Newsletter Will Be the Last in the Quarterly Format

Home , Adrian Pagan, Chris Heyde, Deane Terrell

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 1 News

ISSUE NO. 160 September 2017 From the Acting Editor This issue of the SSA Newsletter will be the last in the quarterly format. It is sad to see a part of the fabric of the Society disappear; however, the Society is convinced Editorial 2 that the new monthly eNewsletter will be more effective and timely in meeting the communication needs of its members. The President of SSA has written elsewhere in this issue about the Newsletter and its Editors. Events 4 Chip Heathcote died on July 18th 2016, and a short obituary written by Alan Welch was circulated to SSA members. Eugene Seneta has written an extended obituary Christopher Robin Heathcote, for Chip, which he has interwoven with a history of Statistics at the ANU and with 1931-2016, and Statistics at ANU 5 some of Eugene’s reminiscences of Chip and the ANU. It is published in full in this issue; we hope you enjoy reading it.

QLD Branch 18 Doug Shaw Acting Editor VIC Branch 19 This is a very special SSA Newsletter SA Branch 19 because it is the last quarterly issue. Editors Since Dennis Trewin and Richard Dennis Trewin Tweedie set up the first Newsletter (May 1977 – November 1980) From the Office 20 in 1977 many dedicated members of the Statistical Society have put in Richard Tweedie considerable effort in managing this (May 1977 – February 1982) important communication tool for the Chris Edwards Society. (May 1978 – August 1980) The list here shows the names of those Bob Forrester who held the position of Newsletter (November 1980 – May 2001) Editor until the end of 2015. After having been SSA Newsletter editor Ian McRae from 1982 until 2001, Doug Shaw (February 1981 – May 1984) then took over as Acting Editor, never Doug Shaw intending to make it a long-term role. (May 1982 – May 2001) Due to the struggle to find new editors, Ray Barge and because of the introduction of the (August 1984 – February 1986) recent eNewsletter, it was decided to discontinue the quarterly SSA Eden Brinkley Newsletter after almost forty years. (May 1986 – May 2003) Thank you to everyone who has Alice Richardson contributed their time, expertise and (August 2001 – November 2014) writing skills. Thank you especially to the Michael Adena Editors, the Branch Newsletter contacts, (August 2003 – November 2014) the many, many authors of articles and the Society members who have Sonia Langford supported this SSA icon over the years. (February 2015 – November 2015) Scott Sisson SSA President

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 2 SEPTEMBER 2017 Issue 160 Section Chairs SSA Central Council Bayesian Statistics Executive Committee SSA Chair: Christopher Drovandi President: Scott Sisson PO Box 213, Belconnen ACT 2616 [email protected] Secretary: Doug Shaw Phone 02 6251 3647 http://www.statsoc.org.au/sections/ [email protected] Email [email protected] bayesian-statistics.htm Website www.statsoc.org.au Environmental Statistics Branch Presidents and Branch Secretaries Editor Co-Chair: David Warton Dr Douglas Shaw, Acting Editor [email protected] Canberra Co-Chair: Jakub Stoklosa President: Robert Clark Correspondence [email protected] Secretary: Warren Müller Please direct all editorial http://www.statsoc.org.au/ [email protected]. correspondence to environmental-statistics.htm Email [email protected] au Business Analytics New South Wales Disclaimer Chair: Mark Griffin President: Michael Stewart The views of contributors to this [email protected] Secretary: Thomas Fung Newsletter should not be attributed to http://www.statsoc.org.au/sections/ [email protected] the Statistical Society of Australia, Inc. business-analytics/ Queensland Statistical Education President: Lee Jones Chair: Peter Howley Secretary: Dimitrios Vagenas [email protected] [email protected] http://www.statsoc.org.au/statistical- South Australia education.htm President: Julian Whiting Secretary: Paul Sutcliffe Official Statistics [email protected] Chair: Stephen Horn [email protected] Victoria President: Jessica Kasza http://www.statsoc.org.au/sections/ Secretary: Charles Gray official-statistics/ [email protected] Biostatistics Western Australia Co-Chair: Sabine Braat President: Alethea Rea [email protected] Secretary: Tom Davidson Co-Chair: Karen Lamb [email protected] [email protected] http://www.statsoc.org.au/medical_ statistics Further contact details for Society Secretaries and Section Chairs can be Young Statisticians’ Network obtained by contacting the Society on Raaj Kishore Biswas (02) 6251 3647. [email protected] http://www.statsoc.org.au/about- young-stats.htm

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 3 Member News Events SSA member Penny Robinson was WIMSIG Conference 2017: Celebration of Women in Australian Mathematical awarded the “Inspiration Award for Sciences Individual Achievement (Adult)” on 24-26 September 2017, Adelaide 27 April 2017 at the Autism Spectrum Australia (Aspect) awards. Developing Your Career to Thrive in a Data-rich, Technology-driven, Reproducible Research Environment https://www.autismspectrum.org. 25 September 2017, Tweed Heads au/news/congratulations-2017- recognition-award-recipients Young Statisticians Conference 2017 26-27 September 2017, Tweed Heads Congratulations, Penny! Joint International Society for Clinical Biostatistics and Australian Statistical Conference 2018 26-30 August 2018, Melbourne

To have your event added to this list, please forward the event details in the above format to [email protected].

ISCB ASC18 26-30 AUGUST 2018 MELBOURNE, AUSTRALIA ISCB

International Society for Clinical Biostatistics HOSTED BY: ISCB

International Society for Clinical Biostatistics

www.iscbasc2018.com Joint International Society for Clinical Biostatistics and Australian Statistical Conference 26-30 August 2018

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 4 Christopher Robin Heathcote, 1931 - 2016, and Statistics at ANU Chip Heathcote, as he was universally Chip Heathcote, a PhD from ANU, known, was born 18th April, 1931, in was appointed Senior Lecturer in Ted Secunderabad, India, and died 18th Hannan’s fledgling department in 1962, July, 2016, in Canberra Australia and later became its Head when he was, with Deane Terrell, one of the two The Australian National University Professors who replaced Ted Hannan (ANU), Canberra, founded in 1946 was in 1971 when Ted moved to the IAS designed to develop postgraduate department. training and research within Australia, and to attract to Australia eminent Through successive incarnations of academics. the SGS Department, Chip provided a leadership and paternal role in In January 1952 the first Professor of research, research mentoring, teaching Statistics was appointed in the ANU’s and administration until his retirement Research School of Social Sciences. in 1996 as Emeritus Professor. The This was an Australian, P.A.P (Pat) history of the Department and people Moran. His main efforts were to be and work it produced is fleshed out directed towards his personal research, within this extended obituary of Chip. which was basically theoretical statistics including probability theory and stochastic models, and the Chip’s Early Life and training of postgraduate students. His Early Academia first two PhD students, from about 1953 Secunderabad was founded in 1806 as to completion in 1955 were E.J. (Ted) a British cantonment and developed Hannan and J. M. (Joe) Gani. Hannan directly under British rule until 1948. was appointed the first Professor of Chip’s father, also Robin, was a military Statistics, in the Canberra University doctor. Chip came from a very devout College, taking up his position in Protestant family. 1960. The College became the ANU School of General Studies (SGS) on The oldest of three brothers, he was 30th September, 1960, when it was bookish and loved horses. As his amalgamated with the original ANU parents moved, he lived in various to form the “new” ANU. The “old” ANU places in what is now India and became the ANU Institute of Advanced Pakistan. He recalls 9 December 1941, Studies (IAS). There were thus two the day the Japanese sank the “Prince Departments of Statistics at ANU, of Wales” and the “Repulse”: Hannan’s primarily undergraduate teaching department in the SGS, and “I was aged 10 living in Karachi when Moran’s research department in the news of the sinkings came through. IAS. Joe Gani (2005) has described the The immediate reaction of my mother history of Statistics at ANU up to 2002 was to have the malis (gardeners) dig in some detail, and I am indebted trenches in the garden. … it made to his important study, to which the my mother feel better. The bulk of present study is in part a complement. the Army, including my father, was in Gani (1994) and Seneta (2017) are Africa. ...Karachi, being towards the respectively obituaries of Ted Hannan, western part of India was about as far and of Joe Gani who died a few weeks from the Japanese as possible.” before Chip Heathcote [From Chip’s manuscript: “Twilight of the Raj: India in the Decade Before 15th August 1947.”]

