Blocking Bad News to Keep Groups Stable

In the paper's words

Social systems frequently encounter truthful yet destabilizing information that can trigger panic, market crashes, or rapid opinion polarization. Existing interventions such as link deletion or bot mitigation are blunt and only partially effective. This paper fills the gap by constructing an analytical multilayer network model of belief exchange and introducing a functional ignorance heuristic that probabilistically suppresses the transmission of high‑impact truths.

The abstract, verbatim; every line below is quoted from the text.

What it claims

that a suppression probability exceeding 2/3 guarantees subcritical cascades for R₀=3—would be falsified if an empirical field experiment measured a transmission probability q>2/3 yet observed sustained supercritical cascade growth (e.g., cascade size growing exponentially over generations). Such an observation would indicate that the mean‑field assumptions or the reproduction‑number formulation do not capture essential dynamics.

Numbers it reports

1. The modest impact of link deletion reported in [4] (50 % cascade reduction after 50 % link removal) aligns with our elasticity calculation and motivates the search for more efficient interventions.

   from 6.3 Relation to Prior Work

2. Elasticity of link deletion: $$ e_{\text{del}} = \frac{r_{\text{del}}}{f_{\text{del}}}= \frac{0.5}{0.5}=1. $$ To achieve subcriticality for $R_{0}=3$ via link deletion alone, one would need to remove a fraction $$ f_{\text{req}} = 1-\frac{1}{R_{0}} = \frac{2}{3}\approx 0.6667, $$ exceeding the 50 % deletion level already identified as practically difficult in [4].

   from 4.4 Efficiency Comparison with Structural Deletion

3. An analytic suppression‑threshold condition $q\gt 1-1/R_{0}$ and its evaluation for $R_{0}=3$ yielding $q^{\star}=2/3$ (Section 4.1).

   from Contributions

4. An efficiency comparison with structural link deletion, highlighting the multiplicative advantage of functional ignorance (Section 4.4).

   from Contributions

5. For the higher‑baseline case $R_{0}=10$, the threshold would be $q^{\star}=1-0.1=0.9$.

   from 4.1 Suppression Threshold

6. Assumptions: cost of suppression $c=0.1$, shock probability $p=0.2$, shock damage $L=1$.

   from 4.2 Evolutionary Stability

How it is done

1. Finally, we compare endogenous suppression with structural link deletion, showing that the latter’s elasticity is approximately one, whereas functional ignorance achieves a multiplicative reduction in cascade growth.

   from Abstract

2. This demonstrates that aggressive link deletion is costly and only partially effective, motivating a more nuanced approach that targets specific destabilizing messages rather than indiscriminately pruning the network.

   from 2. Background and Related Work

3. We define the elasticity of an intervention as the ratio of relative cascade reduction to relative intervention size.

   from 3.7 Efficiency Metric

4. The main claim—that a suppression probability exceeding $2/3$ guarantees subcritical cascades for $R_{0}=3$—would be falsified if an empirical field experiment measured a transmission probability $q\gt 2/3$ yet observed sustained supercritical cascade growth (e.g., cascade size growing exponentially over generations).

   from 6.2 Falsifiability

What would falsify it

1. The main claim—that a suppression probability exceeding $2/3$ guarantees subcritical cascades for $R_{0}=3$—would be falsified if an empirical field experiment measured a transmission probability $q\gt 2/3$ yet observed sustained supercritical cascade growth (e.g., cascade size growing exponentially over generations).

   from 6.2 Falsifiability

2. If a principled mechanism existed that could predict when selective ignorance improves collective outcomes, designers could embed such mechanisms into moderation algorithms, corporate communication policies, or emergency‑response protocols.

   from 1. Introduction

3. If cascade size remains supercritical at $q=0.66$ (contrary to the model prediction), the main claim would be falsified.

   from (c) Smallest Next Experiment

Where the claim stops

1. The model assumes a uniform suppression probability $q$ across agents and layers; in practice suppression may vary by context or platform.

   from 6.1 Limitations

2. Values for $p_{\text{gen}}, d, c, p, L, C, \delta$ are assumptions for illustrative purposes; different settings shift thresholds.

   from 6.1 Limitations

3. Agents do not adjust $q$ over time; learning or norm evolution could alter outcomes.

   from 6.1 Limitations

Where to read

1. Abstract186 words · 1 min
2. 1. Introduction267 words · 1 min
3. 2. Background and Related Work261 words · 1 min
4. 3. Methods354 words · 2 min
5. 4. Analysis311 words · 1 min
6. 5. Results136 words · 1 min
7. 6. Discussion284 words · 1 min
8. 7. Practical Implications and Engineering Pathways344 words · 1 min
9. 8. Conclusion135 words · 1 min
10. Provenance and AI Involvement225 words · 1 min