XPost: alt.ufo.reports   
      
   EXECUTIVE SUMMARY:   
   - There is a large folklore connected with "UFO abduction".  While   
     some psychiatrists are sure this is a modern mass psychosis, some   
     others are not so sure.   
   - From FBI missing persons data we find there are significant   
     correlations between UFO activity and changes the number of missing   
     persons.   
   - Some categories of missing persons are not connected with UFO   
     activity.   
   - Apparently young and disabled persons are more likely to become   
     missing after an up-tick in UFO activity than people with abusive   
     partners or someone whose home is destroyed by an earthquake.   
   - The patterns of which missing person groups are associated with UFO   
     activity and which are not apparently associated suggests planning   
     may be involved.   
      
      
   So far we've looked at several ways UFO's apparently interact with the   
   world, mostly leaving zero or very slight footprints.  Apart from   
   apparently being connected with mass fish and dolphin deaths, increased   
   UFO activity so far does not seem to have any negative connotations.   
      
   But that may now change.   
      
   The connections between UFO's and missing persons is urban lore of   
   relatively long-standing. An unusually large number of people believe   
   they have been visited and sometimes abducted by "aliens".   
      
   While many aspects of the stories appear to be dream-like, and many   
   down-players point out an unusually large number of abduction cases   
   seem to occur when the victims were in bed, some high-0profile   
   psychiatrists have argued there are just so many cases and they have   
   so much detail in common they must refer to something real.   
      
   Harvard prof John Mack was initially interested in alien abductions as   
   an exercise in analysing an unusual mass psychosis perhaps similar to   
   the old European belief in "night mare"/nachtmerrie/nachtmahr -- a   
   demon or goblin that tormented people with bad dreams.   
      
   But after years interviewing patients he says he was reluctantly   
   forced to conclude many cases may be actual, if not misrepresented by   
   the patient, events.   
      
   Mack rode out years of ridicule from his peers for his ideas.  It's   
   said other psychiatrists yelled out at conferences whether he had been   
   declared insane yet.   
      
   But is there any evidence for the idea that UFO activity could be   
   connected with abductions or missing persons?   
      
   It's easy enough to check. The FBI maintains a missing person database   
   with data maintained on different categories of missing people. At   
   least since 2015 there are monthly data available. Before that there   
   are annual summaries back to the 1990s.   
      
   As in other studies I'll use the NUFORC database as a proxy for UFO   
   activity. I'll use a simple time-series regression s/w to match monthly   
   numbers of different categories of missing persons published in FBI   
   reports against monthly sightings from NUFORC.  At this point we wont   
   try to break down sightings by type, but subsequent posts will try to   
   point at characteristics of UFO sightings that seem to relate more to   
   missing persons than others. In some cases certain activity seems to   
   relate to the "return" of missing persons.   
      
   Seeing whether monthly UFO sightings "predict" different types of   
   monthly missing person reports finds:   
      
   Category	Filter	Trans	Binning	R2		\beta   
   INV		1.5	-	40	0.75219078	0.0973474   
   OTHER		1	-	10	0.39129340	0.592367   
   JUV		1	-	10	0.16165318	0.824093   
   DIS		1	log	9	0.06841812	20.3069   
      
   The "category" column is the FBI missing person category. In recent   
   reports there are 6 categories: juvenile, endangered, involuntary,   
   disabled, catastrophe, and other.   
      
   The software used very strong criteria to test whether UFO activity   
   predicted changes (either increase or decrease) in missing persons   
   month by month.  In 4 out of the 6 categories it found a relationship   
   that tested at 90% confidence or more in 2 statistical tests -- a   
   T-test on the TS-adjusted \beta and a rank test on the ordering on   
   data records by "x" compared with "y". In addition, if there are fewer   
   than 10 points in the final binning model it is discarded. The final   
   results are meant to be "convincing" as well as statistically   
   robust. :)   
      
   It seems the only categories of missing persons NOT apparently   
   connected with UFO activity are "catastrophe" -- missing people   
   associated with an earthquake, tornado, or a road train driving their   
   their bedroom by accident. And there is no apparent link with   
   "endangered" victims -- those known to have gone missing in   
   circumstances where they were in some kind of danger or threat from   
   significant other or other reasons.   
      
   The "most certain" model is the top one. "INV" means the missing   
   person disappeared in circumstances that suggest it was   
   involuntary. For every 100 UFO sightings in months between 2015-2020   
   there were about 10 involuntary missing persons. For comparison,   
   the average net INV per month over the period was around 30.   
      
   Let's look at the model in detail to see "how convincing" it looks:   
      
   (AUTO CORR CORRECTION; estimated rho = -0.163019)   
   y = 0.0973474*x + -15.7542   
   beta in 0.0973474 +- 0.0367351  90% CI   
   alpha in -15.7542 +- 23.6245   
   P(beta>0.000000) = 0.999424   
   r2 = 0.75219078   
   calculated Spear man corr = 0.745455   
   Critical Spear man = 0.564000 2-sided at 5%; reject H0:not_connected   
      
   Bin label             av #UFO         av #new        model-estimated   
                         sightings/mo    missing "INV"  new missing INV cases   
      2019.12                242.981       33.625      7.89937*   
      2018.79                299.559          1.5      13.4071   
      2019.29                360.856      16.9231      19.3742   
      2018.54                443.188      27.5833       27.389   
      2019.62                505.259         28.7      33.4315   
      2016.62                564.114           22      39.1608*   
      2019.71                 621.36        59.75      44.7336*   
      2019.88                 688.57      61.3333      51.2763   
      2019.96                739.625        45.25      56.2464   
      2015.88                886.832           73      70.5766   
      
   The s/w is a binning time-series regression. It collects the data into   
   similar bins -- in this instance months between 2015 and 2020 -- and   
   averages all the points allocated to that bin. This produces an   
   improved or smoothed estimate of "x" and "y" for 1 case in the   
   regression. The binning operates to maximize the R2 robustly.  The R2   
   number shows what fraction of the "y" (INV cases numbers per month)   
   data change in the same way as the "x" data (UFO sightings per month).   
      
   The "*" at the end of some lines shows points that are not close to   
   the regression line. It turns out only 3 out of 10 points are not   
   "statistically close" to the trend line. The T-test says this can only   
   happen by luck 0.01% of the time.  The Spearman test says the ordering   
   of the data by the missing person column is so similar to the data   
   ordered by the UFO sightings column it could only happen 10% of the   
   time just by luck.  Together the 2 tests point at a real association   
   between UFO sightings and a small but measurable "involuntary"   
   missing persons per month is highly likely.   
      
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)