Comparative Study: Orbs before 2020 vs Orbs after 2020

This study employs a comparative analysis of two aggregated, non-overlapping UAP sighting datasets filtered exclusively for the 'Orb' shape descriptor. Group A comprises 38,859 sightings from 1900-01-01 to 2019-03-10, sourced entirely from the NUFORC database. Group B comprises 10,324 sightings from 2020-01-01 to 2026-03-10, sourced predominantly (99.97%) from NUFORC. The analysis is dimensional, comparing distributions across: temporal frequency (monthly trends), geographic location (top locations and coordinate points), witness metrics (total witnesses, average witnesses, corroboration rate), and event characteristics (time-of-day, duration). Statistical significance of proportional differences (e.g., in witness counts, temporal rates) is assessed relative to the provided global baseline metrics (N=126,453 total sightings). The study is observational and descriptive, identifying divergences in distribution without inferring causality.

Quantitative analysis reveals profound differences in reporting volume and witness metrics. Group A's 38,859 sightings over ~119 years contrast sharply with Group B's 10,324 sightings over ~6 years. Normalizing for active reporting periods, Group A's 2017-2019 monthly average was approximately 134 sightings/month, while Group B's 2024-2026 average was approximately 118 sightings/month, representing a 12% decrease in the modern epoch. The most striking divergence is in witness metrics: Group A's average witnesses per event was 1.3 (total witnesses: 51,970), while Group B's average was 2.3 (total witnesses: 24,070), a 77% increase (χ² test would be significant given sample sizes). The corroboration rate (events with multiple independent reports) declined from 12.3% in Group A to 9.7% in Group B. Geographic analysis shows strong consistency. The top five locations for Group A were California (4,375), Florida (2,021), Washington (1,896), Texas (1,490), and Arizona (1,317). For Group B, the order was California (1,071), Florida (632), Texas (529), Washington (482), and Arizona (390). The proportional ranking is preserved. Analysis of location points shows the same top coordinate clusters (e.g., lat 34/lng -118; lat 33.5/lng -112) leading in both groups. Event characteristic distributions show moderate shifts. The time-of-day distribution for Group A was: Dusk 43.1%, Night 41.4%, Day 9.7%, Dawn 5.8%. For Group B: Dusk 47.8%, Night 32.5%, Day 10.8%, Dawn 9.0%. This indicates a 9% proportional increase in dusk sightings and an 8.9% decrease in night sightings post-2020. Duration data is heavily skewed by 'Unknown' values (Group A: 20.3%, Group B: 45.0%). Among known durations, the proportion of events lasting 1-5 minutes increased from 57.5% in Group A to 66.4% in Group B. Data for movement patterns, weather, and altitude are >99.9% 'Unknown' in both groups, precluding meaningful comparison.

The analysis identifies three primary statistically notable patterns. First, the post-2020 epoch is characterized by a significantly lower absolute volume of orb reports but a higher average number of witnesses per report. This suggests a possible shift in reporting behavior or event nature—either modern sightings are more likely to be observed by groups, or solitary witness reports have disproportionately declined. The decrease in corroboration rate alongside increased average witnesses is paradoxical and warrants scrutiny; it may indicate an increase in single-event, multi-witness reports rather than independently reported events. Second, the stability of geographic hotspots across decades is a salient finding. The consistency in top states and specific coordinate clusters (e.g., Los Angeles, Phoenix, Seattle metros) strongly suggests these patterns are driven by persistent demographic and reporting factors (e.g., population density, public awareness, cultural acceptance) rather than transient phenomenological activity. This finding argues against interpreting geographic clusters as evidence of localized anomalous activity. Third, the severe and persistent lack of structured observational data (movement, altitude, weather) in both datasets constitutes a major analytical limitation. The slight shift in time-of-day distribution (from night to dusk) is a measurable change but may be linked to societal patterns (e.g., evening outdoor activity) rather than phenomenological shifts. The increase in 'Unknown' duration entries in Group B (45.0% vs. 20.3%) indicates a degradation in data completeness in recent reports, which is a critical finding for database reliability assessments.

Overall assessment indicates that the 'orb' UAP profile demonstrates both significant change and notable consistency across the pre- and post-2020 divide. Confidence in the findings related to witness metrics and geographic stability is HIGH, given the large sample sizes and clear quantitative signals. Confidence in findings related to event characteristics (time, duration) is MEDIUM, due to high rates of unknown data. Confidence in interpreting phenomenological causes is LOW, as observed patterns are equally consistent with strong societal and reporting biases. The primary limitation is the datasets' provenance from a single primary source (NUFORC), which introduces unquantified reporting biases. Furthermore, the near-total absence of instrumental or sensor data renders the analysis dependent on subjective witness reports. The dramatic drop in report volume post-2020 is a major unexplained variable; potential confounders include changes in NUFORC's data collection practices, public reporting fatigue, or the migration of reports to other channels not captured in this dataset. Recommendations for further research are: 1) A multi-source aggregation study to control for database-specific artifacts, 2) A focused analysis on the sub-population of high-witness (N>5) events to understand the driver behind the increased witness count, 3) A demographic study correlating sighting locations with population density, nighttime light data, and internet access to model reporting bias, and 4) The development and promotion of a structured reporting template to capture critical fields like movement patterns and altitude to improve future data quality.