feature image via shutterstock
Every day it’s a new very! Important! Top story! Related to Trump, Hillary and the election — because it doesn’t appear that Trump has threatened anyone’s life or suggested that anyone commit international espionage yet today, the story is that Clinton is ahead in the polls — depending on who you ask, she has a “clear and steady lead”, holds a “commanding lead in Virginia” and “trounces Trump among minority voters;” Trump is described as “badly lagging,” most importantly in high-priority swing states Florida and Virginia.
This is all heartening news for those of us who fear Trump being elected above all else despite any other valid complicated feelings we have about this election. However, you may be wondering (as I was!) how invested you can get in this news — in an election cycle that sees seemingly dozens of contradictory takes per day, each claiming to be equally factual, it’s hard to know what headlines you can get your hopes up about, especially when so much else about this election seems to defy expectations and predictions. So let’s look at what these polling numbers mean and how seriously you can take them.
You may have read this piece from the NYT in June that talks about the problems current polling faces, and why we’re seeing an increase in polls that fail to predict results accurately, like the Republican takeover in the 2014 midterm elections. Essentially, the piece argues, the way that polls have traditionally been run — by pollsters calling random respondents on their landline phones — has been derailed by the ubiquity of cell phones and a decrease in people’s willingness to answer poll questions. This has driven pollsters to explore options on the Internet for survey data, but that presents new problems: namely, the demographics that are primarily reached by online polls don’t match up well with the demographic breakdown of the voting population. On the internet, younger voters are overrepresented and older voters are underrepresented, and it’s very difficult to draw a representative sample. The short version is that it’s increasingly challenging to create polls that have accurate results, which is why we’re seeing less accurate results in some cases.
Does that mean that the polls around the probable Presidential candidate are unreliable? Well, not necessarily. Bloomberg has looked into it:
Polls by themselves can vary widely in quality and reliability, but aggregations of polls including the ones on RealClearPolitics or FiveThirtyEight (which weights results by polls’ previous accuracy and also factors in national polls and endorsements), can help smooth out methodological concerns and small sample sizes and allow trends to emerge… Of the 524 individual poll predictions collected by RealClearPolitics and HuffPost Pollster conducted within one month of a state primary or caucus, 450 of them (86 percent) correctly forecast the eventual winner. When we strip out the two biggest misses for polling this cycle, the Iowa Republican caucuses and the Michigan Democratic primary, where 33 out of 38 poll predictions missed the mark, this increases the overall accuracy rate to 92 percent.
What this means is that while any individual poll can be wrong, and sometimes wrong by a big margin, those big misses are often due to methodological issues with that specific poll (the linked article goes into what happened with the Michigan poll that it discusses above) and not that polling in general is flawed. Flaws in individual polls certainly exist and can cause predictions to be off, but the more polls you look at, the less those small flaws affect the accuracy. So what that means in this case, comfortingly, is that it’s a good thing that polls across the board, with varying methodologies, are all seeing Clinton in the lead and to a fairly similar degree. FiveThirtyEight says that the latest batch of polls we’re seeing are actually “a Trump-leaning bunch” due to “house effect” — polling institutions that have “systematic tendencies… to favor either the Democratic or Republican candidate.” Given that fact, we can actually feel very confident about the veracity of this lead — and if you’re looking at sources that use a model of aggregating a broad range of poll results, you can feel fairly confident that the overall trend you’re looking at is accurate.
Are the polls skewed, though, you may be wondering, or “unskewed?” These are terms that have been flying around over the past couple weeks, with some people claiming that some surveys had uneven respondents or other methodological issues and need to be “unskewed,” because the leads they represent are overblown. This claim, however, does not seem to hold water. FiveThirtyEight traces it back to 2012, when “The main contention… was that the polls had too many Democratic respondents in their samples. Dean Chambers… regularly wrote that the polls were vastly undercounting independents and should have used a higher proportion of Republicans in their samples. But in the end, the polls underestimated President Obama’s margin.” Breitbart has also been arguing that polls showing Clinton in the lead are “skewed,” but when they tried to develop their own poll that they claimed would have more accurate results, “despite their efforts to develop a more Trump friendly survey, according to Breitbart’s poll, Clinton is leading a four-way contest with 42 percent of the vote to Trump’s 37 percent.”
There’s still, it feels, lifetimes to go before November, and it will almost definitely be a very rocky road. But if you’ve been looking for one small thing to feel sure of, that you definitively understand in this confusing hellscape, some slightly heartening poll results can be that for you today. Of course, as this election has shown so far, anything can change and these numbers could be different tomorrow! Most crucially, this doesn’t mean that any of us should stop talking to people around us about how dangerous Trump is as a candidate, or fighting the xenophobia, racism, misogyny and Islamophobia that he’s promoted and will unfortunately persist long after his campaign. Also, though, Trump has announced he’s “not going to pivot” and “Republican leaders [have] sent a letter to the RNC urging them to cut [off] funding to Trump, who many see as a losing proposition with less than three months to go until Election Day.” So there’s that!