> Continued on next page

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 5 Continued from previous page

Around the time of India’s Chip’s early papers (the first was an M.Sc., supervised by Ted Hannan. independence Chip’s father tried to Heathcote & Moyal (1959) were in Later in that year, Sleeba John, from move the family to New Zealand, but queueing theory, a vibrant topic in the Indian Statistical Institute, was the family settled, like many coming applied probability at the time, and appointed Senior Lecturer. Ted was from British India, in Perth. In 1946 Chip were written in the course of his PhD, still away, and I remember Chip asking had arrived in Melbourne at Carey supervised by J.E. Moyal, in Pat Moran’s my opinion about a telegram from Baptist Grammar School as a boarder. Statistics Department in the IAS, ANU. Sleeba John, which read “Latest offer His sporting prowess in all the sports, The title was “Some aspects of the expected...” The “expected” was and most specifically his excellence in theory of queues”, and the Ph.D. was meant to be “accepted”. cricket won him life membership of the conferred in 1960. MCG, which he renounced on moving In those early SGS years, the only to Perth, where he attended the Hale Chip’s CV lists under “Previous research degree supervised in the School 1947-1948. He excelled in tennis Positions” the University of North Department was a Master’s by thesis. (Winning Tennis Squad in 1948), and Carolina at Chapel Hill, 1961-1962. In Choong Kong (C.K.) Cheong, who had in cricket, and was a member of the the 1960’s and 1970’s queueing theory shared my 4 undergraduate years school’s football teams. After leaving was a thriving research area in what at the University of Adelaide, had school he worked in the bush in W.A. had come to be called “Applied come to ANU a year before me, to Probability”. A leader in queueing do an M.Sc. which was supervised by At school, apart from a nascent theory was W.L. (“Wally”) Smith, who Chip. The thesis, dated 1965, was “The interest in probability associated with was then in the dynamic Statistics Application of Results in Fluctuation running horse betting, mathematics Department at Chapel Hill, and the and Renewal Theory to Queueing had not been one of his strengths. But Heathcotes established an ongoing Problems.” In 1965 Chip was also he developed an interest, re-sat his friendship during their time there. supervising another M.Sc. student, “Leaving” exam in mathematics, and Hock Aun Cheong (M.Sc., 1967), like enrolled at the University of W.A. for a After his appointment at SGS, ANU, C.K. Cheong from Malaysia (and who Bachelor of Arts, awarded in 1956. His in 1962 as Senior Lecturer, there is a indeed, though no relation, had the foci were Mathematics and History, break in Chip’s publication output till a same surname, Cheong). Unlike C.K.’s, interests which he was to maintain clutch of papers in 1965, one of which there appears to have been no joint throughout his life. relates back to his stay at Chapel Hill publication resulting from H.A.’s thesis, (Heathcote,1965). He records in his CV which was titled “Some Applications of It was early days for mathematical a Study Leave in 1966 at the University the Saddlepoint Methods in Probability statistics in Australia. M.H. Belz had of Maryland. There he was in contact Theory.” founded the first Department of with Ryszard Syski, a Polish-American Statistics in Australia at the University mathematician, who was to become C.K. Cheong remained at the ANU to of Melbourne in 1948: this provided one of the leading names of queueing do his Ph.D. in the IAS Department undergraduate courses. C. E. theory. Chip’s sojourn at the University under the supervision of David Vere- Weatherburn (1946) had written one of of Maryland may have enabled him to Jones, who had also arrived in 1965. the first textbooks to fill a need in the complete his booklet (Heathcote, 1967). Don McNeil, who had arrived at the area. It was based on a course of sixty His publication activity in queueing same time as I (we met in Pat Moran’s lectures on Statistical Mathematics at theory continued till about 1969. office) was enrolled full-time for a the University of W.A. The book may Ph.D. in the IAS Department. C.K have provided some stimulus for the At the S.G.S., ANU, Cheong later occupied very senior young Chip Heathcote, although when positions with Singapore Airlines, and I mentioned it to him in the context of 1965-1970 with a major Singapore-based bank, my Bicentenary article, Seneta (1988), I first met Chip and Nina at the end and was honoured by ANU. he was dismissive of the book. of January 1965, when I first arrived, I had come to Canberra after newly married, in Canberra, to take Chip had met his wife to be, Nina, submitting an M.Sc., written under up a 1-year appointment as Temporary then a University of Sydney student, the supervision of John Darroch Senior Tutor in the SGS Statistics at a student camp on the Wye River in at the University of Adelaide, and Department, and to begin a PhD part- Victoria in about 1953. They married in had brought with me an interest in time in Pat Moran’s IAS Department. February, 1954 and in late 1955 moved absorbing Markov chains in general, At the time Ted Hannan was on leave, to Melbourne to pursue further study. and the simple branching process in and Chip was Acting Head. The Warren Ewens,FAA, FRS remembers particular. Harris’s (1963) now classic other staff in the department were Chip as tutor in Belz’s department, book, and photocopiers, had recently Warren Ewens, John Morgan, Paul and Chip in his CV lists “non-tenured appeared, and I had read in the Winer, Deane Terrell (who had come status” at the University of Melbourne original Russian Kolmogorov’s and from Adelaide in 1964), and Erika 1956-1957, and a 1958 M.A. from that Yaglom’s foundational papers. Frederiksdottir as Lecturers. Des university. Nicholls was a graduate student writing > Continued on next page

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 6 Continued from previous page

