Oh, happy days are here again!  According to the left-wing media, Democrats may be able to defy the odds and keep control of both the Senate and the House in November.  You can read about it here, here, and here.

There is a hair of truth mixed into Democrat optimism.  Joe Manchin's sell-out on the Inflation Reduction Act, AKA the Green New Deal, was a multi-billion-dollar boon for green energy liberals like Bill Gates.  Likewise, the Supreme Court overturning Roe is a fundraising boon for leftist politicians who want to campaign on abortion.  Even Joe Biden benefited — although it took all of the above plus positive news on unemployment, the passage of the Chips Act, a downward drift in the price of gasoline, and a slight decline in the inflation rate to create a five-point dead cat bounce in his approval numbers.

The recent House special election in New York District 19 for the unexpired term of Antonio Delgado is considered a positive for Democrats.  The interesting part is that Republican Mark Molinaro, the loser, will run against a different Democrat in a redrawn version of District 19 in the November election.  In addition, Democrat Pat Ryan will run against a different Republican in District 18.  With these unique circumstances and turnout at only 25% versus 70% in the 2020 general election, this seems less like a bellwether and more like a who-cares election.

So before the Democrats prematurely raise the hammer and sickle flag over Congress for the next term, we should look at some facts.

Much of the supposed good news for Democrats may not matter.  According to The Washington Post, abortion and climate change were dead last in a mid-July Economist/YouGov poll of 14 midterm election issues.  Jobs and the economy were first.  A CBS poll taken in late June had a similar result.  Voters' primary concern was inflation, followed by the economy, crime, gun policy, and immigration.  Both polls were taken after the Supreme Court decision on abortion.

Ditto for the July jobs report and the July unemployment rate of 3.5%, as reported by the Bureau of Labor Statistics.  Of the 528,000 non-farm jobs added in July, 331,000 were full-time jobs cut to part-time.  And 92,000 additional people were listed as working multiple jobs, possibly to pay the inflated prices for food, rent, and gasoline.

Even though the unemployment rate fell, it is unclear if the country is in a recession or will be forced into one by the Federal Reserve raising interest rates.  According to the New York Post, half of U.S. companies have either enacted hiring freezes or are preparing to lay off employees in fear of a coming recession.

Then there is something most consumers do not know is coming.  The price of natural gas, already in record territory at roughly $9.50 per Mcf, could rise as much as 60% this winter.  To put it another way, if your gas bill was $180–200/month last winter, expect $250–300/month this winter, thanks to Biden's energy policies.  Since natural gas is used to generate a third of U.S. electricity, expect prices for electricity to climb as well.  This means that inflation is not going away anytime soon.

Okay, what about Democrat Senate candidates leading in the polls?  A Marquette University Poll showed Republican Senator Ron Johnson behind Mandela Barnes by 7 points.  And a University of North Florida Poll indicated that Democrat Val Demings had a 48–44% lead over Senator Marco Rubio.

All polling organizations, like the ones above, try to develop accurate models of the electorate.  But they are often affected by various types of bias that skew the results.  Unfortunately, most of the bias is in favor of Democrats.

Survey weighting bias is one example.  You can bias your survey if the model overcounts or undercounts any particular demographic.  This includes categories such as age, sex, race, party affiliation, education, etc.

Polls of likely voters are more accurate than polls of registered voters.  The main reason is that not all registered voters vote.  Pollster Nate Silver believes that midterm polls of registered voters have a 2.6-point Democrat bias because Democrats are not as reliable about voting as Republicans.  The Marquette and the North Florida polls both use registered voters, which may account for part of the Democrat lead.

There is something called social desirability bias, which reflects a person's fear of being judged by others.  Voters may alter their answers to polls so as not to seem racist or judgmental.  In the Marquette survey, 41% of voters would not answer the favorability question for Mandela Barnes, a black man, versus 15% for Senator Ron Johnson.  Is this an example of social desirability bias?

Then there is partisan non-response bias.  This involves an active choice on the part of conservative voters, who may refuse to answer polls or identify as conservative because they do not believe polls are fair or fear harassment from liberals.  For example, would you risk answering a political survey if you are a conservative who works for a woke organization, a union member, or a student attending a liberal college?  This type of bias could amount to 3% or more in heavily conservative areas.

Okay, polls may be biased in favor of Democrats, and socialists may be counting their chickens before they hatch.  Why else do I have faith in a red wave?  There are seven reasons.

First, more than a million voters in 43 states switched to the Republican Party during the past year.

Second, according to Gallup, in early 2021, 49% of the electorate identified as Democrat or leaning Democrat, while only 40% identified as Republican.  By late 2021, the parties had switched places, with 47% identifying as Republican and only 42% as Democrat.

Third, there is something called midterm turnout bias, which favors Democrats when a Republican occupies the White House and vice versa with a Democrat president.  In midterm elections since 1978, Republicans have had an average turnout advantage of 5%.

Forth, according to Real Clear Politics, roughly 70% of voters believe that the country is on the wrong track.  The latest Economist/YouGov poll has a 25/63 split between right track/wrong track, despite having an unweighted 11-point bias in favor of Democrats.  The Reuters/Ipsos poll had a 20/70 right/wrong split, despite having a 46%-to-39% bias in favor of Democrats.

Fifth, although the Real Clear Politics average of the generic congressional vote has Republicans up only by a fraction, the Rasmussen and the Trafalgar polls have Republicans up by five.  Both are polls of likely voters, which are more accurate than the registered voter polls.

Sixth, until I see evidence to the contrary, I firmly believe that polls are missing between 2 and 3% of conservative voters due to partisan non-response bias.  See above.

Seventh, the FBI's raid on former President Trump's home at Mar-a-Lago has increased voter enthusiasm for 83% of Republicans and 72% of independents.  And Republicans didn't have to spend a dime!

So if anyone sees Mitch McConnell crying in his beer, tell him to snap out of it and man up.  That shadow he sees on the horizon is not the ghost of Karl Marx.  It's the red wave rising.

Image: tom.arthur via Flickr, CC BY-NC-SA 2.0.