A team-based approach to HDR supervision 1

A team-based approach toHDR supervision

Rob J Hyndman

Econometrics & Business Statistics, Clayton

A team-based approach to HDR supervision My background 3

My background

PhD at Melbourne Uni (1992), Departmentof Statistics. One of two PhD students inthe department.

I saw my main supervisor three times inthree years.

In 1994 I moved to Mathematics atMonash where the PhD students inStatistics never interacted as a group.

A team-based approach to HDR supervision My background 4

My background

In 1998 I moved to Econometrics &Business Statistics where there were about30 PhD students. They never interacted asa group.

By 2004, I had 7 PhD students and wasspending a large amount of time helpingeach of them with the same issues.

I talked to PhD students in laboratorydisciplines where they all worked in teamshelping each other.

A team-based approach to HDR supervision The team approach 6

The forecasting research team

Begun to save me time and to encourageinteraction between students.

We share ideas, feel part of a group withsimilar objectives, discuss researchplanning issues and technical problems.

We meet as a group once every twoweeks. I meet with each personindividually on the alternate weeks.

A team-based approach to HDR supervision The team approach 7

Forecasting research team

A team-based approach to HDR supervision The team approach 8

The apprenticeship modelI see HDR students as “apprentice researchers”.My aim: help them become fullyindependent researchers before they leave.

settling in a new countryframing the research problemsearching the literaturewriting papersgiving seminarsmeeting other researcherscomputer programmingmathematical typesettingattending conferencesproducing a thesis.

A team-based approach to HDR supervision The team approach 9

Early stages

Before they arrive. . .

Begin working on a research proposal.

Discuss sitting the IELTS test, satisfyingthe university regulations, and so on.

Settling in

Assign an existing student to help withfinding accommodation, learning wherethe nearest shops and other facilities are,meeting people of the same ethnicbackground, sorting out computeraccounts and library access, and so on.

A team-based approach to HDR supervision The team approach 10

Early stages

A team-based approach to HDR supervision The team approach 11

Group meeting topics

how to give a good research seminarexplaining your research to people in otherfieldsnew software toolsdebugging computer programsusing LATEX for writing papers and a thesishow to subscribe to RSS feeds for journalswriting a referee reportwriting a response to referee reportsusing some of the new search tools inresearchtools for managing backing up

A team-based approach to HDR supervision The team approach 12

Blog

Some of the topics discussed in our group aresummarized on my blog:

www.robjhyndman.com/researchtips

On most days, the site receives about 300unique visitors

Largely as a result of this website, I receivemany requests from potential overseasstudents wanting to come to Monash.

A team-based approach to HDR supervision The team approach 13

Research blog

Repository of information for researchteam

Topics discussed in team meetings arewritten up for the website

Using blog software so subscribers canreceive automatic updates

Some entries have received over 1000 hitseach.

Essentially an online book on doingresearch in statistics and econometrics,but always evolving and up-to-date.

A team-based approach to HDR supervision The team approach 14

Research library

Students wanted long-term loans of somebooks.

Others wanted to check a reference, butdid not want to go to the library.

So I’ve developed an extensive library ofbooks of use to my students.

Online borrowing system onwww.robjhyndman.com/borrowed

A team-based approach to HDR supervision The team approach 15

Team teaching

Students teach the group specific skills ortopics in which they have expertise.

Builds confidence and experience in ateam environment.Examples:

debugging tools in the R computer languageusing animated graphics in seminarpresentationsstatistical tests for comparing forecastmodels.remote access of computers.

A team-based approach to HDR supervision The team approach 16

ConferencesAttending international conferences a vitalcomponent of research life.All my students are expected to attend theInternational Symposium on Forecasting atleast once.Every year for the last seven years, theMonash contingent has been the largestfrom any university at the ISF.Students often continue lengthycorrespondence with other researchersthey have met at ISFs.Department funds first trip, studentsobtain grants for subsequent trips.

A team-based approach to HDR supervision The team approach 17

Conferences

A team-based approach to HDR supervision The team approach 18

Conferences

A team-based approach to HDR supervision The team approach 19

Seminar preparation

Before a student presents a seminar orconference paper, they give their talk tothe group and obtain feedback.

Sometimes this leads to lively debateabout what is appropriate, and how best toexpress the key ideas.

By the time a student gives the conferencepaper or seminar, the talk has beenwell-practised and the flaws addressed.