Joe Gani’s Journal of Applied was an unmistakable presence. In Probability (JAP) had appeared in the early years Chip smoked a pipe At the S.G.S., ANU, 1964. David Vere-Jones, who was to and was a keen skier. I remember 1971-1974 become my de-facto supervisor, was him on crutches after one weekend. The Faculty of Economics then made developing the theory for countably There was talk in the tea rooms, the building-up of the Econometrics infinite non-negative matrices both Childers St and later Copland, component of the Department a analogous to the Perron-Frobenius of “Smiggins”, and “Perisher”. The priority. Des Nicholls, who had been theory for finite ones, which was a Copland tea room was luxurious, with a part-time tutor 1969-1971 while on a basic tool for absorbing chains. tea ladies and regular morning and PhD scholarship, with Ted Hannan as afternoon tea times, so there was Chip took me under his wing supervisor, was appointed Lecturer, much opportunity for twice daily social immediately. Throughout his career it and was joined over the next four years interaction. The SGS mathematicians, was characteristic of him to nurture or so by Ray Byron, Adrian Pagan and from another building nearby, also new arrivals. He had a facility for Tom Valentine. came. One could learn much about quick academic and social rapport. what was happening at the Faculty, In 1971, R.A. (Ray) Jarvis was appointed His CV indicates that he had, even the university and the world. Senior Lecturer in Computer Science before my arrival, had an interest in within the Department, and Computer simple branching processes, so we Chip promoted social interaction Science I was offered for the first were very quickly collaborating on a through one or two wine-bottling time that year. Ray was soon joined by joint paper, published in JAP in 1966. sessions at his home. A barrel of Peter N. Cressey as Lecturer. Computer We had, however, been a little hasty wine was provided, and bottles and Science stayed as a subdepartment of in our excitement, and a correction equipment supplied. Occasionally, the Statistics till 1976. note was published a year later. But Department would go to the Canberra the highlight of our collaboration Rex on a late Friday afternoon, for Chip was Editor of the Australian was the paper Heathcote et al. drinks and billiards. Journal of Statistics (AJS) 1971-1973, (1967), where the three of us found succeeding the Foundation Editor, H.O. minimal conditions for Kolmogorov’s C.C. (Chris) Heyde, later FAA, Lancaster, (1959-1971). During this time and Yaglom’s results to hold in the joined the Department as Reader he invited Chris Heyde to write Heyde subcritical case, and had the paper in September, 1968 (Seneta & Gani, (1972) for AJS, a ground-breaking published in the premier Russian – 2009). Chip was Reader by this time. paper, and one which Chris later based probability journal. The position of Econometrics and regarded as a favourite. Chip, at the time an applied probabilist Operations Research within the SGS Chip had become Dean, Faculty of whose emphasis was stochastic Department of Statistics, and of this Economics, in 1973. He was again Dean processes (a common focus of ANU Department within the Faculty of in 1978. Ph.D.’s), held the book of William Feller, Economics had been issues for some Chris Heyde succeeded Chip as Vol. 1, in high esteem, and was excited time. Change was made feasible by Editor (1973-1978) of AJS, but left the by the imminent appearance of Vol. Ted Hannan’s moving to a second department in 1974 to join the CSIRO 2. It appeared as Feller (1966), and Chair of Statistics in Moran’s IAS Division of Mathematics and Statistics, Chip, pipe characteristically in place, Department. A decision was made to of which Joe Gani had just become declared it “marvelous stuff”. It was too advertise two Chairs in the continuing Chief. It was located just a short way difficult for me, to whom it seemed a SGS department, one in Mathematical from the Copland Building, in Hobart book written to impress Feller’s peers. Statistics and one in Econometrics. Chip Heathcote was appointed to the Place. At the time of my arrival in Canberra, first, and Deane Terrell to the second. Chip’s CV records two periods spent at the SGS Statistics Department was Chip was first Head, from 1971, of a the London School of Economics: 1969- located in one of the barracks at planned rotating Headship. 70 and 1974. The first of these would Childers St. These barracks, opposite have enabled Chip to finish his book the military drill hall building which is He had earlier been offered the Chair (Heathcote, 1971), in which he records still there, housed temporarily various of Statistics at the University of NSW. teaching in American and British, as departments of the SGS. A year or so He had declined probably because he well as Australian universities. The later, the Faculty of Economics and did not want to leave Canberra, and Preface says: Commerce, to which the Statistics possibly because he did not want to affect the chances of his great friend Department belonged, moved into the “The present text is meant for at that university, Jim Douglas. then-new Copland Building. undergraduates beginning the study of mathematical statistics and probability At Childers Street, we in Statistics after completing a full first year course shared a tea room with the Faculty on calculus and real variables. … of Law. Kep Enderby, later Attorney- General in the Whitlam Government, > Continued on next page

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 7 Continued from previous page

Material for a course of approximately not, hitherto, of technical, let alone time series analysis of economic data. thirty six lectures is contained in the mathematical, material. Chip, in In my first year as Lecturer, in 1966, the unstarred sections of Chapters 1,3, persuading me to write a book on only course I was asked to teach was 4 and 5. … a few important and non-negative matrices for the same Sample Surveys. (Both Ted Hannan and essential theorems from analysis and publisher, explained that someone Chip, who was Head of Department measure theory have been stated senior at Allen and Unwin with whom “episodically 1971-1996” had a policy without proof … An attempt has always he had come in contact (perhaps of giving maximum time for research.) been made to motivate these results through his Study Leave at LSE, and Time series analysis was Ted Hannan’s … . ” his cultural interests), was seeking to specialty, in which he had acquired broaden the output of this publisher. world fame. Des Nicholls eventually The real variable theory reflects an My book duly appeared as Seneta taught time series analysis with great applied probabilist’s inclination to (1973). There is an acknowledgement clarity, and motivated my introduction non-negative random variables, as dated 1972 to “the head of the of such an undergraduate course at reflecting physical quantities, and Department of Statistics at the the University of Sydney. consequently probability generating Australian National University, C.R. functions and Laplace transforms Heathcote, for providing moral, and From the late 1960’s Chip was being the corresponding real-variable arranging material, support ”. The evolving a research interest in the transforms, via expectations, of their person at Allen and Unwin who was empirical characteristic function (ECF) probability distributions. the idealistic driving force behind constructed on the basis of a random broadening their publishing direction sample, as a sampling reflection of Chip’s book at some remove which died in the 1970’s, and with him such the characteristic function (t) reflects initially the static approach characterizes a probability distribution. plans. However, Chip’s book was to mathematical statistics of If the underlying distribution is reissued in 2000 by Dover, New York. Weatherburn (1964), but with Chip’s symmetric about the origin, the A second edition of my book was own subsequent strong deviation to complex part of the characteristic stochastic processes. It embodies his published in 1981 by Springer. function is zero for all t, so the complex concern for good teaching, and for In his nurturing way, Chip thought part of the ECF can be used to test for reading correctly the level of students’ that I was able to supervise my first such a null hypothesis in the manner capabilities by a lecturer. 4th Year Honours project (nowadays H0 : (t)= (t0), using a fixed t=t0. The Chip taught at ANU, amongst other called by students a “4th Year thesis”) first of Chip’s publications on the ECF things, a one semester second-year in 1971, when the only 4th Year Honours is Heathcote (1972), received October course in probability. The “thirty six” student was Kathy M. Kang. I gave 30, 1970. The idea of ECF was to lectures mentioned in the Preface her the task of sorting out from the pervade the work of Chip’s students of his 1971 book clearly relate to this, standpoint of mathematical statistics and his later collaborations. Some of since 36 corresponds to 3 per week the controversial topic: Path Analysis, the key ideas already appear in this for 12 weeks: about the length of a and she did it brilliantly, under difficult paper, namely: testing for location semester’s teaching time. The year conditions. Our joint paper was of a symmetric distribution; the 1974 saw entry into our second-year delayed until 1978. relationship of an ECF- based statistic mathematical statistics courses of with the Cramér-von Mises test for Although the SGS Statistics a cohort of National Undergraduate testing a specific form of distribution Department was in the Faculty of Scholars which included Iain Johnstone, function: H0 : F(x) = F0(x); and the Economics, its courses were also Barry Quinn, Rosemary Karmel and idea of studying the long-tail structure available to students in the Faculties Mark Ramsey. Chip at the time was of distributions with such a long-tail of Arts and of Science. The first year enthusiastic about the Keller Plan of structure, but with a simple structure courses were largely of a service teaching, which enabled talented of the characteristic function (such nature, with differentiation between students to proceed at their own pace, as that of the symmetric stable law) statistical and econometric directions. but it was demanding on staff time through the ECF. A number of undergraduate students because of its 1:1 contact nature of were from the ABS, then housed in a The first successor paper to Heathcote students and staff. Chip recognized building “just down the street” from (1972) with a student was Heathcote this, and it was dropped within a year. the Childers Street barracks. Cadets & Pitman (1972), received 30 July, 1971. Barry Quinn, Professor of Statistics at or employees sought to complete a J.W. (Jim) Pitman, son of the great Macquarie University, recently showed degree with one major in Statistics; Australian mathematical statistician me his (second) well-worn copy of and other students on completion E.J.G. Pitman, was among the talented Chip’s book. Iain Johnstone has long of 4th Year Honours went into students who passed through the SGS been Professor in Statistics at Stanford employment with the ABS, where at Statistics Department., and went on to University. the time two of the leading statistical glittering careers in Statistics. The book was published by George figures were Ken Foreman and Ken Allen and Unwin, London, a very Brewer. Focus was on sample surveys > Continued on next page well-known “literary” publisher, but (sampling from a finite population) and

