Here is the link to my poll: https://assets.bwbx.io/documents/users/iqjWHBFdfxIU/rklCDpOEK78Q/v0
The poll that I am doing is
Bloomberg Politics National Poll. What this poll got right was that Clinton was
in the lead for the majority of election and especially the days leading up to
the election. Another thing that they got right was that the majority of people
who said they would vote, voted for Clinton. This was backed up by the fact
that Hillary Clinton won the popular vote by more then 2 million votes. What
this poll missed was declaring the correct candidate. They had Clinton’s
percentage of the likely hood she would win the Presidency at 84%. This was the
single farthest off statistic that they published during the election season.
This
polling could have been better and more accurate if it was not by live phone,
but posted online. It is more likely that someone who is talking to you on the
phone will lie about his or her pick. People get embarrassed, so they will most
likely mislead you. Also they could have targeted one area that’s more
prominent with liberal residents then conservatives or vice versa. If I wanted
to better understand my candidate’s choices I would ask questions involving
their previous choice for President and if the current President is favorable
or unfavorable. This would give me a clear understanding of the voter’s most
likely pre-conceived notions and where their loyalty (currently) lies. Also another
important question would be about the demographics behind my poll. The
demographics of an area would include asking questions involving wealth, number
of children they have, and how many jobs the parents hold. It is important to know as
much information as possible about the people you will survey.
Polling
is so important because it gives the general public a clear view of the
front-runner throughout the election season. It also informs the voters what
the other citizens of the United States are thinking, so that they can then
form their beliefs and opinions around them. One of the limitations of polling is
that it can be misleading or highly inaccurate. Any time a poll is relying on
people’s honesty there will be errors. For example, most of these election polls
are misleading because Trump supporters became to embarrass to tell pollsters
their choice, so they told them they would vote for Clinton instead. Another
limitation of polls is how they are handled. One poll might have Trump winning
by 5% points, but that’s because the pollsters only surveyed people in a
conservative dominant area. The location and number of people surveyed is
critical for the poll to be as statistically factual as possible.
https://assets.bwbx.io/documents/users/iqjWHBFdfxIU/rklCDpOEK78Q/v0
https://static01.nyt.com/images/2016/01/28/world/28trumpbelgium-web2/28trumpbelgium-web2-facebookJumbo.jpg
https://staticseekingalpha3.a.ssl.fastly.net/uploads/2016/11/957061_14786266408835_rId7.jpg
It’s good that you touched on the significance of a wide demographic when conducting polls. I agree that the lack of diversity within most of the polls during the race played a substantial part in why they were off.
ReplyDeleteAlthough, I don’t think that the phone polling was that big of a factor. There most certainly were voters who were ashamed to admit that they were leaning towards, but there wasn’t an overwhelming amount of people who were in that boat.
I also agree that a wide demographic is crucial when conducting a poll, especially when it is done over the phone. I certainly agree with the assumption that some people may have been embarrassed and misled the pollers. In this election in particular misleading could have been more prominent than past elections because of how polarizing the candidates were.
ReplyDeleteI agree that reaching a wide demographic is really important so I think a combination of online and over the phone methods would be necessary. While reaching a wide base is necessary, it wouldn't solve the issue of weighting that many of these polls missed. Although, I really like that you pointed out the issue of social desirability and respondents lying when asked over the phone.
ReplyDelete