FORT WAYNE – The major shift in the WTHR/Howey Politics Indiana Poll taken this week from earlier polls is crystal clear: Todd Young has closed the wide margin that Evan Bayh once had in the Senate race. The Gregg-Holcomb race for governor remains extremely close.  Donald Trump, as other polls have shown, has a large lead in Indiana.
    
Two months ago, when Evan Bayh first announced that he was going to run for the Senate, the first poll had him up 54-33%. Since that time he has been under heavy criticism for his Washington D.C. residency and lobbying ties in all forms of media. A Monmouth poll had Bayh’s margin down to 48-41%. The WTHR/Howey Politics Indiana Poll shows it has now fallen to 44-40%. However, the underneath numbers suggest more serious problems ahead for Bayh.
    
This poll surveyed 600 Indiana voters. Trying to get an appropriate mix to have an accurate poll and have even somewhat meaningful breakouts is difficult to do. Among the goals are reasonably accurate mixes based upon geography, age, party affiliation, race, sex, and education.  
    
Furthermore, the balances such as race and sex can dramatically make regional numbers worthless. For example, the breakout for the Fort Wayne metro area says that Gregg and Bayh will win our region, with Young doing slightly better than Holcomb.  I doubt either will win in northeast Indiana.      
    
Having read many polls for my own campaigns, those of Dan Coats and Dan Quayle, and many other candidates it is my opinion that this poll has all the categories fairly distributed with one exception. The education variable has become important because of the Trump phenomenon. This poll again shows how dramatically Trump support declines as education increases. It is my opinion that the pollsters, in part to test this, over-sampled those with college degrees dramatically.
    
I also know from past business polling I have been involved with that this variable has the greatest “lying” factor by those who take the polls: They exaggerate the amount of education they have. It is a classic problem. It is also always been assumed that those with more education will vote in greater numbers, an assumption that has increasingly been challenged. This poll included 28% with no more than a high school diploma. Official stats show that while Hoosiers have made gains in education, 44.7% of residents have not attended any college. Furthermore, it was closer among those with some college (27% of poll, 22% actual) than for those with college degrees (45% of the poll, 33.3% of the actual).  (“Actual” is from the 2010 census data.)
    
In other words, since Donald Trump does best among the less educated, his 43-36-11 lead over Hillary Clinton in this poll may be understated. The Monmouth poll last month had Trump ahead by 47-36%. In other words, the Trump-Clinton race may be frozen at this point.  6% of the Presidential voters claim to be undecided. If they are still undecided in this race, it is my opinion that those are likely non-voters.
    
This becomes a critical assumption. Ten percent of those called were undecided in the race for governor, and 16% of the voters are undecided in the decision for governor.  The way I used to break this out in analyzing my own polls, which when adjusted proved to be astoundingly accurate in spite of the Democrats and my primary claiming otherwise, is to look at the hard numbers as well as the percentages.  When one also does this with the cross-tabs, usually multiple key variables become clarified.    
    
In this case there are several.  One important variable is that more Clinton voters are decided on the Senate and gubernatorial races than Trump and Johnson voters.  This makes sense because Young and Holcomb are less well known to voters. But just because a point seems obvious, doesn’t always mean it is always true.  This poll establishes that the logic holds true in these races.
    
One would assume that this means that both Young and Holcomb have more to gain in the next two months than do Bayh and Gregg.  While the raw numbers are way, way too small to make definitive conclusions based upon the breakout data, they do wave some major caution flags for Republicans.
    
One can only speculate why (I have some theories but they can’t be proven by this data) but the numbers state that Bayh and Gregg both lead among the least educated. This could be good news for them since this poll under-sampled those voters. However, they both had much bigger leads among those with graduate degrees, who were over-sampled in even greater numbers.  In other words, it is a probable wash.
    
It does, however, illustrate where the battleground will be because across the board, more of the less educated voters appear to be undecided on the other races and make up a larger percentage of the undecided voters.
    
Here are some other comments from this Howey Politics-WTHR poll:
    
1. This poll dramatically shows that across the board that more libertarian votes are coming from Democrats, not Republican voters. This has certainly been hinted at by the national polling that shows Clinton doing better when she is one-on-one with Trump. This is counter to Republican assumptions.
    
2. The right-direction-wrong direction polling, especially for Indiana, has some small nuances that are interesting.  Holcomb gains among voters who think Indiana is doing well.  Young does better than Holcomb among those who think things are bad.  The core question there is: do some voters blame Bayh more than they do Gregg?
    
3. The surmised unfavorable (as in close to “detesting”) ratings for Hillary among all those not solid Democrats is clearly overwhelming the unfavorability ratings of Trump.  
    
As a final conclusion, when I went through the raw data breakouts, if I had to use this poll (which was a “moment in time” poll) as the predictor of what is to come, I would put Senate race (forcing the undecideds) at Bayh 48.6%, Young 46%, and 5.4% for the Libertarian. The gubernatorial race would be separated by less than 1%.  
    
In other words, Republicans should be thrilled with these results given the well known Democrats running against comparatively unknown Republicans. Young and Holcomb also appear to have more room to grow their numbers.
    
Souder is a former Republican congressman from Indiana.