Avoid Bad Decision Making August 2011
Relevant data that individuals, businesses, and other organizations require to make sound decisions can at times be overwhelming in scope and complexity, requiring some sort of interpretation or curation to organize the data into usable intelligence. Unfortunately, the tools available for data sorting and selection do not always represent the original information properly, which can have negative effects on decision-making processes.
The way opinions are conveyed can be just as important as (if not more so than) other decision-making parameters.
US psychologist Barry Schwartz claims that conventional cost-benefit analyses are often less useful than people traditionally believe. Decision makers often view the world in terms of closed systems with easily identifiable costs and benefits. However, the closed-world assumptions are almost always incomplete, and the complexities of the state of the world and the state of the people in it necessary to understand the true cost-benefit parameters are not accounted for. Schwartz argues that without having all the knowledge in the world, cost-benefit analyses cannot replace a system of moral rules that guide decisions. However, many important decision makers still use cost-benefit approaches, be it because of their perceived objectivity or for their acceptance in the business world.
On the other hand, relatively simple approaches to filtering data, such as popularity, are not always sufficient for complex or sensitive issues. The method of popular voting is used to decide elections, complex ballot measures, scholarship winners, and the winners of television contests like American Idol. Unfortunately, the decision makers in these approaches (the voters) often lack the expertise to weigh the technical merits of the presented choices, and incomplete information can make it surprisingly easy for individuals to change, alter, or discount beliefs.
To make technical decisions, people must often rely on experts and scientists who possess the knowledge and skills necessary to offer informed opinions on their areas of interest. However, the way opinions are conveyed can be just as important as (if not more so than) other decision-making parameters. For example, researchers at the University of California, Berkeley, found that presenting people with dire warnings regarding the consequences of global warming in fact reduced individuals' likelihood to believe that climate change is happening, whereas people presented with possible solutions to climate issues were more likely to believe in associated problems and to take action to reduce their carbon footprint. Beliefs around a wide range of issues are malleable, in part, because it is extremely difficult to absorb the issues' complexities.
In other cases, relevant—and correct—data can generate distorted views (either deliberately or otherwise) in ways that make it difficult for people to understand the situation and draw appropriate conclusions. A recent New Scientist article highlighted—as an example for a wide range of similar distortions—how in 2008, Fox News (News Corporation; New York, New York) reported on a scientific study that stated in its abstract that throat cancer among white men had risen more than 400% in 30 years. This startlingly large percentage increase, however, reflected a rise from 1.01 to 5.69 incidences of throat cancer per 100 000 person-years—a figure that is relatively small in absolute terms, providing a somewhat different notion about the development's urgency. Although the media often showcase the most eye-catching element of a story, in this case, the scientists themselves actually highlighted the dramatic percent increase. While not a misrepresentation of the data, its delivery was seemingly designed to provoke an emotional response and perhaps garner added attention for the study.
In some cases, though, even the research to obtain information can be controversial, leaving decision makers without relevant data to make informed decisions. Former scientists at the US Centers for Disease Control and Prevention (CDC; Druid Hills, Georgia) claim they cannot answer basic questions about the causes of injuries and death related to firearms in the United States because their funding was cut off in the 1990s by legislators the National Rifle Association (NRA; Fairfax, Virginia) had lobbied. The NRA in turn argues that the studies the CDC was conducting in the 1990s were political opinion rather than scientific exploration.
Additional approaches to facilitate and focus information acquisition are curation and recommendations. Curation of data either by experts or by analytic software tools creates its own set of pitfalls, though, by promoting a preselection of information potentially without providing users of the information with the credentials or intentions of the curator. Internet search engines and recommender engines aim at sorting through vast amounts of data to aid users in decision making by limiting information to the data perceived as most relevant to the individual. But the mere act of selecting a subset of information has led experts to critique the process because of vanishing serendipity—information the individual was not aware of—leaving him or her with a limited, reinforced worldview. For solutions like internet-based movie-provider Netflix's (Los Gatos, California) recommender engine, preselected movie recommendations may represent appropriate customer service; however, when internet search engines like Google's (Mountain View, California) base results on previous searches and website navigation, the bias can have much wider implications. Eli Pariser, the cofounder of the liberal political organization MoveOn.org (http://front.moveon.org), related his experience of websites that contain news and opinions from conservative commentators disappearing from his internet experience because social media and internet-search sites prioritized links that they calculated would be more in line with his interests and views. Pariser lamented the loss of perspectives that challenged his worldview; the curated and recommended media he encountered only serve to reinforce his existing beliefs. As modern societies increasingly fragment into smaller and smaller subsegments whose sources of entertainment and information become tailored and personalized to their views (as the internet makes possible), evaluations and decisions could increasingly lack a broad, balanced perspective.
Considerations of scientific research and interpretation as well as stakeholders' interests have always played a role, but instant availability of an overwhelming amount of data and the need to rein in the deluge of information is a fairly novel phenomenon. A wide range of tools exist: all of them valid in certain applications, all of them misleading and inappropriate in other situations. A decision maker's first decision has to be selecting the most appropriate tool and approach. Then the decision maker must weigh the benefits and shortcomings of that tool.