By Umaiza Anwar and Dr. Sheheryar Banuri
Introduction
Why does negative news seem to linger longer and influence us more deeply than positive headlines? Psychological research suggests that people tend to weigh negative information more heavily than positive information—a phenomenon known as negativity bias. While this tendency is well documented among the general public (see for example, Rozin and Royzman, 2001; Vaish, Grossmann, and Woodward, 2008;or Norris, 2021), less attention has been given to whether policy professionals—individuals responsible for shaping and generating policy advice—are susceptible to such biases. This post examines whether policy professionals’ beliefs about the effectiveness of the Sustainable Development Goals (SDGs) are disproportionately influenced by negatively framed public information, relative to a neutrally framed, or a positively framed benchmark. This is of critical importance precisely because policy professionals are expected to base their judgement on evidence and analysis (i.e. engage in evidence based policymaking). If professional expertise protects decision-makers from cognitive biases, we would expect them to be less susceptible to socially transmitted signals about policy success or failure. Our findings show that policy professionals are influenced by negative framing, revealing an important cognitive vulnerability in high-level decision-making.
Negativity Bias
Negativity bias is a psychological phenomenon in which individuals give greater weight to negative information than to equally strong positive information (Baumeister et al, 2001). Research in cognitive and social psychology consistently shows that losses loom larger than gains—a principle closely related to loss aversion in prospect theory (Kahneman and Tversky, 1979). Negative information is processed more deeply, remembered more vividly, and perceived as more urgent than positive information of the same magnitude (Rozin and Royzman, 2001). While this tendency has been widely documented among the general public, it is critical to examine whether policy professionals—who are trained to evaluate evidence objectively—are similarly susceptible. For example, Banuri, Dercon, and Gauri (2019) demonstrated the susceptibility of policy professionals to framing effects (and specifically, loss aversion) and confirmation bias. We extend this work to test for the presence of negativity bias.
One explanation for negativity bias lies in evolutionary psychology. From a survival standpoint, sensitivity to threat increased the likelihood of survival. As a result, humans developed heightened responsiveness to potential risks (Baumeister et al, 2001; Rozin and Royzman, 2001). Neuroscientific research suggests that negative stimuli activate threat-detection systems in the brain, including regions associated with emotional salience (LeDoux, 2000). This heightened attention makes negative information feel more immediate and consequential than positive information, even when both are equally plausible (Vaish, Grossmann, and Woodward, 2008).
Negativity bias also interacts with social information and heuristics. When prominent news sources signal that a policy may fail, this information becomes more cognitively available, increasing its perceived likelihood through the availability heuristic (for more on the availability heuristic, see Tversky and Kahneman, 1973). In addition, framing effects suggest that how information is presented—particularly when framed as potential loss or failure—can significantly alter judgments (see, for example, Banuri, Dercon, and Gauri, 2019). Policy professionals exposed to negatively framed narratives may therefore interpret those narratives as stronger evidence, even if objective data had not changed.
Finally, cognitive efficiency plays a role. As cognitive misers, individuals rely on mental shortcuts when processing complex policy information. Negative signals may be treated as diagnostic cues that require less further scrutiny because they appear to signal risk. In high-stakes policy environments where decisions must be made under uncertainty, this can lead to disproportionate weighting of pessimistic narratives (Skowronski and Carlston, 1989). Thus, rather than merely conforming to popular opinion, policy professionals may exhibit heightened susceptibility to socially transmitted negative information.
Survey Design
The experiment presents a direct test of negativity bias with a sample of policy professionals. We operationalize this by asking policy professionals about their judgement on the impact of the Sustainable Development Goals (SDGs). In the Control condition, policy professionals (management level officials serving in the World Bank) are simply asked about their personal opinion on the effectiveness of the SDGs. In the Negative treatment group, policy professionals are informed that prominent news outlets believe the SDGs will have no long-term impact and that this opinion is gaining traction. In the Positive treatment group the opposite perspective is presented: policy professionals are informed that many prominent news outlets anticipate a very high long-term impact from the SDGs, and that this view is becoming more widespread. The participants were asked to rate their personal beliefs about the SDGs’ effectiveness on a scale from 1 (no impact) to 5 (extremely high impact). By comparing the responses across these three conditions, we can observe whether policy professionals’ opinions are influenced by news items.
Survey Questionnaire
- Control: Personally, what do you think the SDGs will achieve?
- Negative: In recent times, many prominent news outlets have stated that the SDGs will have no impact in the long term. Furthermore, this opinion is getting stronger. For example, see: https://psmag.com/environment/why-are-the-uns-sustainable-development-goals-stalling. Personally, what do you think the SDGs will achieve?
- Positive: In recent times, many prominent news outlets have noted that the SDGs will have very high impact in the long term. Furthermore, this opinion is getting stronger. For example, see: http://www.telegraph.co.uk/connect/better-business/business-sustainability/business-engaging-un-sustainable-development-goals/. Personally, what do you think the SDGs will achieve?