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 8 Continued from previous page

of Mathematical Statistics, University conservative persuasion, a bit like my At the S.G.S., ANU, of Sydney, to explore possibilities of Dad in that respect. ...I was only 21 1975-1979 graduate work. He eventually began when I started working with him. ….” In June, 1975 Raymond (Ray) Chambers research in the SGS Department under Chris Heyde, and enrolled for an M.Sc., Chip’s Study Leave was at submitted his MSc (Statistics) thesis, Mathematisch Zentrum, Amsterdam, written under Chip’s supervision. submitted two years later while he was in Oxford completing a D.Phil. under 1979, and University of California, Its title was A Linear Model with Berkeley, 1979-1980. Stable Residuals, and it was about John Kingman, FRS. Peter returned to asymptotics when the residuals in a the Department as Lecturer, in 1978. Brenton continues: linear regression follow a symmetric Terry O’Neill, an Adelaide graduate and Stanford Ph.D. also joined in 1976, “Peter’s contribution was to get me stable law. Chip asked me (I expect to look up results on limit theorems he thought it was time) to be one and with Peter Hall gave it a boost in research in mathematical statistics. since he was doing his book on such of the two examiners. It was my first with Chris Heyde. But Peter’s greatest experience of examining a research Closely following Chambers & attribute was a positive attitude re thesis, although I had already Heathcote (1975) was the paper of research. I think we all benefitted from supervised Malcolm Quine’s Ph.D.. Feigin & Heathcote (1976), with its Chip’s running of the department even The two papers: Chambers & title clearly describing one of the though I was too young to appreciate Heathcote (1975, 1978) were largely a issues raised in Heathcote (1972). Paul that at the time.” Feigin’s (Ph.D., 1975) thesis: “Maximum product of work associated with the The title of Brenton’s PhD dissertation thesis, motivated by the fact that for Likelihood Estimation for Stochastic Processes- A Martingale Approach”, is “Robust Estimation: Limit Theorems distributions such as the symmetric and Their Applications”. M-estimators, stable, with a simple characteristic acknowledges both Chip Heathcote and Chris Heyde as supervisors, and a generalization of maximum likelihood function, moment-based inference via estimators, and robustness, all come the ECF was feasible. Their Biometrika Gani (2005) indicates joint IAS/SGS supervision. within the ambit of approach of Chip’s paper, Chambers & Heathcote (1981), ECF interests. some years later, was a culmination Brenton Clarke (Ph.D., 1980) in an initial inasmuch it pointed out that a loss email to me of August 2016 mourns the Chambers & Heathcote (1981) was function using the ECF could be passing of both his supervisors, Chip immediately followed, in the same constructed. Therefore estimation of Heathcote and Peter Hall, within such a volume of the same journal, by one location and scale parameters was short space of time. with another of Chip’s students, in a possible, in rough analogy with simple similar vein (Heathcote & Silvapulle linear least squares when the residual In a condolence email, of 4 August (1981)). Mervin Silvapulle (PhD 1981) lists distribution had moments. 2016 to Chip’s daughter, Sarah, on as his supervisors Chip Heathcote and Chip’s passing, Brenton had written: Terry O’Neill. His thesis title was “The Ray Chambers was still with the ABS Minimum w2-Method of Estimation.” in the early 1980’s. He was awarded “I knew him … when he took me on to a Commonwealth Public Service be a tutor in his department and then In the year 1981, the Hungarian Scholarship Tenable Overseas in to do a PhD with him. We used to meet mathematician Sandor Csörgö was a 1979, with very strong support from once a week to talk as he had a very Visiting Fellow in the Department of Ken Foreman, and went to Johns busy schedule as Dean of the Faculty…. Statistics, The Faculties, from 14 July Hopkins University, Baltimore, where I remember getting frustrated because to 18 September. Csörgö had been he completed a Ph.D. (1982) under the PhD problem I had decided to one of several authors interested in the supervision of Richard Royall. The work on was not giving me numerical and publishing on the ECF. He may title was “Robust Finite Population values that made sense, even after have been the referee for Chambers & Estimation”. Ray writes: one year, …it worked out well in the Heathcote (1981), on which the invited end as he took a long sabbatical from paper Heathcote (1982), in a Special “My research there was very much ANU … and came back to advise me Issue honouring P.A.P Moran, focussed. influenced by my training in outlier well on what to leave out of my thesis, Contact had been established and robust inference that I had received so that I could submit my doctorate. Chip had arranged the visit. It from Chip. … ” I was very appreciative of the PhD was spectacularly successful. The Chip and Ray Chambers were both referees that your Dad had lined up for paper Csörgö & Heathcote (1982), “refugees” from the British Raj in me while he was away. The problem received 30 October 1981, presents India, which enhanced their academic that did not work out in the first year the estimation theory (free of its linear closeness. was solved and it became the last regression setting) of Chambers & example chapter of my thesis. ...in 1991 Heathcote (1981) and Heathcote (1982) Peter Hall had come to ANU in [I] worked with your Dad on a paper (at that time still to appear) within December 1973, having completed his which extended the last chapter of my 4th Year Honours in the Department thesis …Your Dad was a kind man of > Continued on next page

