Three lessons from statistics for interdisciplinarians and fellow travellers

Community member post by Gabriele Bammer

Gabriele Bammer (biography)

In last week’s blog post on recognising interdisciplinary expertise I argued that forming a new i2S discipline could help embed interdisciplinarity and related approaches (transdisciplinarity, systems thinking, action research, T-shaped research and others) in the academic mainstream. But how would such a discipline work? What are the challenges to establishing an i2S discipline and how could they be overcome?

The discipline of statistics provides three productive analogies. Key to success in both statistics and i2S are: collaboration, dedicated journals to publish advances in concepts and methods, and lobbying for effective application of the discipline.

1. Collaboration

For both statistics and i2S to be effective, collaboration is essential. Both enhance research across a wide range of problem areas—health, the environment, education and more. Statisticians contribute to teams that require expertise in quantitative reasoning. They assist teams in designing studies and in interpreting the outcomes.

Integration and implementation scientists contribute to teams requiring expertise in exploring a complex problem more comprehensively. They assist teams in figuring out how best to approach the problem and its interconnections, which disciplines and stakeholders need to be involved, how to bring together the various disciplinary and stakeholder perspectives, how to take into account what’s not known about the problem, how to support those charged with acting on the problem, and other related issues.

2. Dedicated journals to publish advances in concepts and methods

If a statistician’s work in a project team leads to advances in the statistical concepts or methods they brought to the table, the innovation is reported in the statistics disciplinary literature, not in the literature about the problem they were working on. The advance then becomes available to all other statisticians to apply, as appropriate, in the full range of problem areas, be they in education, environment, security, health or elsewhere.

i2S needs similar journals to share relevant concepts and methods among researchers undertaking interdisciplinary explorations of various complex problems. If an integration and implementation scientist in a team working on illicit drug use developed an innovative scoping technique to identify all the relevant disciplines and stakeholders, for instance, there is currently no journal through which this advance could be made available to others tackling, for example, biodiversity loss, obesity, or poverty reduction.

Instead, relevant tools tend to be published in the literature about the societal or environmental problem, in the grey literature, or not published at all. As well as making concepts and methods hard to find, this means that any research team’s knowledge about the tools that are already available is poor. In turn, that leads to a lot of reinventing of concepts and methods for, for instance, engaging stakeholders and translating evidence for practitioners. It also means that existing concepts and methods tend to stagnate rather than being continuously improved.

Laying the foundations for a future journal of i2S concepts and methods is one of the purposes of this Integration and Implementation Insights blog.

3. Lobbying for effective application of the discipline

The importance of collaboration means that it is not enough for statistics and i2S to simply exist; each needs to be effectively brought into the relevant partnerships. In recent years, statisticians have banded together to ensure that statistical understanding and tools are appropriately deployed. For example, because analysis cannot fix poor design, they have effectively lobbied and educated for inclusion of statisticians at project start-up, rather than the previous practice of just bringing them in at the later analysis stage. As a result, the statistical approaches in grant applications and papers now come under close scrutiny, with funders and journal editors often requesting specific reviews by expert statisticians.

Similarly, it will not be enough for i2S just to be established; action will also be required to ensure that its members have influence in research on complex societal and environmental problems. The role of i2S is to raise the bar in teams tackling these issues. An immediate task is to move teams away from reinventing concepts and methods to employing and building on those that already exist.

A major hurdle to promoting such adoption of i2S is again fragmentation—this time the lack of a unified academic community to drive uptake of i2S, which was described in more detail in my blog post last week.

What would success look like?

Imagine multiple research teams tackling complex societal or environmental problems—such as a team project on fisheries depletion in Sweden, a government research team investigating firearms control in Brazil, a graduate student program tackling obesity in the USA, and multiple research groups in a public-private partnership addressing poverty in sub-Saharan Africa.

At least one member of each team is an integration and implementation scientist, and belongs to a vibrant international academic community striving for continuous improvement in ways to tackle complex societal and environmental problems. Their work has a strong basis in relevant concepts and methods.

