Jessica: It goes back to a comment I made earlier: if one doesn't ask the right questions, on the right people/subjects, at the right time .. the results can be wrong, biased or non-generalizable. Today, in many areas, disruptive technology has to be considered. And, that not only affects voting polling, it can affect any statistical modelling exercise
as @zimana mentioned, it's interesting to see the difference between this and previous elections in terms of polling. It was 8 years ago that the Republicans were convinced their polling was right. This time it was the Democrats. It's a caution to us all in terms of cognitive bias.
Another old reliable (ok, maybe 20 years old) was the whole idea of key precincts, where news organizations and pollsters did exit interviews at polling places that were predictable based on their history. If the interviews matched the previous results you could forecast. That didn't work this time.
And how does that perspective image not just businesses, but other institutions. I like that Google, Twitter, FB have agreed to not do a muslim registry database if requested. How does FB and Twitter operate against government interest if they are treated as a media company but the valuable media - data - is not easily regulated.
@atabachneck. Good point about social media. The experts who swear that it's a good predictor are dedicated users themselves. Those who say it isn't accurate often aren't users themselves. So, there is bias on both sides
Very True Pierre, Zuckerberg doesn't want to be labeled a Media company, but I think he no longer has a choice. Users dictate how the goodies of FB will be used. So whether he likes it or not ....he will have to committ resources to make sure that the News within his site is really news. Good luck with that......
Hey, here's a link to the story about the Nate Silver presentation I went to last year: http://www.informationweek.com/big-data/big-data-analytics/nate-silver-predicts-2016-presidential-race-at-salesforce-world-tour/d/d-id/1323205
Louis - I think social media will struggle with the idea of not being media companies - that's a big shift in business model and introduces new ethical concerns - Social media was not designed to be a news community.
Good point Pierre, Twitter, FB and all the rest have diluted Truth and Facts. Never thought I would live to see the day when facts mean nothing. The death of critical thinking ? Well I will always practice it ....so it is not quite dead yet.
Lyndon: I view that as very much a data problem. Are the right questions being asked of the right people and using the least biased method possible. Obviously, they should have been more concerned with the electorial votes than the popular votes. If they didn't account for that, they were in effect simply asking the wrong population to be able to generalize the results.
@Lyndon_Henry Excellent point. Take the state of Virgina for instance, went strong for you know who. Can't seem to face the fact that Coal is no longer a viable resource. Ignoring change doesn't stop it.
I agreed with Jim - I also unfollowed people when they were becoming ongoing with political commentary rather than encouraging insightful discussion. The reaction is indeed scary - we have more access to commentary to a public event that we've ever had.
Actually I've seen reports that the polling results, as it turned out, were basically correct. Clinton DID win the actual public vote by a fair margin (I've seen the figure of over 4 million votes). The problem was electoral votes, which apparently have been somewhat skewed by vote counting problems in several key swing states.
That said ... I am persuaded by the argumnent that pollsters tended to focus overwhelmingly on surveying urban rather than suburban and rural voters.
Interesting. Thinking like that ...voting with the party is how we got this result. In the end, discard reason and just vote what you think your party used to be.... I am still wondering how one can vote for someone who has used them as a tool, a tool they insulted along the way.....mind boggeling.
There are no easy answers when it comes to data privacy, but there are some best practices, some hard rules that can't be violated, some technical considerations that help, and some ways to understand the right and wrong of managing data. This series will guide you and your organization around some of the pitfalls you could face on the path to success.