First he takes the polling into account:
This week's generic ballot tracking poll from Gallup shows Republicans with a 5-point lead, tying their previous best showing from 1994. The previous two weeks revealed a 6-point lead (the largest in 50 years of Gallup tracking) and a tie.
SNIP
Rasmussen's tracking poll showed Republicans opening up a 10-point lead in the generic ballot. This is significantly larger than the .4 average that RCP currently shows. But Rasmussen is the only pollster who has imposed a likely voter screen since mid-May. Other pollsters will add likely voter screens later in the year, and that usually moves the ballot toward Republicans.
SNIP
But the worst news for the Democrats comes from NPR's recent polling. Using a top-notch Republican and a top-notch Democratic polling firm, NPR polled 60 districts represented by Democrats that it considered the most competitive. It isn't a simple generic ballot - it named the actual candidates where incumbents were running. It further broke these down into two "tiers:" Tier I (the 30 Democratic districts it considered "most competitive") and Tier II, (the next most competitive 30 Democratic districts). NPR also polled the ten districts represented by Republicans that it considered the most competitive.
Then Trende factors in Obama's approval ratings:
Obama's approval is horrible in both "tiers" of districts represented by Democrats. 53% of voters disapprove of him in the "Tier I" districts, while 56% disapprove in the "Tier II" districts, including a near-majority who strongly disapproves of the President. He's 50-50 in the districts held by Republicans, but Obama averaged a 16.15% victory in these districts in 2008 (the districts polled that are held by Democrats were carried by McCain by a point or so on average in 2008).
Against this landscape, it should not be surprising that NPR finds a Democratic debacle in the making in these districts. Voters prefer, on average, a Republican to the Democrat by 9 points in "Tier I" districts and by a 2 point margin in the "Tier II" districts. Among the most enthusiastic voters, it is even more ominous for Democrats: Republicans lead by 14 points in the 60 districts represented by Democrats. In the districts held by Republicans, by contrast, the Republican lead by 16 points over all, and by 21 points among the most enthusiastic voters.
Before getting to the conclusions he draws from the polling, I am skipping down to another area of concern for Democrats, which is their message itself:
What is perhaps the worst piece of information for Democrats is that they are not likely to march into November with a message that they will be able to rally voters around. NPR tested Republican and Democratic themes, and found the Republican theme winning in almost every instance. Near-majorities of voters in both tiers of districts represented by Democrats "strongly" believe that Obama's economic policies have done nothing to stop the recession, and have only run up the deficit. The Republican message on spending beats the Democratic message by twelve points, on the economy by about ten points, on health care by about ten points, and on Wall Street reform by about ten points.
His conclusions, factoring in all the points above, seem reasonable and justifies his predictions:
Three conclusions follow from this. First, very few Republicans seem likely to lose in 2010. Second, if we use a spitball estimate that Republicans will win 2/3 of the competitive Democratic districts, which seems reasonable if the President is averaging a 40% approval rating there and Democrats are behind by 5 points on average in these districts, that nets the Republicans 40 seats right there.
But perhaps most importantly, what does this say about the next 30 or 60 Democratic districts, which were not polled? If the Republicans are winning handily overall in both the seats commonly thought to be highly competitive and in the seats thought to be less competitive, that seems to imply that there are probably a lot of races that aren't on anyone's radar screen that are competitive. I'm not saying that Republicans would win 2/3 of these districts or that they are winning there by ten points, but even a 30% win rate in the next tier of seats gets the Republicans to 50 seats overall. And I'm guessing their win rate would be higher, given how poorly Democrats are faring in their Tier I and Tier II seats.
Read it all, I left out certain specific examples proving his points and the whole piece is well worth the read.
.