Through text books, journals, and an open-access online knowledge bank, they have access to, and help upgrade, a repository of concepts and methods for dealing more comprehensively with complex problems. They provide their teams with practical assistance in choosing and applying the best concepts and methods for their investigations, and use that experience to evaluate and improve those tools.

Their performance for tenure and promotion is assessed in an analogous way to that of statisticians: including by the concepts and methods they have developed and improved and by their ability to contribute to the team in progressing understanding and action on complex societal and environmental problems; all evidenced through publications and grants. Critically, the assessment is conducted by peers from the i2S discipline.

Do you think that establishing an i2S discipline could provide a break-through in the way interdisciplinary and related researchers are embedded in the academic mainstream? Are there other ways to recognise the specialist skills required to 1) weave disciplinary and other insights into a more comprehensive understanding of the problem as a whole, and 2) provide effective options for action? What would it take for you to identify as an integration and implementation scientist? Are there other lessons or analogies that building an i2S discipline could draw on?

To find out more:
Bammer, G. (2017). Should we discipline interdisicplinarity? Palgrave Communications, 3 (article 30). Online (DOI): 10.1057/s41599-017-0039-7

Biography: Gabriele Bammer is a professor at The Australian National University in the Research School of Population Health’s National Centre for Epidemiology and Population Health. She is developing the new discipline of Integration and Implementation Sciences (i2S) to improve research strengths for tackling complex real-world problems through synthesis of disciplinary and stakeholder knowledge, understanding and managing diverse unknowns and providing integrated research support for policy and practice change. She leads the theme “Building Resources for Complex, Action-Oriented Team Science” at the National Socio-Environmental Synthesis Center (SESYNC).

14 thoughts on “Three lessons from statistics for interdisciplinarians and fellow travellers

  1. Thanks for your post, Gabriele.

    From the perspective of a practitioner in modelling, policy, forecasting in relation to capital intensive utilities, I am always sensitive to the way people fail to appreciate limitations in interpreting statistical results – if I may mention them in a cryptic fashion, they relate to correlation (so often misunderstood as causation), confidence (placed in atypical data), non-linearities, and missing factors. In all cases broader collaboration, criticism and education are needed to help correct false conclusions.

    • Thanks Graeme – indeed. That also provides another area of analogy with Integration and Implementation Sciences (i2S) methods, which, of course, also have limitations.

  2. Yes, I read quite a lot around what I call ‘collaboration’ and I find not only much reinvention but also much restarting from scratch: the endless rediscovery of ‘new’ insights.

    The problem experts in collaboration experience is much the same as that experienced by most management consultants who are expert in process rather than content: it is much easier (and often preferred by many influential clients and sponsors) to perceive and appreciate tangible content expertise rather than abstract process expertise, which through necessity and intention (and its nature) tends to morph into new shapes and guises according to context.

    Indeed, the inbred client-focus of process management consultants tends to magnify the contributions of and the results achieved for and by the client and minimise the process expertise of the management consultant. As you allude to above, the latter is at best included as supporting information integrated into a final overall report (or at worst added as a small footnote or afterthought appendix within the same).

    Initiatives, such as yours, which seek to take action to identify collaborative (or interdisciplinary) good practice and record and disseminate it clearly and widely are invaluable in raising the profile, consistency, quality and credibility of collaborative expertise.

    To my mind, however, it is equally important that individuals possessing collaborative expertise begin valuing and promoting it more explicitly, doing their best to ensure that it is given a chance to take a proper (and clearly recognisable) bow beside the results achieved and successes gained by their clients and sponsors, etc.

    After all, it is through the everyday actions of influential individuals that change is not only achieved but also sustained.

    • Thanks for these thoughts. I just want to make clear that i2S is not just about collaboration (or team science) expertise – important as that is. It’s also about making sure the team is well equipped to think productively about the complexity of the problem, for example, to use the most effective systems and other methods and to be able to take context into account.