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 9 Continued from previous page a context of a limiting Gaussian preprint-available, paper of Csörgö University of California, Santa Barbara, stochastic process, with a covariance (1983), played an important part in of Ross Cunningham and Joe Gani, matrix expressed in terms of the real Alan’s research. with Joe titled Visiting Fellow at ANU, and complex parts of a characteristic and agreeing to do some teaching. In the Preface to his thesis “Some function, where “t” is a time parameter. Then a further Visiting Fellowship, Problems in Adaptive Estimation”) The full highly technical detail is in the- again “without grant”, for 6 months Alan writes: “… it is known that the tail then to appear Csörgö (1983). from January, 1990; and similar for behaviour of a distribution function 12 months from 1st September, 1991. I left the Department in mid-1979 is reflected in the behaviour of the These visits were formally facilitated as Reader (from 1974) to go to the characteristic function near the origin by Chip. Joe was at the time transiting University of Sydney as Professor and ...Here we will mainly concentrate on to retirement amidst his old ANU Head, Department of Mathematical the empirical characteristic function friends, especially Ted Hannan and Statistics. Alan Welsh was a student approach … ” Chris Heyde, from Santa Barbara, in my Sydney department at the time. where he was, in his last years there, He obtained First Class Honours, with There is a break in Chip’s publication spending only about 3 months each University Medal in 1981. record between 1983 and 1987, perhaps related to his continuing recovery from year. Subsequent requests for The “School of General Studies” at his first triple by-pass operation. It extension of Joe’s Visiting Fellowships ANU was renamed “The Faculties” then picks up in 1987, resuming his were taken up by Chris Heyde. Chip from 1980. collaboration with Sandor Csörgö on spent a period of Study Leave at Santa symmetric distributions, and with Alan Barbara in 1992. Welsh on functional least squares. Department of In 1992 the Mathematical Statistics Statistics, The By 1982 the size of the Department group, headed by Heathcote as had increased to thirteen (Gani, Professor, in the Department included Faculties, ANU, 2005), with Heathcote and Terrell as D.F. Nicholls as Reader, and R.B. Professors. Cunningham, T.J. O’Neill, and A.H. 1980-1996 Welsh as Senior Lecturers. By 1980 Chip had given up smoking. His In 1988 Peter Hall, through intensive angina had become so severe that he efforts by Chip Heathcote, was Late in 1987 the Head of Actuarial could scarcely walk from the carpark appointed Professor. In “The Faculties” Studies at Macquarie University to his ANU office. Given weeks to live, promotion to Professor at the time was approached the Faculty of Economics he had triple bypass heart surgery, it a very rare event. From 1986 Peter and Commerce. The Department of seems in 1981 when he was about 50, spent half of each year at IAS, and Statistics was enthusiastic, and Chip at Royal Prince Alfred Hospital, Sydney. was involved in the fledgling Centre and Des Nicholls designed a course It was early and anxious days for this for Mathematical Analysis. In 1991 he for a B.Ec.in Actuarial Studies, in kind of operation. After the operation reverted to a full time appointment which Macquarie University would Chip became an avid walker. in The Faculties, with secondment to participate. The course proceeded the Centre of Mathematics and its successfully till 1992, when Macquarie Alan Welsh (Ph.D., 1984) had arrived Applications. University announced that it would no early in 1982 to begin his PhD. having longer participate in 1993 and beyond. intended to work with Chip on the Ray Chambers left the ABS in 1983, An arrangement was then made for the ECF. In 1982 ANU was starting to have but stayed in the Australian Public University of Melbourne to teach the a supervisory panel of three for its PhD Service (APS). In 1991 there was a actuarial units by videoconferencing. students. Alan’s panel consisted of Senior Lecturer position open in the In 1996 ANU introduced an Honours Chip, Des Nicholls and Peter Hall. SGS Statistics Department for which, programme in Actuarial Studies, and in at Chip’s suggestion, he successfully Chip, recovering from his triple bypass, 1999 the Faculty decided to run its own applied as a straight transfer from Actuarial Studies course. used to come in in the mornings, and APS. In 1995 there was an offer from then go home to rest. In 1995 the Department was renamed Southampton University, at the time Alan’s first paper, following on from the Department of Statistics and building considerable strength in Econometrics, according to Gani ideas of the Chambers and Heathcote Statistics, and with Chip’s blessing papers, was with Peter, on a test for (2005). By then Chip had been Ray made that jump. He is presently succeeded as Head. normality. His first paper joint paper Professor of the University of with Chip (Heathcote & Welsh, 1983) Wollongong. In publications from 1988 to Chip’s applying the Chambers and Heathcote retirement in 1996 there is a decided regression ideas to a time series For the period mid-September, 1989 turn to epidemiology: issues of disease, context, autoregression, which was – 31 December 1989, there was an Chip’s initial proposal for a thesis topic, exchange between the Department > Continued on next page was next. The then still-to-appear, but and Joe Gani’s department at the

