Recent data on political violence often relies on media reports, potentially introducing reporting bias. This issue is especially relevant for studies involving communication technology since these technologies can influence both conflict outcomes and how they're reported.
Using the effect of cellphone coverage expansion as a case study, this paper provides a quantitative assessment of reporting bias at a micro-level. By comparing event reports from media outlets with data from military sources, we find that higher levels of cellphone access in certain regions correlate with increased media reporting of violence, without necessarily increasing actual violent incidents.
Key Finding: Reporting rates vary significantly between regions with and without widespread cell coverage, even when controlling for actual violence.
Our analysis suggests that previous findings linking cellphone infrastructure to conflict may be partially driven by this reporting effect rather than direct causation. We offer a straightforward diagnostic approach to help identify such biases in similar studies focused on technology's impact across Africa.
Method: Quantitative Assessment Comparing Media & Military Sources
This simple procedure helps distinguish whether observed effects are genuine or artifacts of reporting patterns when examining cellphones and violence dynamics throughout the continent.
| A Closer Look at Reporting Bias in Conflict Event Data was authored by Nils B. Weidmann. It was published by Wiley in AJPS in 2016. |
