‘Bias’ and natural resource management

Acknowledging that environmental managers are only human

People in all walks of life – from town planners to judges and financial regulators – are subject to bias in their perceptions and judgements. Of course, this applies to environmental managers and natural resource managers too. We recently explored the influence of bias in natural resource management (NRM) and found that we may be able to improve our performance if we recognise these influences and work to reduce them.

Sayed Iftekhar, on the right, listens to ecologist Geoff Kay in a grassy woodland (a threatened ecosystem). Sayed has investigated how NRM managers are often influenced by unacknowledged biases in their decision making.

Sayed Iftekhar, on the right, listens to ecologist Geoff Kay in a grassy woodland (a threatened ecosystem). Sayed has investigated how NRM managers are often influenced by unacknowledged biases in their decision making.

Many shades of bias

Decision makers do not always perceive things accurately. It has been shown that, in making judgments dealing with uncertainty, decision makers are susceptible to different types of biases – beliefs that are inconsistent with reality or behaviors that compromise the achievement of objectives.

There is some research around which demonstrates that people are subject to a range of biases. However, the influence of bias has received little attention in the conservation literature. We set out to explore the consequences of these biases on NRM in general and adaptive management in particular (Iftekhar and Pannell, 2015).

Based on our survey of the economics and psychology literature we explored the impacts of action bias, the planning fallacy, reliance on limited information, limited reliance on systematic learning, framing effects, and reference-point bias.

Dealing with the planning fallacy

Each bias can have an adverse impact on our capacity to undertake effective adaptive natural resource management. The ‘planning fallacy’, as one example, is the tendency of project planners to be excessively optimistic about the performance of a project that they are developing. It’s a very common bias and we suspect that it has led to some very poor decisions being made about major NRM investments.

A strategy to reduce the planning fallacy is to ask managers to forecast the completion time, cost, or benefits for a range of comparable projects rather than a single project. This strategy, known as Reference Class Forecasting, has been effective in reducing time and cost overruns of large infrastructure projects.

Where the planning fallacy is in evidence, adaptive management may help to reduce its adverse consequences. Adaptive management, involving information collection and refinement of project design, helps in correcting decisions that were initially made on an excessively confident or optimistic basis. If necessary, targets can be modified or the project can be terminated following the collection of improved information.

A strategy to reduce the planning fallacy is to ask managers to forecast the completion time, cost, or benefits for a range of comparable projects rather than a single project.

The consequences of bias

Based on our survey of the economics and psychology literature, we believe that environmental managers and natural resource managers should be on the look out for a range of common biases that have the potential to adversely impact NRM and specifically adaptive management (see the box ‘Nine shades of bias’).

Based on what we know about these biases there is evidence to expect that:

  1. Managers are likely to undertake on-ground actions even when these are not worthwhile.
  2. They could suffer from the cognitive illusion of being more in control of the system than they actually are.
  3. They could be overconfident about the expected outcome of their decisions.
  4. They may be overly optimistic in terms of expected completion time of the project.
  5. They might rely on a partial set of information for decision making even when more complete information is available.
  6. They might rely on trial-and-error learning and repeating their past successful choices instead of collecting and comparing information about the full set of decision options; and
  7. Managers could try to achieve predefined goals rather than the best possible outcomes from a project.
Project planners are often excessively optimistic about the performance of a project that they are developing. It’s important to aim high but making judgments about a planned activity that are systematically over-optimistic (including underestimating project completion time, underestimating costs, or overestimating benefits) can lead to some very poor decisions.

Project planners are often excessively optimistic about the performance of a project that they are developing. It’s important to aim high but making judgments about a planned activity that are systematically over-optimistic (including underestimating project completion time, underestimating costs, or overestimating benefits) can lead to some very poor decisions.

Minimising the impact

Bias is a part of human life. The take home message from this study is that NRM agencies need to be aware of the influence of biases when management decisions are undertaken. There are many things they can do that will help minimise the impact of bias.

First, agencies need to promote a culture of learning. It needs to be recognized that both successful and failed projects generate valuable information about the future state and expected impacts of the management interventions. This could be done by providing appropriate incentives (tangible and intangible) for the managers and decision makers to consider the full range of options before making any decision, or asking managers to justify their decisions to external parties.

Second, adoption of a decision support system could facilitate retention and storing of relevant information. It may also make learning from past projects easier and help in systematic evidence-based decision making. Of course, relevant staff should be adequately trained and properly incentivized to use such systems.

Third, conducting benefit-cost analyses of planned options would help to refine and prioritize the options during the design phase of an adaptive management cycle. Benefit-cost analysis provides a systematic and objective framework to include all relevant costs and benefits (both market and nonmarket goods and services) related to a project. In the process of identifying benefits and costs, it also helps in identifying whether there is complementarity among them (to avoid double counting) and the time lag and uncertainty attached to realization of each benefit and each cost. Thus, benefit-cost analysis could be used as a tool to comprehensively assess the expected merits of a project.

Fourth, involvement of external third-party reviewers may also help in designing more realistic and feasible projects.

And, finally, scenario analysis should be conducted as part of the assessment and design phase to anticipate the expected outcomes of different options. It is advisable to consider the likely impacts of different types of biases, and the effectiveness of potential remedial measures before making any final recommendation for use in decision making for natural resources.