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 10 Continued from previous page mortality, life expectancy, insurance, 1. “Twilight of the Raj: India in the and in particular coronary heart Retirement Decade Before 15 August 1947.” disease. There is a publication gap until Chapters 3 and 4 of Heathcote and 2. “First Shots in the Cold War: The Chip includes in his CV the following Higgins (2001a, 2001b). Warsaw Uprising of 1 August 1944.” periods spent in France: INSERM, Montpellier, summer 1992; INED, Paris, These begin a sequence of papers Sarah told me that Chip’s political winter 1995-1996. In an email dated which continue epidemiological views, over his lifetime, had run the 19 September, 2016, sent to iMaCh- themes of a decade or so before full spectrum, beginning with left- users, and copied to Sarah Heathcote, Chip’s retirement. Notable is wing; and in the end were, according Nicolas Brouard writes: continuing collaboration with Terry to Joe Gani, very conservative. In O’Neill, Tim Higgins, Brett Davis, Borek the last years of his life he pondered “During the early 90’s Chip travelled to Puza, and Steven Roberts, all but the purpose and meaning of human France … He spent a semester at INED perhaps one members of what had existence, in some intellectual sense at our research unit been until 2001 called the Department and measure perhaps reconnecting to the religiosity of his childhood ‘Mortalité, santé, épidémiologie’ and of Statistics and Econometrics, when environment. we worked on mortality in France it was merged with the Department of as well as estimation of confidence Finance to form the School of Finance intervals in interpolated Markov and Applied Statistics, with Terry Reflections and Last chains.” O’Neill as its Head. According to Gani (2005), the School’s staff in 2002 Things Brouard and his family visited Chip at consisted of eighteen members, nine These notes are written to illuminate ANU. Brouard says of Chip at this time specializing in Statistics and Actuarial the background to the previous that Chip couldn’t be more happy, and Studies, namely O’Neill and Nicholls as sections. remembers him as saying that if one Professors, M.A. Martin as Reader, D. Chip and Nina’s children, Robin and has gardening at home and reading at Pitt, D. A. Service and S.E. Stern as Sarah Heathcote, were born in 1961 the university, little more is needed in Senior Lecturers, T. Higgins and B. Puza and 1970 respectively. Chip has 4 life. Brouard continues: as Lecturers, and Helen Johnson as grandchildren through Robin. Sarah Associate Lecturer. “For several years, he and his wife Nina left Canberra in 1995 to work in spent springtime and early summer in In about 2006 it became the School of Geneva. She now continues the ANU Europe, flying back to the Southern Finance, Actuarial Studies and Applied connection as Associate Professor in Hemisphere to avoid cold time in Statistics, and in about 2009 the Law. Europe … probably the best time to live Research School of Finance, Actuarial John Besemeres and his wife Anna on this Earth … He is one of the main Studies and Applied Statistics. In 2016, Wierzbicka, soon after their arrival authors of our iMaCh software and “Applied” was dropped from the title. from Warsaw early in 1973 to pursue articles related to it ...” The last of Chip’s papers to appear academic careers at ANU, met Some of Chip’s contribution in this was Heathcote et al. (2009). Thus he the Heathcotes through Nina, who respect is encompassed in Lièvre et al. had been publishing over a period of had developed a specialty in the (2003). 50 years. politics of the European Community. and was a Research Fellow in the His last paper published before his Statistical issues in health science were Political Science Department of the retirement is Heathcote et al. (1995). only a part of his retirement activities. Research School of Social Sciences. He indulged in many interests. His Sarah Heathcote remembers, of some This Department was housed in the daughter Sarah told me about a years before his retirement, Chip Coombs Building, where Nina and John new-found interest in pottery, and the waking at 4am to work, and Chip had offices close to each other. University of the Third Age where he taking 8am classes. These were first took courses on Kipling (a reconnection John, an expert on political aspects year classes, the good teaching of with his Anglo-Indian heritage) and on of demographic developments in which the Enlightenment. the Soviet Union and the communist Chip regarded as very important. Alan countries of Europe, and a gifted There were two talks, of which he Welsh took over these classes in 1994. linguist, continues in an email shared the materials with me, to the He remembers Chip mentioning taking attachment of 15 August 2016: Probus Club of Canberra, in August his dog for walks along Red Hill in the 2010 and August 2012 respectively. “Though Chip had the gravitas of a very early mornings. These were historical with background professor, I remember finding him Chip retired at the end of 1996. From connection to his own and Nina’s – despite his quiet characteristic January, 1997, he was Emeritus Prof- background. Both bear the notation: essor and Visiting Fellow, Centre for “C.R. Heathcote, for the family.” The > Continued on next page Mathematics and its Applications, ANU. respective titles are:

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 11 Continued from previous page reticence – very warm and He had a heart attack on 26 January References communicative from the outset. … 2016. Sarah Heathcote writes (7 Feb., Chambers, R. L. & Heathcote, C. R. (1975) A He seemed to read widely … I quickly 2017): “About a month later he fell linear model with errors lacking a variance. realised that it was important to have and fractured his ‘C2’ disc, commonly I. Austral. J. Statist. 17, 173–185. a bottle of respectable whisky for a known as a broken neck.” Chambers, R. L. & Heathcote, C. R. (1978) A post-prandial … He also enjoyed a linear model with errors lacking a variance. glass or two of red wine, and … in his Joe Gani died on 12th April, 2016, and II. Austral. J. Statist. 20, 161–175. last months ... [I] took one of my better a Memorial Service for him was held on Chambers, R. L. & Heathcote, C. R. (1981) bottles to leave with him in memory of 22nd April in the Great Hall, University On the estimation of slope and the identification of outliers in linear regression. happier times. House, ANU (Seneta, 2017). At the refreshments following, I was told that Biometrika 68, 21–33. One of the pleasures of my retirement Chip had been there, with the walking Clarke, B. R. & Heathcote, C. R. (1994) was that I had more time to see Chip frame on which he had become reliant Robust estimation of k-component univariate normal mixtures. Ann. Inst. and Nina, and we would meet them since Nina’s death. He had looked very Statist. Math. 46, 83–93. sometimes for lunch. Chip by then unwell, and Sarah had taken him home had gone through his second bypass before I could see him. I spoke with Csörgö, S. (1983) The theory of functional least squares. J. Austral Math. Soc., Ser. A, operation [about 2003], and while he him later that day, for the last time, 34, 336-355. made light of it, it was evident that he by phone. As usual he made light of Csörgö, S. & Heathcote, C. R. (1982) Some felt it had been less successful than it, summarizing by saying that he had results concerning symmetric distributions. the earlier one.” broken his neck, but that this was not Bull. Austral. Math. Soc. 25, 327–335. as bad as it sounded. There would be I (ES) used to visit Canberra about Feigin, P. D. & Heathcote, C. R. (1976) The no time to meet since Sarah was about empirical characteristic function and the once a year to see old friends, and to take him for a holiday. Cramér-von Mises statistic. Sankhyā, Ser. A. visits to ANU were always followed 38, 309–325. by a meeting of Joe Gani’s informal Ray Chambers and his wife, Pat, visited Feller, W. (1966) An Introduction to coffee club. I never found Chip there, Chip at home in Forrest in May of 2016. Probability Theory and Its Applications. Vol. and so we drifted apart. Ray writes in an email to me: “He was 2, Wiley, New York. his usual gentle civilised self then, frail, But before one of my visits to ANU, in Gani, J. (1994) Obituary: Edward James but very determined to keep going.” Hannan. Austral. J. Statist., 36. 1-8. 2011 or 2012, I told Joe that I would like to see Chip again. I was happy that Chip, like Nina, died in his sleep in their Gani, J. (2005) Fifty years of Statistics at the Australian National University. Historical he came, by car directly to the coffee bedroom at their home. The official Records of Australian Science. 16, 31- 44. house. cause of Chip’s death on 18 July, 2016 Harris, T. E. (1963) The Theory of Branching was heart attack. Whether the neck, Nina died in 2012. Chip, Sarah and Processes. Springer, Berlin. which didn’t heal, played a role is family organised a commemoration in Heathcote, C. R. (1965) Divergent unclear. And so: their beautiful garden. single server queues. In: Proc. Sympos. Congestion Theory (Chapel Hill, N.C., 1964). The tumult and the shouting dies – It was a mild sunny day, and Chip pp. 108–136. Univ. North Carolina Press, delivered a eulogy, after climbing The Captains and the Kings depart – Chapel Hill, N.C. onto a garden bench by way of an Heathcote, C. R. (1967) Complete improvised podium. Recessional. Rudyard Kipling Exponential Convergence and Some (1865-1936). Related Topics. Methuen’s Supplementary After that Chip and I had an annual Review Series in Applied Probability, Vol. 7. Methuen, London lunch in Canberra, four in all, always Eugene Seneta over wine, generally his own, generally Heathcote, C.R. (1971) Probability: Elements of the Mathematical Theory. George Allen in a favourite restaurant. He would and Unwin, London. arrive to pick me up from New Acton Acknowledgements Heathcote, C. R. (1972) A test of goodness punctually at 12 midday, and had of fit for symmetric random variables. a walking frame in the car which My grateful thanks for memories, Austral. J. Statist. 14, 172–181. information and materials are due to he needed after arrival near the Heathcote, C. R. (1982) Linear regression by restaurant. He had another health Sarah Heathcote, John Besemeres, functional least squares. J. Appl. Probab. issue developing at the time of the first Ray Chambers, Des Nicholls, Brenton Special Vol. 19A, 225–239. (Essays in two lunches, and I was advised not to Clarke, and Alan Welsh. Statistical Science.) delay my next visit for too long. But Heathcote, C.R. & Higgins, T. (2001a) when next we met, medical or dental A regression model of mortality, with intervention had been successful, and application to the Netherlands. In Forecasting Mortality in Developed he was quite cheerful. After the last Countries. eds. E. Tabeau, A. van den Berg of our lunches as we talked at Chip’s Jeths, and C.R. Heathcote. pp. 59-82 home over wine, he was effusive over (Chapter 3). Kluwer, Amsterdam. the electronic tablet which Sarah had got him. > Continued on next page

