How information design shapes our decision-making?

In a perfect world, people make rational decisions based on having complete information to evaluate alternatives and choose the best possible course of action. In the real world, risk and uncertainty are part of our everyday life, making decision-making a complex cognitive process where intuitive thinking and emotions come into play. The way we solve several problems is heuristic in nature: we use mental shortcuts to make a decision, pass judgment, or solve a problem quickly and with minimal mental effort. Thus, the quality of a decision is heavily contingent on the kind of information available. For instance, the same piece of information can be presented in numerical, linguistic (words, phrases) or pictorial (symbols, graphs, images) formats [1]. As individuals are likely to process these formats differently, their decisions may differ. This phenomenon is known as “ framing effect” was first studied by Amos Tversky and Daniel Kahneman who showed how participants’ choices were susceptible to marginally different phrasing of the same task in a hypothetical scenario [2]. Information structure can also shape decision outcomes as well. The order or sequence in which information is presented to decision-makers impacts the way information is processed [3]. Reaching a decision is a complex process, given that it is shaped by what information is available, its format, order and its framing.

Based on these insights, we wanted to further understand how information display affects the decision process when it comes to decisions on a plant with novel trait(s) (PNT). We have done for from a collection of survey responses of 501 Canadians and 118 experts from around the world. Respondents were surveyed on their responses to the different forms of information (linguistic or a combination of linguistic and numerical formats) and its structure (information presented piece by piece or all together at once), and how this influenced participant responses regarding new agriculture technology applications. Participants were invited to submit their decisions on whether a PNT can be released for cultivation in Canada. In Canada, PNTs have a unique classification of regulation based on the novelty of the final product. Respondents had to recommend if a plant was PNTs or not based on agronomic, environmental and safety considerations as described in the decision documents of the Canadian Food Inspection Agency (CFIA). Depending on how the information was structured, the following scenarios were randomly allocated to participants. While they may look slightly different, there is no difference in the meaning of the qualitative and quantitative descriptions of the treatments.

The way information is presented matters for the public more than for experts

Our surveyed public, when presented with a qualitative description they made statistically different judgments in terms of recommending the PNT from those participants presented with a quantitative description. The unfamiliarity with or limited understanding of the task—of a technical nature—may explain why non-experts make biased decisions grounded on social aspects rather than scientific proof. Expert decisions were found to be less—or not at all—influenced by the way information was framed (with words or numbers). The consistency in the expert decision is a result of extensive knowledge and accumulated experience that supports informed decision making.

Our findings show that non-experts are more vulnerable to how information is presented to them compared to experts in the context of agricultural innovation. In short, for our PNTs, the decision is description-variant for non-experts and description-invariant for experts. Experts appear to better process information owing to their experience with different types of evidence in decision-making. Our result lends support to previous studies comparing experts and novices decision-making under uncertainty .

With the exception in a few circumstances, our results show that surveyed Canadians and experts were not generally influenced by information structure. For instance, for one of the food safety criteria, similar proportions of Canadians (83% and 86%) and of experts (54% and 50%) did not recommend PNT under bundled and non-bundled treatments, respectively. That is, similar proportions within each sample were not in favor of PNT regardless of the presentation structure [6].

Policy recommendations

The key take-home message is that the way information is presented matters, especially to the general public. Information framed with numerical and linguistic content led to different responses about biotechnological innovations (more favorable in the case of PNTs) compared to narrative format. As the current decision documents released by the CFIA, and many other national regulatory organisms, tend to be more qualitative—using numbers sparingly, our result should help regulators to rethink how technical information should be presented and communicated to end-users to avoid biases in decision outcomes.