What looked like a typical bit of Pentagon mis-direction has been   
   widely publicized across the usual and fringe media -- in "a secret   
   report" at least one woman claims to have an unwanted pregnancy as a   
   result of a UFO or alien encounter.   
      
   Like many people (who consider themselves "serious minded" :) I   
   dismissed the "I have an alien baby!" story. Most of the release of   
   material from the relevant FOI seemed to be the usual noisy junk   
   needing some serious analysis to extract anything useful from.   
      
   But this is where "AI" meets the kind of criteria written up by   
   Harvard's Avi Loeb. "AI" offers science or mil int the prospect of   
   some kind of "unbiased" or at least "uncluttered" analysis of any   
   topic, no matter how far fetched.   
      
   While I've had some experience using AI for serious analysis of   
   complex topics and find it's not quite as "unbiased" as many might   
   hope, we can offer here an example of "unprejudiced" analysis of the   
   sensational topic of possible "alien" babies. ;)   
      
   If it is the case that "UFO's" have something to do with "unwanted   
   pregnancies" can we find any evidence for that in hard   
   publicly-available datasets?   
      
   The answer is always yes. :)   
      
   Here's we look at state abortion rates per 1000 women as an indicator   
   or proxy for unwanted pregnancies. And we just do some simple   
   regressions against different types of UFO data to see if there is any   
   kind of statistically significant hit.   
      
   And the s/w immed picks up some hits. Depending on how much the data   
   is "tortured" by advanced stats there is almost any level of support   
   for *some* kind of link.   
      
   At this point we can't claim there is evidence for a "causal" link,   
   but at least we all have a mental model of how a UFO encounter *might*   
   result in an unwanted pregnancy.   
      
   But here are the stats.   
      
   The UFO data I will present here are rates of UFO sightings across US   
   states from the NUFORC dataset. I thank again the NUFORC for proving   
   these numbers online without let or hindrance or requests for monetary   
   assistance.   
      
   The NUFORC created a web report form in 2006 possibly related to some   
   event in 2004. The character of the UFO reports changed radically with   
   the advent of the form, so I will split the state-by-state data into 2   
   sets -- one pre 2006 and the other post 2006 (up to the end of 2021 in   
   the present dataset).   
      
   Each finds there is a strong statistical link. The more UFO's each   
   state sees per km2 the more abortions per 1000 women each state   
   records (the abortion rates come from 2022, but we assume at this   
   point they are indicative of historical rates).   
      
   Both the pre- and post-2006 data gets a strong hit. But "strong" I   
   mean my usual standard 2 tests -- a T-test on the relevant regression   
   \beta, and a rank test on the ordering of the data by the dependent   
   ("Y") and independent ("X") values of the data.  If both tests pass at   
   less than 10% "could be just luck" then we have a "3 9s" to "4 9s"   
   likelihood there is some link.   
      
   The pre-2006 data looks like:   
      
   MODEL:   
   y = 1.75914*x + 7.53293   
   beta in 1.75914 +- 0.566151  90% CI   
   alpha in 7.53293 +- 1.23503   
   T-test: P(beta>0.000000) = 0.999998   
   Rank test: calculated Spearman corr = 0.556110   
   	Critical Spearman = 0.432000 2-sided at 1%; reject H0:not_related   
   r2 = 0.361354   
      
   State		UFO sightings/mn km2  Abortion	  Model   
   				      rate/1000w   
   Alaska                 0.0244491         8.6     7.57594   
   West.Virginia          0.0710519         1.3     7.65792*   
   Wyoming                0.0710519         1.3     7.65792*   
   New.York                0.131067         7.9      7.7635   
   Missouri                0.136541         8.3     7.77313   
   Rhode.Island            0.150194         3.1     7.79715*   
   Idaho                   0.231004         3.9      7.9393   
   Montana                 0.234582         5.5      7.9456   
   Tennessee               0.245581         4.4     7.96494   
   Nebraska                0.247947        16.4     7.96911**   
   Kansas                  0.295641        12.2     8.05301   
   New.Jersey              0.301669        11.7     8.06361   
   Iowa                    0.329347         6.3      8.1123   
   Colorado                0.415429        10.9     8.26373   
   North.Dakota            0.441903         6.2      8.3103   
   Michigan                 0.45743        10.1     8.33762   
   Arkansas                0.479191         5.5      8.3759   
   Ohio                    0.490571        11.9     8.39592   
   South.Dakota            0.494522         9.4     8.40287   
   Minnesota               0.566035         4.3     8.52867*   
   Washington              0.607172         5.9     8.60103   
   Wisconsin               0.607172         5.9     8.60103   
   Louisiana               0.618185        10.6     8.62041   
   Alabama                 0.677642         6.4       8.725   
   Arizona                 0.718026         9.2     8.79604   
   Mississippi             0.747786           4     8.84839*   
   Maine                   0.752897         8.8     8.85738   
   Texas                   0.763022        11.4      8.8752   
   South.Carolina          0.778738         9.2     8.90284   
   Georgia                 0.864147        16.9     9.05309*   
   Virginia                0.892359         4.4     9.10272*   
   Massachusetts           0.898225        14.2     9.11304*   
   Utah                     1.01999        10.2     9.32724   
   Hawaii                   1.02434          12     9.33489   
   Kentucky                 1.06059         3.8     9.39866*   
   New.Mexico               1.09047        14.6     9.45122*   
   Pennsylvania             1.21787         5.3     9.67534*   
   Indiana                  1.28285         5.9     9.78965   
   Illinois                 1.34668        16.6     9.90193*   
   Vermont                  1.43503        12.1     10.0574   
   California                1.7572        16.4     10.6241*   
   Florida                  1.82027        18.6      10.735*   
   Oklahoma                 1.91142        13.1     10.8954   
   North.Carolina           1.96389         9.4     10.9877   
   Nevada                   2.35382         9.2     11.6736   
   Delaware                 3.25733        10.5      13.263   
   Maryland                 5.01171        13.5     16.3492   
   New.Hampshire            6.55215          28     19.0591**   
   Connecticut               7.1742        17.7     20.1534   
   Oregon                   8.24588        16.7     22.0386*   
      
      
   The 2 stats tests say there is a strong link. We cant see UFO's   
   *cause* some unwanted pregnancies but the R2 is the important value   
   here. It says the state-by-state difference in UFO rates/km2 predict   
   about 36% of the state by state variation in abortion rates.   
      
   And the R2 is important. Because that's almost exactly the same   
   statistic from the post-2006 dataset. Despite the huge changes in UFO   
   observing methodology, the "correlation" has remained the same,   
   presumably over the past ~100y.   
      
   Here's the data for the post 2006 stuff:   
      
   MODEL:   
   y = 0.142532*x + 7.39118   
   beta in 0.142532 +- 0.0440032  90% CI   
   alpha in 7.39118 +- 1.23117   
   T-test: P(beta>0.000000) = 0.999999   
   Rank test: calculated Spearman corr = 0.561585   
   	Critical Spearman = 0.432000 2-sided at 1%; reject H0:not_related   
   r2 = 0.380761   
      
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)