Nine shades of bias

Here are nine behavioral biases that we believe can potentially affect adaptive management

Action bias: Tendency to take actions even when it is better to delay action

Framing effect: Tendency to respond differently to alternatively worded but objectively equivalent descriptions of the same item

Reference-point bias: Tendency to overemphasize a predetermined benchmark for a variable when estimating the level of that variable

Availability heuristic: Tendency to give more weight to events that can be recalled more easily

Planning fallacy: Making judgments about a planned activity that are systematically over-optimistic, including underestimating project completion time, underestimating costs, or overestimating benefits

“Satisficing rule”: Tendency to stop searching for a better decision once a decision that seems sufficiently good is identified

Loss aversion: Tendency to value losses more highly than similar gains

Reliance on limited information: Tendency to use a subset of information even when full set of information is available

Limited reliance on systematic learning: Tendency to use information from past successful efforts rather than using information from both successful and failed efforts.

Studying behavioral bias

Both psychology and economics have rich literatures on the influences of different types of bias on behavior. Experimental economics serves three main purposes: testing theories, building new theories from observing experimental outcomes, and testing policy and management options. Behavioral economics also integrates insights from psychology to explain economic decision making. It studies the effects of psychological factors such as emotional, social, and cognitive factors on many decisions and economic processes. A related field is behavioral decision theory, which studies how people make decisions as well as how they should make decisions.

More info: Sayed Iftekhar mdsayed.iftekhar@uwa.edu.au


Iftekhar MS & DJ Pannell (2015). “Biases” in Adaptive Natural Resource Management. Conservation Letters. doi:10.1111/ conl.12189 http://onlinelibrary.wiley.com/enhanced/doi/10.1111/conl.12189/

4 comments on “‘Bias’ and natural resource management”

  1. Tim Hosking Reply

    Great article, though as a ‘manager’ I would disagree with some of it.

    Managers are working with the pressing challenges of today, which is typically 3-5 years in advance of the outcomes of the next research or monitoring project to directly back them up. Better to act, based on best available information and a working systems experience with a chance of success, than not act with nil chance of success. Waiting the 3-5 years for a science project to be published which may never come, may not find enough data or be nuanced enough to inform decisions or where it will possibly be outdated by the time it arrives is not a gamble many will take. Government funding cycles where they are driving actions also will not typically wait.

    Also, BCA/CBA cannot be done for many NRM decisions where non-market valuations for all the possible benefits and costs would simply take too long, be too expensive or be inconclusive anyway.

    • David Pannell Reply

      Hi Tim

      Thanks for your comments. These are interesting issues. I’d like to respond and clarify our points a bit.

      I’m interpreting that your main point relates to “action bias”: action is taken but inaction would be better. There are several possible reasons why inaction could be better: (a) there is sufficient evidence to judge the merits of current action and that evidence indicates that it is not worthwhile, (b) action now is irreversible and rules out other options in future that are could be better, or (c) we don’t have enough information to judge whether action is worthwhile. I think your focus is on the last of these. My first comment is, don’t forget about the other two possible reasons, especially (a). Money is always tight, so if you take costly action in an area where the case for action is not compelling, you are necessarily giving up the opportunity to do other things, some of which may be more compelling. My experience working with many environmental organisations is that they do often tend to pursue projects without taking the time to consider in a critical way whether they really will deliver sufficient benefits to be better than other possible uses of the money and staff effort. Such critical consideration doesn’t have to rest on top-knotch research, but it should at least involve being transparent about what judgements you are making about the difference(s) the project will make relative to business as usual, and how significant those differences are.

      My next comments relate to having adequate information (case (c)). When thinking about taking action, there is a continuum between having no information about the issue and having full information. We’re certainly not saying that that managers should wait for full information before proceeding with a project or action. On the other hand, no sensible manager would take action based on zero information. The question is, how much information is enough to justify action? Our point is that at least some environmental managers may be committing to actions based on information that is too close for comfort to the “zero information” end of the spectrum. There is evidence that this happens in various fields, so it would be surprising if the environment was immune to this problem. And again, I’d comment that I’ve seen examples of it in practice.

      The scarcity of funds is also relevant here. You’ve characterised the situation as: we do a project or we do nothing. However, at some level the reality is: we do a project or we do a different project. If we’ve got more reliable info about the other project, that should influence our choice. It’s not the only factor by a long shot, but it is certainly one of the factors.

      The fact that government funding cycles will not wait certainly adds to the problem. It probably reflects some form of action bias at the political level.

      I agree with you that a full-blown BCA of every possible environmental project, including original studies of estimating the non-market value, is miles away from being sensible or even possible. There is a relationship between the size and importance of a set of investments and the amount it is worth spending on the decision making process. When working with environmental organisations I try to be explicit about this: quick and dirty analysis for very small projects, more thorough analysis for bigger ones. Even then, I’d argue that a Benefit: Cost Analysis mindset is the right way to think about the decisions, even for relatively small projects. The information you include might be simple guesswork, but at least you can get the best possible environmental outcomes based on that relatively weak information.

      Our tool INFFER (Investment Framework for Environmental Resources – http://www.inffer.com.au) is based on this sort of philosophy. There are versions of different levels of complexity. The fuller versions can take non-market values as an input if they are available, but that is rarely the case. Usually information about environmental values is based on judgements by the people who are using the tool.

      Dave Pannell

  2. Austin Milt Reply

    Try replacing ‘manager’ with ‘everyone’ and remember this when doing NRM science. A great reminder not one of us is objective.

  3. Max Bourke AM Reply

    Good article and thoughtful. As someone who has worked for decades as both a scientist, not very good, and a land manager, slightly better, I think you have missed one form of bias that sometimes drives people like me crazy. It is the tendency of scientists/researchers to end every publication with “but this needs further research”. This is not just a phrase but a real approach/bias/cause of inaction which can lead to nothing being done…ever. It certainly drove me nuts when we were doing the first State of the Environment Report in the 1980s, I think we might still be writing that report if I had not put my foot down, and said “well lets just take a punt”. Best wishes Max Bourke AM

Leave a Reply