I've been looking into this a little bit but I wasn't actually planning on posting anything about it bc I am missing too much info. but here's what I've thought about (forewarning: I do not make any conclusions).
Seems like nowadays, there are 40ish, maybe as many as 50ish (if you include children never recovered), stranger abduction+murders of children under 14 a year on average in the U.S. (it used to be somewhat higher in previous recent decades).
Using the percentage of population of Texas of the US, statistically, we would expect about 4 stranger abduction murders per year in that state. These are presumably clustered geographically in the population centers. There are 4 major centers in TX, one being Dalls/Ft Worth (ranking 2nd of the 4), so we would statistically expect roughly about 1 of these crimes a year in that general area.
I don't know the actual statistics for the area. A significant deviation from baseline would be interesting.
The Opal Jennings case is very close to Alanna's geographically, with the same age and similar appearance, but the amount of time between them makes it less compelling. But there is possibly a good reason to account for a hiatus: a jail term. In crimes like these, "The suspect in the case had been arrested for prior offenses against children 53% of the time and the most common crime was sexual in nature." (
http://www.trutv.com/library/crime/criminal_mind/psychology/child_abduction/3.html) so it's quite possible the criminal is an RSO who was in jail for most of those gap years.
Regarding Amber's case, once you open up the geography and age more broadly, you would probably expect to turn up cases over the years that were that similar. Brunettes are common.
But one thing I don't know, that is very relevant , is how many crimes, on average, a child abductor commits. Are serial killings the norm? That makes it more likely that any two cases are connected. Also, if we wanted to analyze these specific probabilities more closely, one needs to know the probabilities that each of these things remain constant for serial killers (appearance, age, geography, etc) as well as a likely time horizon.
In sum, I didn't get anywhere with this line of thought, but I decided to throw it out there in case anyone else wants to go further.