A team-based approach to HDR supervision The team approach 20

Writing issues

Of the 27 HDR students and post-docs Ihave supervised, only four are nativeEnglish speakers.We cover English grammar and otherwriting issues in team meetings.

Common confusions: affect/effect, later/latter,stationary/stationery, discreet/discreteDefinite articles, singular/plurals.Words to avoid: “Obviously”, “Note that”, . . .

Structuring a paper

Writing a good abstract and introduction.

I employ a copyeditor to help the studentswith their English.

A team-based approach to HDR supervision The team approach 21

Writing issues

Activities1 Rewrite the abstract of a published paper

so that it is clear.2 Critique published papers — practicing

referee reports.3 Identify features of good papers.4 Keep a productivity diary.

A team-based approach to HDR supervision The team approach 22

Social interaction

A team-based approach to HDR supervision The team approach 23

Social interaction

A team-based approach to HDR supervision The team approach 24

Social interaction

A team-based approach to HDR supervision The team approach 25

Building a research group

Consider forming a group between 2 or 3supervisors working on similar topics.Try to guide students towards topics thatfit within a defined field.Encourage research students to give talksto the group on their project.Provide great facilities.Group meetings covering broader issuesGet visiting scholars to address researchgroup meetings and encourage questions.Involve post-docs and other researchassociates and assistants.

A team-based approach to HDR supervision Facing difficulties 27

Difficulties

Psychiatric problems

Depression

Cultural expectations

Family expectations

Weak students

Beware of surrogate parenting

Personality clash

Losing contact

A team-based approach to HDR supervision Facing difficulties 28

Reasonable expectations

A supervisor can expect a PhD student to

turn up to appointments, prepared forthem;

write regularly and share the draftmaterial;

tell the truth about work done and notdone;

keep in touch — socially, practically andacademically;

do the research tasks that have beenmutually agreed and scheduled.

A team-based approach to HDR supervision Facing difficulties 29

Reasonable expectations

A student can expect a supervisor toprovide

regular supervision. At least fortnightlyone-on-one meetings. More in the firstyear.

written feedback on draft material.

guidance on the value of proposeddirections and methodology

guidance through the existing literature inthe area.

(Adapted from Delamont, Atkinson and Parry, 2004)

A team-based approach to HDR supervision Producing the thesis 31

Survey of examiners

Results of a study involving interviewswith 30 experienced examiners

Expect students to pass.Have formed a view about the quality of the thesisby the end of Chapter 2.Are very, very reluctant to fail a student with mostexperiencing considerable distress if they do so.Poor thesis equates with sloppiness.Are fiercely independent in their views.Have an inclusive view about the writing of ESLstudents.Have surprisingly ‘catholic’ approaches tomethodology.Are influenced by papers published by students,especially in science.

A team-based approach to HDR supervision Producing the thesis 32

Producing the thesisWrite early and oftenProduce a table of contents at end of firstyear, and revise it periodically. Think aboutthe flow of ideas. Tell a story.Literature review should not be encyclopedic.Keep it focussed on the relevant literature.Papers that are only cited in one chaptercan be discussed in that chapter.Get students using bibliography systemearly. e.g., EndNote, BibTeX, ZoTero orMendeley, with indexing of printed papers.Make successful theses available (butavoid the outstanding!).

A team-based approach to HDR supervision Producing the thesis 33

Producing the thesisMake sure they have backups in more thanone location.Try to publish sections as they arecompleted as working papers and submitto journals. Allows early feedback,simplifies examination, easy to convert tochapters, builds confidence.Discuss writing issues with group ofresearch students. They can help eachother improve, and it saves repeating yourself.Watch tenses, forward referencing, definitions.English editing is always required. I paysomeone to do it.

A team-based approach to HDR supervision Producing the thesis 34

Producing the thesis

Finishing a thesis takes about twice thetime a student estimates.

Discuss potential examiners with students.Ask if there is anyone they do not want.

Choose examiners who are sympatheticand not picky. Personally known is best.Must know what an Australian PhD is.Older and wiser is usually better.

Read what the examiners have written,and make sure you cite them favourably.

Informally approach the examiners.

A team-based approach to HDR supervision Producing the thesis 35

Final word

Research teams are better for

students and supervisors.

Think of students as apprentice

researchers and teach them all

aspects of being a researcher.

Be the researcher you want your

students to become.