  3. One important aspect of learning is noise. I’d like to inject what might be considered noise into this thread by suggesting the following paper:

    Engeström, Yrjö (2009). “Wildfire Activities: New Patterns of Mobility and Learning”.
    International Journal of Mobile and Blended Learning. Volume 1, Issue 2. IGI Global.

    That is, if one considers that humans, like all living systems, are complex, adaptive, and anticipatory, those ideas related to how we collaborate are necessarily complex. John H. Holland, Yrjö Engeström and others write in this space; mostly, they talk about aspects of collaborative ecosystems.

    Let me ask this question: how can any academic mainstream, which entails its own competitions such as the Nobel Prize, and funding (!) ever really penetrate the depths of interdisciplinarity?

      • That’s an interesting set of ideas – thanks. Are there any tools your can share for effective ways of “listening to noise”?

    • Thanks, there are some interesting ideas in the paper, although I have to confess that overall I don’t find the argument very convincing (eg the idea that disaster relief is self-organising ignores the fact that, in the example used, it’s organised by the Red Cross, who use project managers to oversee the process). Nevertheless, what I did find useful to be reminded of is that our ideas don’t enter an empty playing field and more importantly, the standard literature review that researchers conduct is a very limited summation of previous knowledge. There are multiple messy earlier ‘tracks’ that our ideas may hook onto. Maybe some of the other literature on collaborative ecosystems will be more helpful.

      Your closing question is an important one, to which I’d say ‘don’t let the perfect be the enemy of the good’. Sure, entrenching interdisciplinarity into the academic mainstream will have limitations, but it will also have strengths in attracting more smart people to important interdisciplinary work. As we gain a more influential voice in the academy and in research (and education) policy decision making, we need to ensure that we use it to counteract the negative effects of mainstreaming.

  4. Such an important issue. I think these are great points, to which I’d add one more. Your second point about the lack of publications gets at this piece, but it’s more than just that. I use this analogy because both statistics and IDS/i2s are methodologically focused.

    Rather than being topic experts, both statisticians and integration scientists provide expertise in processes. These basic processes and tools are accessible for all to use, but they can become incredibly sophisticated and thus complicated in particular scenarios. Bringing in a (process) expert is just one more way teams can ensure that they have all of the right expertise represented, and are able to use the best strategies for conducting effective interdisciplinary work.

  5. Interesting comparison. A few things about statistics worth noting in this context (based on the ideas of Ian Hacking, in his book, The Emergence of Probability):

    (a) statistics developed around the notion of probability and
    (b) probability could be discussed formally (with as much precision as required) based on the demonstrated properties of statistical populations and sampling procedures.

    This type of formalism and the resulting precision was found to be transferable across contexts, which lent support to the science of statistics.

    Whether a similar basis could be proposed for i2S remains and interesting topic for discussion.

      • Following up on the idea that i2S doesn’t have a core formalism like statistics does – would it be fair to say that i2S is an umbrella discipline rather than a core discipline?

        To take another analogy, computer science also has a methodological focus, and as a discipline facilitates collaboration, has dedicated journals to publish advances in concepts and methods, and lobbying for effective application of the discipline. However, it is so broad a discipline that you need to ask what area of computer science someone works in, and there are a large variety of journals in which to publish that are not necessarily connected.

        In the same way, can I say that I work in i2S if I specialise in integrated modelling (or action research, systems thinking etc.)? I would still attend my own specialised conferences and associations, but I would also be a member of the umbrella association(s) and attend large scale multi-stream conferences where all the sub-disciplines of i2S get together?

        In that case, might the next step be less about creating new journals and more about getting existing associations to band together around their common interests in ways that recognise their individuality?
        Instead of having to agree on new (integrated) theory, different existing theories can co-exist and compete within the broader umbrella discipline?

        • Thanks Joseph, that’s a useful idea. I don’t know much about computer science, but will follow it up, because what you are saying sounds about right.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s