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 12 Continued from previous page

Heathcote, C.R. & Higgins, T. (2001b) Forecasting mortality from regression REGISTER NOW FOR OUR WORKSHOP models; the case of the Netherlands. In Forecasting Mortality in Developed Countries. eds. E. Tabeau, A. van den Berg Jeths, and C.R. Heathcote. pp.83-104 Developing Your Career to (Chapter 4). Kluwer, Amsterdam: Heathcote, C. R. & Moyal, J. E. (1959) The Thrive in a Data-rich, random walk [in continuous time] and its application to the theory of queues. Technology-driven, Reproducible Biometrika 46, 400–411. Heathcote, C. R. & Pitman, J. W. (1972) An inequality for characteristic functions. Bull. Research Environment Austral. Math. Soc. 6, 1–9. with presenter Professor Di Cook, Department of Econometrics and Heathcote, C. R., Puza, B. D. & Roberts, S. Business Statistics , Monash University P. The use of aggregate data to estimate Gompertz-type old-age mortality in held on 25 September 2017 in Coolangatta, at the Oaks Calypso Plaza, heterogeneous populations. Australian and New Zealand J. Statist. 51 (2009), no. 4, 99 Griffith St, Coolangatta. 481–497. Professor Di Cook is a Fellow of the American Statistical Association, and Heathcote, C. R., Rachev, S. T. & Cheng, Ordinary Member of the R Foundation. Her research is in data visualisation, B. (1995) Testing multivariate symmetry. J. Multivariate Anal. 54, no. 1, 91–112. exploratory data analysis, multivariate methods, data mining and statistical computing. She has developed methods for visualising high-dimensional data Heathcote, C. R., Seneta, E. & Vere-Jones, D. (1967) A refinement of two theorems in using tours, projection pursuit, manual controls for tours, pipelines for interactive the theory of branching processes. Teor. graphics, a grammar of graphics for biological data, and visualizing boundaries Verojatnost. i Primenen. 12, 341–346. in high-d classifiers. She has experimented with visualising data in virtual Heathcote, C. R. & Silvapulle, M. J. (1981) environments, participated in producing software including xgobi, ggobi, cranvas Minimum mean squared estimation of and several R packages. Her current work is focusing on bridging the gap location and scale parameters under between statistical inference and exploratory graphics. She is currently doing misspecification of the model. Biometrika experiments using Amazon’s Mechanical Turk, and eye-tracking equipment. Some 68, 501–514. of the applications that she has worked on include backhoes, drug studies, mud Heathcote, C. R. & Welsh, A. H. (1983) crab growth, climate change, gene expression analysis, butterfly populations The robust estimation of autoregressive processes by functional least squares. J. in Yellowstone, stimulus funds spending, NRC rankings of graduate programs, Appl. Probab. 20, 737–753. technology boom and bust, election polls, soybean breeding, common crop Heyde, C.C. (1972) Martingales: a case population structures, insect gall to plant host interactions, Melbourne pedestrian for a place in the statistician’s repertoire. traffic sensors, soccer and tennis statistics. Invited Paper. Austral. J. Statist., 14, 1-9. The workshop will be about Lièvre, A., Brouard, N. & Heathcote, C. (2003) The estimation of health 1. Reproducible research, for publications, talks, and web sites using Rmarkdown expectancies from cross-longitudinal surveys. Math. Popul. Stud. 10, 211–248. with the R ecosystem.Many journals now require you to submit code to reproduce the results reported in your paper, and this will become the norm Seneta, E. (1973) Non-Negative Matrices. An Introduction to Theory and Applications. over the next few years. Maintaining a public profile is important for many George Allen and Unwin, London: careers, and new tools using hugo enable web site construction, including Seneta, E. (1988) Silhouettes in early blogs. Australian Statistics. Austral. J. Statist., 2. Why and how to organise data. Concepts of tidy data, and learning to 30(B), 2 – 22. rearrange data will be covered. It is often said that the data cleaning stage, [Bicentennial History Issue, C.C. Heyde and often neglected in statistics education, takes 95% of your time. Being efficient E. Seneta, eds.] in data handling can allow you to spend more time thinking about the problem Seneta, E. (2017) Obituary: Joseph Mark to be solved. Gani A.M. DSc FAA FASSA. J. Appl. Prob., 54, 1-11. 3. Making effective plots, grammar of graphics, good practices: Mapping data Seneta, E. & Gani, J.M. (2009) Christopher to graphical elements in plots using ggplot2. Simple plots, scatterplots, bar Charles Heyde. Historical Records of charts, time series, profiles, boxplots. Using cognitive principles to improve Australian Science, 20, 67-90. plots. Advanced graphics, layering, maps, interactivity: Layering different data Weatherburn, C. E. (1946) A First Course sets, drawing maps, exploring model fits, multivariate plots. Simple interactive in Mathematical Statistics. Cambridge graphics. University Press, Cambridge. For more information and to register please check out our website: http://www.statsoc.org.au/events/ssai-events/cpd85-developing-career- thrive-data-rich-technology-drive-reproducible-research-environment/.

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 13 QLD Branch SSA-QLD Career Pathways seminar

Left to Right: Ms Lee Jones, President of SSA-QLD branch; Ms Angela Emblen-Reeves and Ms Jo’ Anne Langham, Australian Tax Office; Mr Chris Galvin, Queensland Health; Prof Gita Mishra, The University of Queensland; Ms Jeeva Kanesarajah, 2017 SSA-QLD Young Statistician Representative.

