(For most postings in the series see: ).   
      
   EXECUTIVE SUMMARY:   
      
   - Carefully comparing UFO sightings of various types shows US military   
     spending seems to deter subsequent activity.   
   - It also seems certain types of activity see increased military   
     spending subsequently.   
   - The result appears to be either a knowing or unknowing low-level   
     cold war between the US military and UFO's.   
   - Results seem to be consistent across the various datasets subject to   
     some countries living inside the US military umbrella and some   
     outside; some countries have a more favored position under the   
     umbrella than others, possibly relating to assets on the relevant   
     national territories.   
   - There is preliminary data showing there has been a "gear change" in   
     the past 10-20y. Where increased military spending might have   
     deterred certain activity in the previous 50y more recently it can   
     have the opposite effect.   
   - The data seems to support oftentimes unattributed claims or rumors   
     of confrontations since at least the 1930s between military or   
     civilian aircraft and unknown objects and even close-order   
     "fistfights" between various militaries and UFO's or unidentified groups.   
      
      
   We've looked at the relationship between US military spending and   
   patterns in UFO sightings before.  At one point we found military   
   spending seems to predict certain types of sighting in certain   
   locations, suggesting UFO sighting data does seem to include aircraft   
   the US military may have been developing over the past 50+years. There   
   was no indication the suspected aircraft used any unconventional   
   technology, but this might be another thing to examine at a later date. :)   
      
   After the AI s/w threw up an anomaly I've managed to smooth out the   
   results linking military spending on UFO sightings in several   
   countries, and UFO sightings on military spending.   
      
   It seems there is a consistent effect of military spending on UFO   
   sightings -- it generally declines in parallel to increasing in   
   spending and increased when spending decreases. Whether this is a   
   deliberate strategy is not clear. But there seems to be a de facto   
   cold war between the US military and UFOs of almost all kinds.   
      
   A wrinkle in the pattern comes from the US's close military ties with   
   some countries and lesser ties with others. "Allied" countries   
   generally see the same reduction in UFO activity when US military   
   spend goes up. But we can also detect a "move on" effect where the   
   opposite is true for some countries that don't apparently come under   
   the US military umbrella.   
      
   On the flip side we can see increased sightings of certain types are   
   apparently treated as "threats" by the US govt and invariable result   
   in greater military spending in later years. When sightings go down as   
   a result of "whatever" that spending goes toward, then spending tends   
   to relax again.   
      
   These patterns are very consistent across all types of UFO's -- I   
   generally divide them up by location-of-observation (state or country),   
   color of whatever lights are seen, and shape of whatever objects.  The   
   time of day (in my work generally "dawn" aka eastern sky, "dust" aka   
   western sky, "daylight" and "nighttime") is another way the classify   
   sightings that produce interesting groups of events.   
      
   In the work below I'm mostly using the NUFORC dataset upto the end of   
   2021 now.  Some other data comes from curated data series for the UK,   
   Canada and France (GEIPAN).   
      
   As this project proceeds -- it's only approx 1y old now :) -- I'll   
   continue to comb the 'net for sources of other sightings or category   
   of event. As we know, many unusual things tend to link up and are   
   suspected of being connected in one way or another with UFO activity.   
      
   The s/w I'm using here takes year-by-year event counts of various   
   classes of UFO sightings and compares them with US military spending   
   as a% of GDP. The data is massaged to remove various problems   
   (e.g. the NUFORC dataset changes gears in March 2006 when it moved   
   from telephone/mail reporting to a web-based report form; not only did   
   report numbers go up almost immediately an order of magnitude but the   
   type of things reported radically changed) and the resulting files   
   compared using very severe statistical s/w.  If a result is not   
   better than 99% statistically certain the s/w will ignore it.   
      
   The s/w allows for a final level of "noisy data" (i.e. years that seem   
   to be much different from the overall trends are dropped, making sure   
   only a small proportion is finally ignored) and time-shifts data to   
   find the best alignment. The effect of some change may take several   
   years to show up to full effect; the s/w tends to ignore the   
   time-shifts that are less than the best it can find.   
      
   The best way to rank results is by the so-called "explanation power"   
   of the underlying statistical model. The "R2" statistic shows what   
   proportion of the "effect" variable is predicted by the "cause"   
   variable (possibly time-shifted).   
      
   Let's first look at the best overall results here.   
      
   Model                Lag     R2              Beta    (90% CI)   
   (mil spending vs:)   
   Other                4      0.81047451   -0.224577  0.0226569   
   white                1      0.79688464   -0.235496  0.0242267   
   AUSTRALIA            6      0.74702442   -1.31413   0.287154   
   gr[ae]y              3      0.73177617   -0.202808  0.0258248   
   dusk                 2      0.71641154   -0.166723  0.0217102   
   AZ                   9      0.67127955   -0.268146  0.0444731   
   Circle               1      0.65660639   -0.183243  0.0276467   
   Sphere               4      0.65109060   -0.150306  0.0229554   
   MO                   5      0.61291093   -0.163303  0.029277   
   notlight             1      0.61047755   -0.172894  0.0277193   
      
   We first note the top10 results include sightings based on US states,   
   one is the national total for Australia compiled from the independent   
   assessments of groups in each state, some are based on (mostly   
   US-based) NUFORC object color (incl US and UK spellings) or shape, and   
   the last is a category that excludes from the NUFORC annual totals   
   those objects curators classified as "lights in the sky" -- supposedly   
   the most error-prone UFO type.   
      
   We see in all cases that increases in US mil spending are associated   
   with decreases in those types of sighting. The Beta shows approx what   
   percent/fraction reduction is associated with a 1% GDP increase in US   
   military spending. E.g. for 1% GDP extra US military spending the   
   category "Other" (any shape that doesn't fit a usual NUFORC shape   
   category) sees a 22% decline after 4 years (the Lag column).   
   Contrariwise a decline of US spending generally sees an increase in   
   those kinds of sightings.   
      
   The effect of US military spending on sightings in Australia is even   
   more profound. The Beta of "-1.3" actually indicates an almost 4-fold   
   decline for each 1% of GDP in spending or similar increase for reduced   
   spending. The lag for this link is around 6 y. It seems Australia   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)