which he had carefully conserved. One of my colleagues described "wasting 2 years of his life" working full-time to collect this data, and without the consultant statistician saving a copy, it would almost certainly never have seen the light of day again. We managed to find some comparable data from 1997-98, and now have a paper ready for submission to a journal reporting some interesting patterns that are apparent.

Recent training opportunities have involved Mike Camden running courses in the 3 main cities for DOC staff on using SPSS for exploring data; and having 25 DOC staff attend a course on mark/recapture with Richard Barker and Shirley Pledger in July. My next event is running a workshop on 16 & 17 December for those with experience in using distance sampling, led by Len Thomas and David Borchers from University of St Andrews, who will be here for the SEEM4 conference in Dunedin the week before. I hope to catch up with NZSA members with an interest in ecology and conservation at that conference.

A trip to Australia in July was successful in making contacts with Australian statisticians interested in conservation. At ASC16 in Canberra, I met up with the statistician in the Queensland Environmental Protection agency, along with making contact with some university-based people with ecological interests. Then I visited the Conservation and Land Management department in Western Australia -where they have twice as many statisticians as DOC - 2! I hope that we may gradually build up an informal network of statisticians working in conservation in Australia and NZ.

As I am the only statistician at DOC, I find it very important to keep in contact with other statisticians I know - at universities, AgResearch and other CRIs, and former colleagues at SNZ. If you are interested in helping conserve NZ’s native biodiversity, please make contact with me at iwestbrooke@doc.govt.nz.

A couple of industrial highlights to report since moving to Christchurch. Particle size problems have been solved with help from Nick Fieller’s software ShefSize (ver 1.0). I had data from the very fine (micron only) to the very large on separate projects. Nick’s log hyperbolic model, including mixtures works well for me. Another highlight has been my exposure to Six Sigma projects for business improvement, and watching Minitab being used. There is a surge in the use of statistics by companies adopting the Six Sigma approach, which I expect to flow through to our own community.

David Fletcher is looking forward to the challenge of working as a statistical consultant in ecology and environmental science fulltime. He will still be based in Dunedin, working through the company that he jointly runs with Darryl MacKenzie, who recently graduated with a PhD in mark-recapture methods from Otago. One of the major contracts that the company has at present is with the US Geological Survey. It involves developing methods for estimating the proportion of sites occupied by a species when the species is not guaranteed to be detected even if it is present. The Marsden Fund is also supporting this project. See the company website (http://www.proteus.co.nz) for more details.

August saw the re-establishment of the Dunedin branch of the NZSA. The aim is to provide a social forum for statisticians from other University departments (e.g. Preventive and Social Medicine and Finance), and for statisticians working at independent research institutes (e.g. AgResearch, Proteus Consulting). The hope is that it will also appeal to local secondary school teachers. At September’s meeting George Seber discussed the notion that ‘statisticians count more than sheep’ to an appreciative audience.

The first few weeks of the second semester saw several visitors from overseas who gave seminars:

In addition, Marti Anderson (Auckland University) visited Dunedin and presented the seminar "Canonical analysis of principal coordinates: a new point of view".

John Harraway attended ICOTS6 in Cape town, South Africa, 7-12 July. He presented an invited paper in the session Statistical Training and Education in Environmental Settings, and organised the session on Multivariate Statistics. Both the Local Organising Committee and the International Programme Committee ran a well organised conference with 500 delegates and a wide range of papers. John was also asked to join the Advisory Committee for the International Statistical Literacy Project. He has responsibility for developing a web page containing resources which can be used by journalists to help them report on statistics and statistical ideas correctly and in an understandable manner. If you are aware of any relevant addresses please contact jharraway@maths.otago.ac.nz. On the way to South Africa John spent four days at Monkey Mia 800km north of Perth seeing the field work being carried out on dolphin behaviour by a PhD student with whom John has previously collaborated.

In the second half of September, Richard Barker will fly to North Dakota for the annual meeting of The Wildlife Society. Darryl MacKenzie finished his PhD in May which investigated computer intensive methods for assessing mark-recapture data. Near finishers are Ruben Roa and Jorge Navarro who have returned to Chile and Mexico, respectively, which is sad for us but happier for their families.

Markus Neuhäuser joined the department in January. He is from Germany, received his doctorate from the University of Dortmund, and subsequently worked as a biostatistician in the pharmaceutical industry. His research interests are nonparametric methods, multiple comparisons as well as location-scale tests and their application in life sciences, especially in drug development, ecology, and ornithology. He recently visited the colleagues at Auckland University and presented recent research about adaptive designs.

The department is busy getting ready for SEEM4, the fourth conference in their series on Statistics in Ecology and Environmental Monitoring, which takes place 9-13 December this year. The theme is "Population Dynamics: The Interface Between Models and Data". The invited speakers are Hal Caswell (Woods Hole Oceanographic Institution, USA), Jean-Dominique Lebreton (CNRS, France) and Byron Morgan (University of Kent, UK). There is also a pre-conference workshop on "Matrix Population Models" from 4th to 6th December, to be run by Hal Caswell and Jean-Dominique Lebreton. Further information is available from Irene Goodwin (casm@maths.otago.ac.nz) or athttp://www.maths.otago.ac.nz/SEEM4/.

Richard Barker and Claire Cameron held a workshop at VUW on analysing mark-recapture data using the computer program MARK. The workshop was co-taught with Shirley Pledger from VUW. Because of demand for places, two workshops were run back-to-back making for an exhausting week. More than 40 took part with about half the participants from the Department of Conservation, with the remainder coming from Universities and Research Institutes from all over New Zealand and Australia.

David Duganzich retires from Ruakura after almost 44 years working there as a statistician. Dave’s laconic humour will be missed by his fellow statisticians and clients. We wish him a happy retirement . We welcome Katarina Domijan, who has just completed an MCMS in statistics at Waikato, to the stats team at Ruakura, and Ray Littler for the duration of his sabbatical leave from Waikato.

Seven of the AgResearch statistics group attended the NZSA one-day conference at Waikato and the whole group participated in the Data Mining workshop on the following day. Then followed our annual retreat (or should it be called an advance), where we considered our strategic direction, did a spot of team building in the Waimangu Volcanic Valley (hoping for no practical demonstrations of the power of nature), and devoted a day to presentations of current projects and research.

David Baird and Harold Henderson attended the SSAI meetings in Canberra in July. David presented "The analysis of microarray data using mixed models" at the Microarray workshop at ANU that followed. Presentations are at http://CBiS.anu.edu.au/workshop/. Peter Johnstone attended the IBC in Freiburg and enjoyed a walk in the Black Forest.

Our group now includes 3 bioinformaticians stationed at Invermay, Nauman Maqbool, Mark Schreiber and Jonathan Warren. Allan McRae has also had a half-time position in the statistics group at Invermay from February to September. He has been investigating the analysis of linkage disequilibrium for helping locate genes in animal mapping populations. He has been awarded a Commonwealth scholarship to study towards a PhD at Edinburgh University.