Preparing for working with us
Biostatisticians are academics who have specialised in study design and statistical methods, and who have expertise in quantitative health-research in general. To get the most out of working with a biostatistician, whether this is a one-off meeting or a long-running collaboration (we encourage the latter), you will need to understand some statistical concepts or, at least, have insight into thinking statistically.
The Centre offers short courses that can provide some of this background—the Introductory Biostatistics 2-day course is an excellent starting point. Any other training you have had in biostatistics, statistics, epidemiology, bioinformatics, data science, or research methods in general can also be helpful. If you do not feel confident having a conversation about the design and statistical methods of your research project, we encourage you to bring someone with you who does. We insist that masters' and doctoral thesis students contacting the Centre bring their supervisors to at least their first meeting with a biostatistician (see 'Allocated support for masters' and doctoral students').
The following key policies are important in clarifying the role of an academic biostatistician.
Biostatistics policy summary
- All requests for collaboration, including one-off meetings, must be entered into our IRIS database.
Login to IRIS database - Biostatisticians do not provide one-on-one teaching of any kind. There are courses and other resources available to cover this.
- Biostatisticians will neither provide code that they have written nor perform analyses to order.
- Masters' and doctoral thesis students can have up to two hours of our time per calendar year for general advice. They must bring at least one supervisor to their first meeting.
- When part of the supervision team for masters' and doctoral thesis students, a minimum of 20% supervision (and the same for EFTS) is expected.
- Irrespective of their role, biostatisticians will not carry out any statistical analyses for any student (including sample size calculations or checking / confirming results from statistical analyses). This policy is consistent with those of the University of Otago.
- Masters' and doctoral thesis students supervised by a biostatistician in the Centre, can attend any workshop provided by the Centre at no cost.
- We expect our contributions to collaborative research to be recognised appropriately, including through being named investigators on funding applications and authorship when justified.
General
Students
Funding and authorship
For more detail on each policy:
Understanding our role
We expect that researchers wishing to work with a biostatistician from our Centre will have familiarised themselves with our policies and other relevant material on our webpages.
Our biostatisticians are not data managers, which is an important and specialised methodological area in itself. When data management (including but not limited to combining data files or calculating scores or other values from data) is a necessary part of the research, biostatisticians may agree to take responsibility for some or all such tasks. It should not be assumed that these tasks are the biostatistician’s responsibility unless they have explicitly agreed. Similarly, biostatisticians are not health economists, bioinformaticians, graphic designers, programmers, or epidemiologists. Biostatisticians may suggest that specialists in these or other areas (particularly including Māori or Pacific health) are needed.
Working together effectively
In general, we find that meeting in person produces the best outcomes but we can also meet via Zoom or Teams and will take all reasonable steps to support our collaborators with disabilities.
We expect that all potential conflicts of interest will be declared and managed appropriately from the outset.
The role of the biostatistician in a particular project should be clear to all members of the research project, including the biostatistician. Good communication is key. In some cases, this role may be advice only and in others the biostatistician might be responsible for performing the statistical analyses. The writing of the statistical methods will generally fall to the person who performs the analyses but the writing of the results should be carefully discussed as this can require more subject-matter knowledge than the biostatistician will possess.
When collaborating with a biostatistician, it is expected that they will be invited to all meetings about the project (this includes being sent agendas and minutes—but note that the biostatistician may decline to attend some meetings) and included in all emails about the project. We do not expect non-biostatisticians to always be fully aware of the potential biostatistical implications of decisions made during study design, data collection, data coding, and when planning communication of findings. Keeping the biostatistician fully apprised of the research is the best way to avoid complications in analyses or planned analyses being ruled out.
Timelines, including key dates, which must be clearly communicated, should be discussed and determined by agreement. Clear communication around expectations (including timelines) and responsibilities avoids many disagreements. If a disagreement arises, careful and respectful discussion should occur with the biostatistician in the first instance. If the matter cannot be satisfactorily resolved, it should next be escalated to the Director of the Biostatistics Centre in a timely manner.
Guidelines and ethical practice
Our biostatisticians follow American Statistical Association (ASA) Ethical Guidelines. Biostatisticians also follow the University’s relevant guidelines, including the Responsible Practice in Research—Code of Conduct. In general, biostatisticians reserve the right to decline to perform any analysis that they personally feel violates ethical principles, including their own personal ethical principles, and this includes, but is not limited to, concerns about how results might be interpreted or used and issues of data sovereignty.
The use of artificial intelligence (including Large Language Models) in biostatistics is an emerging topic and, at the present time, the Biostatistics Centre is still developing its policies around this.