Participants provided their response on a simple 5-point Likert scale where:
1=”No impact”
2=”Very low impact”
3=”Neither low nor high impact”
4=”High impact”
5=”Extremely high impact”
The sample size for this study consisted of a total of 526 survey respondents, with 178 respondents for Control, 163 for Negative, and 185 for Positive.
Analysis and Results
Analysis
We conduct 2 sets of analyses:
- Proportion of respondents by response and group (control, positive, negative). This is to evaluate whether respondents skew positive or negative compared to control.
- Distribution of scores by group and a comparison of medians and means. This is to evaluate if there are any differences in the median or mean score among groups.
Results
Figure 1 shows the proportion of respondents by group where the x-axis shows all 5 answer choices, and the y-axis shows the percentage of respondents in a group (with a bar for each group). For each response score, the left most bar shows the control question, the middle bar shows the positive framed question, and the rightmost bar shows the negative framed question. The “very low impact” response was selected by only 10.27% of the positive treatment group, 17.42% of the control group, and 25.77% of the negative group. This shows that the way the question is presented does have an impact on the response people give. More respondents chose the answer with a negative connotation when asked the negatively framed question than not. The same phenomenon happens in reverse on the “high impact” answer choice. The fewest respondents are for the negative group (30.06%), followed by the control group (37.08%), and with the highest response rate (41.62%) from the positive group.
Figure 1. Proportion of respondents by score and respondent group
The results show evidence in favour of negativity bias: This can be directly observed based on the difference in proportions. Starting with the “very low impact” responses, we observe that the difference between the control and the positive groups is 7.15% but between the control and negative groups the difference is 8.35%. This means that the negative frame question had more of an overall effect than it did for the positive frame question. A similar phenomenon is observed for the “high impact” responses. The difference between control and positive is 4.54%, but the difference between control and negative is 7.02%. The respondents in the negative group demonstrate higher shifts in their responses relative to the positive group.
The results from the first analysis are supported by the second analysis. In Figure 2 we plot the distribution of scores (boxplot) and mean by group. We observed that the median response is comparable across all groups. However, the mean scores indicate that the negative group had a lower score than both control and positive groups (with no meaningful difference between control and positive). The greater difference between the negative and control averages (~0.25) than the control and positive averages (~0.07) demonstrates that the negative question has more effect than the positively framed question.
Figure 2. Distribution of score by group (red dot = mean of group)
Conclusion
This study set out to examine whether policymakers are influenced by socially framed information about the Sustainable Development Goals (SDGs). While the findings do show that framing affects responses, the pattern of influence was not symmetrical. Policy professionals reacted far more strongly to negative framing than to positive framing. The increase in “very low impact” responses and the larger shift in average scores under the negative condition demonstrate that pessimistic narratives carried greater influence than optimistic ones. This imbalance indicates the presence of negativity bias, where negative information carries more cognitive and emotional weight than positive information of equal strength.These results are consistent with negativity bias in responses to socially transmitted policy signals.
These results suggest that policy professionals are not simply conforming to perceived popular opinion; rather, they appear especially sensitive to adverse signals about policy effectiveness. Even among trained professionals, negative information may be interpreted as more diagnostic, more urgent, or more credible than positive information. In high-stakes policy environments, this susceptibility has important implications. If negative narratives disproportionately shape judgment, policy decisions may become overly influenced by pessimistic framing rather than balanced evaluation. Recognizing this bias is essential to promoting more evidence-based and resilient decision-making processes.
References
Banuri, Sheheryar, Stefan Dercon, and Varun Gauri. “Biased policy professionals.” The World Bank Economic Review 33, no. 2 (2019): 310-327.
Baumeister, Roy F., Ellen Bratslavsky, Catrin Finkenauer, and Kathleen D. Vohs. “Bad is stronger than good.” Review of general psychology 5, no. 4 (2001): 323-370.
Hák, Tomáš, Svatava Janoušková, and Bedřich Moldan. “Sustainable Development Goals: A need for relevant indicators.” Ecological indicators 60 (2016): 565-573.
Kahneman, Daniel, and Amos Tversky. “Prospect theory: An analysis of decision under risk.” Econometrica 47, no. 2 (1979): 263-291.
LeDoux, Joseph E. “Emotion circuits in the brain.” Fear and anxiety (2013): 259-288.
Norris, Catherine J. “The negativity bias, revisited: Evidence from neuroscience measures and an individual differences approach.” Social neuroscience 16, no. 1 (2021): 68-82.
Rozin, Paul, and Edward B. Royzman. “Negativity bias, negativity dominance, and contagion.” Personality and social psychology review 5, no. 4 (2001): 296-320.
Skowronski, John J., and Donal E. Carlston. “Negativity and extremity biases in impression formation: A review of explanations.” Psychological bulletin 105, no. 1 (1989): 131.
Tversky, Amos, and Daniel Kahneman. “Availability: A heuristic for judging frequency and probability.” Cognitive psychology 5, no. 2 (1973): 207-232.
Vaish, Amrisha, Tobias Grossmann, and Amanda Woodward. “Not all emotions are created equal: the negativity bias in social-emotional development.” Psychological bulletin 134, no. 3 (2008): 383.
World Bank. World development report 2015: Mind, society, and behavior. The World Bank, 2014.