The SSA-QLD Career pathways seminar was held on the 1st of August 2017 with a record turnout of 50 early career statisticians from universities, industry and government agencies. Ms Jo’Anne Langham and Ms Angela Emblen-Reeves spoke about their experiences as analysts and employers in the Australian Tax Office. The spoke about the “Smarter Data” initiative by the ATO, and the need of analyst to be able to translate key findings to suit a lay audience. “Numbers resonate with people who are making decisions” Ms Langham noted. Mr Chris Galvin from Queensland Health spoke of his diverse background and being able to link his clinical background with his statistical work gave him a competitive edge in his career. He highlighted the importance of being able to meet work deadlines and being reliable as an important aspect government employees need for career progression. Professor Gita Mishra from The University of Queensland discussed the importance of working with great people as a stepping stone during the course of her career. She highlighted the need for statisticians to develop content knowledge of their research area e.g. biology, epidemiology, beyond statistics alone. She emphasised learning as a life-long process, and the importance of statisticians in research to be first authors or senior authors for career progression as an academic statistician. She spoke of the rise of social media and how researchers need to use this to their advantage as a method for research translation, maximizing research impact, and to communicate the importance of the research to the broader public. All presenters emphasised the importance of not only having subject knowledge but also the importance of interpersonal skills, the ability to write and speak well. There was robust discussions at the Q & A, followed by a networking session.

Jeeva Kanesarajah

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 14 VIC Branch SA Branch Victorian Branch July South Australian SSA 28 June 2017 Meeting meeting: An actuary, a data scientist and DAGs, a statistician walk Counterfactuals into a bar… and Data generation Ever wondered what the difference is between an Actuary, a Data Scientist and a Statistician? Murthy Mittinty process

Actuaries do it until death, disability or The speaker at the June 2017 meeting of the SA Branch was Murthy Mittinty from withdrawal the School of Public Health, the University of Adelaide. His primary methodological Data scientists do it with models research was to distinguish between association and causation in the health Statisticians probably do it. sciences and the study of the mechanisms by which causal effects arise. His talk On July 27, the Victorian branch jointly “DAGs, counterfactuals and data generation process” was a piece of his current hosted an event with Data Science methodological research work. In his presentation, Murthy raised the point about Melbourne and the Actuary Institute, longitudinal data which brings many challenges to modelling. He illustrated some where a member from each profession of these challenges using directed acyclic graphs, and also presented some recent (plus someone who identified as methods that have been proposed using counterfactuals theory to study such data all three!) discussed what it is that generating mechanisms. they actually do. Amanda Aitken At the beginning of his talk, Murthy described the Directed Acyclic Graph (DAG) represented the actuaries, Gareth in detail with Simpson’s paradox, local directed Markov properties, Global Markov Dickenson the data scientists, and the properties and factorization of joint densities according to the structure of DAGs. Victorian Branch’s very own Sandy Murthy also discussed the G-methods: the marginal structural models (MSM, Clarke represented the statisticians. G-formula), MSM-inverse probability treatment weights (MSM-IPTW) and the In addition, Mika Alo talked about his structural nested mean models (G-estimation) that have been suggested using work as an actuarial-data scientist- counterfactuals theory. At the end of the DAG discussion, he illustrated with an statistician. Through their entertaining example-DAG of a real world problem: the causal relationship between obesity and informative talks, each of the and pre-eclampsia, highlighting the time dependent confounding using joint, speakers gave an insight into how they marginal and conditional distributions. In conclusion, the DAG should be used ended up in their profession, what cautiously, otherwise mistakes can be made and conclusions can be adversely they enjoy about it, and what it is that affected, he said. they spend their days doing. Although these three careers may sometimes be A dinner was held right after the meeting at an Indian Restaurant in Adelaide. thought of as being somewhat distinct, Shahid Ullah the message that we’re far more similar than we are different came through strongly.

Jessica Kasza

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 15 From the Office SSA Job Board Things at the SSA Office are usually Have you seen the SSA Job Board (http://careers.statsoc.org.au/) yet? It was pretty quiet, but for the past few added to our website just over three years ago and while it had a slow start, people months it’s been intense, with have certainly caught on now. The Job Board currently has 87 registered employers preparations for YSC2017, hosting and 172 registered job seekers. 52 resumes have been uploaded and can be viewed several workshops and getting to the by potential employers. tender stage for the refurbishment of our website. Many meetings, usually Our Job Board has many helpful features, such as tips on how to prepare a resume, by teleconference, are required for the what NOT to include in your resume, how to prepare for your interview, how to day-to-day running of the Statistical advance your career and how to manage in a digital world (http://careers.statsoc. Society, but as events like our Young org.au/jobseeker/resources/). Job Seekers can create job alerts so that they get Statisticians Conference approach, notified when a new ad is listed. we’re suddenly looking at weekly Employers can view the uploaded anonymous resumes for free. If they see one that meetings, which certainly keeps me they find interesting, they can contact the Job Board managers to let them know on my toes. Our teleconferences are that they would like to contact the person who posted the resume. The Job Board always a good opportunity to get to managers will then get in touch with that person and if he or she would like to know our members- even if it’s just follow it up, they will connect employer and potential employee for a fee of $35.00, their voice. payable by the potential employer. As mentioned before tenders have From the (passive) job seekers’ view they have nothing to lose. Their resume just sits gone out to about ten different there waiting for the right employer to see it. website designers, asking for submissions of proposals for a new look We are lucky to have this great tool on our website. It did not cost the Statistical for our website. At this stage we have Society anything to have the Job Board installed. We now share the fees with the received three proposals and I can Job Board managers, “Your Membership”, based in the US. Our Job Board will not already tell that the selection process make the Society rich or even contribute significantly to its income. However it does will be very difficult indeed, because provide an extra service for our members and I hope that one day it will develop the different companies all have so into the hub for all things to do with the statistical job market. much to offer. This is a very exciting time and I can’t wait to see the final product. Hopefully we’ll have a new website up and running within the first quarter of next year, but before this can happen a lot of people will be working awfully hard to make it all happen. We can only hope that our members will like what we will have in stall for them. As new tasks are added to my list, there’ll be one less for me in the future: writing a “From the Office” column for our quarterly newsletter. It’s been sometimes challenging coming up with topics to write about, but at the same time I’ve liked the idea of being able to call myself a bit of a newsletter “columnist”. I suppose if I miss this task terribly I can always take it up again for the eNewsletter. Thank you to all the readers who have supported the SSA quarterly newsletter for so many years. Marie-Louise Rankin Executive Officer

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 16 35% SSAI Member Discount Promotion with Wiley

Members of the Statistical Society receive a special discount of 35% on general online purchases with Wiley. A discount of 25% applies to textbooks. This offer excludes school books. To take advantage of this benefit, please go to http:// au.wiley.com/WileyCDA/ , click on My Account at the top of the screen and start shopping. At check-out please enter the Promo Code SDP92 to activate the discount.

For assistance, please contact Wiley: Toll free phone (from within Australia only) 1800 777 474 Toll free phone (from New Zealand only) 0800 448 200 Other overseas phone + 61 7 33548455 E-mail: [email protected]

STATISTICAL SOCIETY OF AUSTRALIA NEWS • SEPTEMBER 2017 17