«STATISTICS FOR POSTGRADUATES AND RESEARCHERS IN OTHER DISCIPLINES: CASE STUDIES AND LESSONS LEARNED John A Harraway Department of Mathematics and ...»
ICOTS8 (2010) Invited Paper Harraway
STATISTICS FOR POSTGRADUATES AND RESEARCHERS IN OTHER DISCIPLINES:
CASE STUDIES AND LESSONS LEARNED
John A Harraway
Department of Mathematics and Statistics, University of Otago, New Zealand
Postgraduates and researchers in many disciplines use advanced statistics procedures. Statistics backgrounds often extend to at most an introductory course on statistical methods. Effective ways of providing training in these advanced procedures must be found. Emphasizing content, prerequisites and target groups, a summary of specialized courses offered at this level over the last two years and advertised internationally is presented. Then local four day intensive workshops on advanced topics for ecologists are described. These workshops draw on research contexts familiar to participants and use appropriate software. Menu driven packages or self written programs may be used. Participants in the workshops can bring their own data or data are chosen from their discipline. The teacher is introduced to the researchers which may result in future collaboration.
Student evaluations of the workshops are reported leading to recommendations for further training.
A survey of 913 research graduates in employment with PhD and Master degrees in the biological sciences, psychology, business, economics and statistics from all New Zealand Universities (Harraway & Barker, 2005) identified gaps between techniques learned at University and techniques used in the workplace by these graduates. The respondents reported deficiencies in statistical preparation and made a series of recommendations about how to remedy the gaps. There was a consensus for focused workshops to be developed by education establishments or industry on a wide range of topics including regression and generalised linear models, multivariate methods, survey design and power analysis, clinical trials,new statistical software and Bayesian methods.
Methods proposed for implementing this process included the expansion of advanced statistical service courses within universities, the development of specialist workshops in appropriate contexts for groups of postgraduate students in these areas and workplace retraining.
Some taught workshops are discussed in the next section. This includes a description of several sets of workshops provided on the web during 2009. One of the programmes involves a complete set of courses which either could lead to the equivalent of a degree in statistics or could have courses chosen selectively to support specialist subjects. Other programmes described are provided either in the home university of a particular group of students or the workplace. In the third section five workshops taught for ecologists within their university will be discussed in detail and accompanied by course evaluation comments by those participating leading to a summary of lessons learned for the future.
The courses or workshops provided outside regular university teaching which are aimed at promoting professional development for researchers appear to fall into one of three categories;
those that are introductory, those provided for researchers in a wide range of applied areas where only one introductory course has been taken and thirdly those that represent an advanced specialised topic often aimed at the statistics specialist or a person in another subject who has experience in some advanced or specialised statistics activity. Generally, a mathematical background is not needed but for some of the advanced courses facility with theoretical statistics is a benefit if not essential.
In C. Reading (Ed.), Data and context in statistics education: Towards an evidence-based society. Proceedings of the Eighth International Conference on Teaching Statistics (ICOTS8, July, 2010), Ljubljana, Slovenia. Voorburg, The Netherlands: International Statistical Institute. www.stat.auckland.ac.nz/~iase/publications.php [© 2010 ISI/IASE] ICOTS8 (2010) Invited Paper Harraway
A REVIEW OF STATISTICS WORKSHOPS AND ON-LINE COURSESThis review is not an attempt to provide a complete report on statistics courses on offer by way of workshops in the two years beginning January 2009. The aim is to give a sample overview on what is on offer as a way of identifying appropriate material for inclusion in future workshops.
There are differences between those courses taught to specialist groups of students within a university and those offered to a general audience through, for example, the web as part of professional development. There is usually a substantial charge for courses of both types although those aimed at students within a home university generally do not cost as much and sometimes may be built into a research programme or seminar series.
The Postgraduate Statistics Centre at Lancaster University has a list of professional development courses at different levels on offer for 2010. (www.maths.lancs.ac.uk/psc) For a general audience a series of two day courses include an Introduction to Bayesian methods, Data Mining, Structural Equation Modelling, Methods for Missing Data, Generalized Linear Models, Genomics, a variety of courses on R software, STATA and SPSS and an Introduction to Statistics for the Life Sciences. Three day specialist courses for statisticians cover Pharmacological Modelling, Survival Analysis, Adaptive and Bayesian Methods in Clinical Research and Statistical Methods for Ordered Categorical Data.
In Australia many Universities include workshops for target groups of researchers. For example, at least six workshops of two days duration on aspects of the package R at the introductory or intermediate level and two workshops on Bayesian methods at the intermediate level have been provided. Ten workshops at the intermediate or specialised level have been taught on areas related to epidemiology, clinical trials, survival analysis and longitudinal data ranging from one to three days. These courses have been provided by schools of population health or epidemiology at various universities such as the National Centre for Epidemiology and Public health at the Australian National University, the School of Population Health at the University of Melbourne, Deakin University, the University of Queensland, the University of Sydney and Queensland University of Technology.
The Statistical Consulting Centre at the University of Melbourne has offered two four day introductory courses on Design and Analysis of Surveys, Design and Analysis of Experiments, a six day course on Statistics for Research Workers and a one day course on Producing Excellent Graphics Simply (www.scc.ns.unimelb.edu.au/courses.html).
