Your right Iowa was 7%, not 8% but that actually strengthens my point. If only 2% out of 7% is Jill Stein/write-in (and write-in isn't indicative of anything other than dissatisfaction with the process) and the others are all libertarian and right-wing then how do Democrats come up with this lopsided amount of 2016 third party voters?
While there are some 3rd party voters who are well informed and know who they are voting for, there are quite a few others who are lower-information voters more likely to be disengaged/uninterested in the political process and they vote 3rd party as a protest vote, without necessarily agreeing with who they are voting for. 3rd Party votes can be and often are simply rejections of main parties rather than affirmations of the 3rd party, and voters may simply select whatever random 3rd party they have heard of (name ID). There are others who are more informed and normally in the past have voted for one party, but can't stomach voting for them in a particular election (for example, former Rs/R leaners who voted Libertarian). Such a voter may vote Libertarian or even Constitution Party or something like that, but after a few years have passed (and after it has become clearer to them that it is a 2 party system and if they don't want R, in practice they have to vote D) and they find themselves feeling more and more "former" R, then even if they don't necessarily like everything about the Dems they may be much more likely to vote Dem than to go back to voting for the Rs that they are increasingly unhappy with.
The reason 2018 was the way it was because Democrats turned the out and Republicans did somewhat but disproportionately the people who sat home were the Obama/Trump types, skewing without a college degree and higher approval than Trump than the electorate that voted.
This is true or false to varying degrees in different states. In the case of Iowa, it is not true that most (much less all) of why 2018 was the way it was is because of turnout.
About 60% of the difference in the result of IA-GOV 2018 as compared to the 2016 Presidential case resulted from changes in candidate/party preference. Only about 40% of the change resulted from turnout differences. True, 40% is still 40%, and that is definitely significant and meaningful, but it was not the only (or dominant) factor.
See here for the data the above is based on:
https://medium.com/@CatalistAnalytics/what-happened-in-the-iowa-gubernatorial-election-b4638ae596b9The electorate in Iowa in particular wasn't actually that favorable for Dems in particular in terms of age. The % of young voters was definitely low in comparison to 2016. This is particularly important in Iowa because IA has a large college/university presence, and college towns/students form a relatively big part of the Dem base in Iowa as compared to many other states. IA is actually the #1 state (2nd if you count DC as a 'state') in terms of student population as a share of the overall population. Young voter turnout was pretty good for a midterm (i.e. a bit better than 2010/2014), but it was still midterm turnout (much lower than 2016) and the age 65+ vote share shot way higher than in 2016. IA also has a larger than average senior population, which basically makes IA one of the most polarized states in terms of age - there are lots of youngs and lots of olds, but not many middle aged people. So if young turnout is not good, then relatively the Senior vote share spikes, and given that Rs are doing very well with Seniors, that basically means a GOP win.
https://www.huffpost.com/entry/states-college-student-population_n_561b2ed4e4b0082030a30bfc
