Considering the R trend we've seen in Tennessee over the past 20 years, that's not surprising.
It's not really an R trend issue. Kids mostly vote like their parents until around age 18-21, when they have in the US recently developed a pattern of shifting rapidly to the left upon adulthood.* So R+14 among kids just reflects their parents - which isn't necessarily representative of Tennessee as a whole, either, since (i) parents are more Republican than non-parents and (ii) it's counting by child rather than by parent, so parents with multiple children are overcounted (and having more children correlates with being more Republican, mainly because it correlates with religiosity). Plus, there's of course no effort to make these types of polls of children representative in general, so they probably overpoll middle class and wealthy and heavily white school districts and children and underpoll poor and heavily minority school districts and children.
*This is also why we get the constant "The next generation is super Republican!!" memes. Because 18-year-olds tend to be closer to their parents than 19-year-olds, and 19-year-olds more so than 20-year-olds, etc., so that there is always a dip back towards the Republicans among the very youngest voters, yet that dip isn't sustained as young voters age into their 20s.