The University of New England at Armidale is teaching a three day course in January 2010 for postgraduate students and other professionals on Application of evolutionary algorithms to solve complex problems in quantitative genetics and bioinformatics and a five day course in February 2010 on Bayesian methods in genome association studies (email@example.com).
The Institute of Health and Biomedical Innovation at Queensland University of Technology has taught a one day course for intermediate level researchers on Multivariate Data Analysis and SEM using SPSS and AMOS (www.ihbi.qut.edu.au).
The New Zealand Social Statistics Network (www.nzssn.org.nz) provided in January/February 2009 and is providing again in 2010 a set of 10 five day and one four day introductory workshops in the School of Government at Victoria University of Wellington. Course titles include Data Analysis using SPSS, Research Synthesis for Policy and Practice, Advanced Analysis of Linked Data, Qualitative Research Techniques, Introduction to Survey Design, Using Mixed Models in Research and Program Evaluation, Introduction to Structural Equation Modelling using AMOS, Introduction to NVIVO, Introduction to Case Study Design and Introduction to Survey Design.
A four day advanced workshop Modelling Patterns and Dynamics of Species Occurrence (http://www.proteus.co.nz/home.html) was provided by Proteus Wildlife Research Consultants at the University of Otago.
Two introductory two day short courses, Basic Statistics/Analysis of Variance and Simple Regression and Analysis of Covariance, have been offered by Saville Statistical Consulting Limited (firstname.lastname@example.org) at both Rotorua and Lincoln in New Zealand.
In the United States, Statistical Horizons (www.PaulDAllison.com) provided repeat presentations at different locations. Three continuing education workshops, Longitudinal Data Analysis, Missing Data and Survival Analysis Using STATA, were two days duration at an International Association of Statistical Education (IASE) www.stat.auckland.ac.nz/~iase/ ICOTS8 (2010) Invited Paper Harraway intermediate to advanced level. Two five day workshops, Event History and Survival Analysis and Categorical Data Analysis were also at an intermediate to advanced level. These courses will assist with professional development in the relevant areas of statistics use.
There are many programmes which provide distance learning through internet workshops.
Details of one set of courses can be found at email@example.com At the introductory level there are Creating Effective Graphic Presentations, Learn Statistics Through Applications and Introduction to R each of which involve about 30 hours work over three weeks. At the intermediate level are Power and Sample Size Determination, Applying Resampling Methods and Modelling with R also taught for 30 hours over three weeks. A Manager’s Guide to Design and Conduct of Clinical Trials is a four week specialised course of 40 hours.
Another extensive set of on-line workshops and courses for professional development in statistics and areas of application of statistics can be found at statistics.com. It is possible with these courses to construct a programme in advanced statistics study equivalent to a degree in statistics or just register in specific courses as desired for one’s own specialty. There are over 100 courses on offer with most being four weeks duration. They range from the introductory/beginner level through intermediate level to advanced/specialist level. As well as introductory statistics and a set of advanced course on statistical method they include courses relevant to the life sciences, engineering, the social sciences, the environment, business and the statistical package R. It is stated who the courses are aimed at. For example, a course on Structural Equation Modelling is aimed at market researchers, education researchers, sociologists, psychologists, political scientists, economists and survey researchers, a course on Advanced Logistic Regression is aimed at the life sciences, business, social sciences, environmental science and engineering and a course on Cluster Analysis is aimed at market analysts, computational biologists, environmental scientists and IT specialists.
This brief review of courses for continuing education and professional development provides confirmation that particular methods tend to be used in different specialties (Harraway et al 2001) and consequently the review is a guide to how retraining workshops should be targeted and context developed.
THE LOCAL EXPERIENCEThe workshops now described are aimed at a group of ecologists, internal student researchers at the University of Otago rather than external students. The Ecology Research Group at Otago is a diverse group involving thesis Master and PhD students as well as staff members from a range of subjects including Botany, Zoology, Marine Science, Geography, Geology and Chemistry. Each year for the last seven years a sum of money has been available from the convenor of the Ecology Programme to help researchers in this group. Each year the students have decided to fund an intensive three or four day workshop on selected statistics topics related to procedures they are currently using or wanting to use. The topics covered have reflected closely the requested procedures for the workplace listed in Harraway and Barker (2005).
In 2002 and again in 2004 a Statistical Ecologist from outside the University of Otago was invited to teach workshops on multivariate statistics. The workshops were instructive but difficult with an emphasis on some of the underlying mathematics and limited hands on experience for the participants. The number of people attending was around 50 but response was not totally positive.
The courses, based on a textbook being developed by the presenter, were scheduled out of term time and not in the summer as many of the ecologists at this time would be out in the field.
In 2005 I was instead approached to teach a course on multivariate statistics with a new group of research ecologists as those attending the previous courses had for the most part completed their study. Before agreeing to teach this workshop I outlined my plan to see if it would be acceptable. My main objective was to produce a group of researchers who would be capable of using the techniques in their own work. To achieve this I explained that half the three day workshop would involve lecture sessions of one and a half hour length and half would involve hands on laboratory sessions using data which I had accumulated from my consultations with marine science and zoology students as well as data generated by local students or other widely available data in the ecology context. I said that participants could bring their own data for analysis.