hey everybody welcome back it seems like
people these days are super skeptical of
opinion polls especially if they don't
like the results is this mistrust
Justified smash that like button and
let's find
out remember this unit is called
American political ideologies and
beliefs these lessons are all about how
political scientists try to measure
people's beliefs the primary method is
through scientifically valid polls so
let's start by talking about polling
methodology and elements of scientific
polls first the sample must be random in
other words everybody in the population
must have an equal chance of being
selected think of like a random number
generator or something like that this
means that you can allow people to
select themselves to participate as
that's not random that's a good start
but the sample also needs to be
stratified meaning that the population
is divided into subgroups and weighted
based on the demographics of the
population okay hoay words but what does
that mean let's say you want to measure
a pinion of students at your school if
seniors are 22% of the school population
then they should be 22% of your sample
your poll shouldn't over or under sample
certain groups otherwise this will
negatively impact the validity of the
poll we're thinking here of things like
race gender party affiliation education
level income age location they should
each be the same proportion of the
sample that they are of the population
now even if you've done everything
properly your poll will still contain a
sampling error that's because you only
pulled a sample of the population you
didn't ask everybody so some of your
results may be the result of chance
variation meaning that if you did
everything the exact same way again the
results might be slightly different a
sample size of a thousand people is
considered scientifically valid and a
sampling error of plus or minus 3% is
generally acceptable but this means that
the results of the poll could be 3
percentage points higher or lower than
what's indicated any scientific poll
will report a sampling error this is
something that can contributes to its
validity additionally keep in mind that
the wording of the question must be
neutral clear and unbiased as any effort
to push respondents towards a specific
response makes the poll invalid and if
people don't understand the question the
results aren't valid either Additionally
the question needs to be framed
neutrally which isn't just the wording
of the question it can be about the
surrounding questions on the survey for
example let's say that a pollster wants
to find out if people support drilling
for oil in a wildlife refuge
even if that question is neutral
pollsters could manipulate respondents
for example if the first 10 questions
were about the environment and animals
they've primed the respondent to answer
in opposition to drilling on the other
hand if the first 10 questions were
about high gas prices and economic
problems they've primed respondents to
answer in support of drilling the type
and format of questions and answer
choices matters for example if the
question is open-ended or multiple
choice or r Choice Etc all these things
will affect the results lastly the
results must be accurately reported
should be clear to understand and should
only draw conclusions that can be
supported by the data this is typically
a bigger problem with media sources than
with pollsters themselves it's very easy
to report poll results and make stronger
assertions about public opinion than the
data support okay so let's switch gears
and discuss four types of scientific
polls each of these can be placed in the
broad category of mass survey
we're talking about interviewing or
pulling a large sample of the population
typically contacting them via phone or
sometimes internet these days first up
opinion polls which are delightfully
self-defining since they are used to
measure you guessed it opinion on some
issue let's say we wanted to know how
people feel about legalizing marijuana
this is the poll that we would conduct
by the way remember that people's
opinions change over time so a poll
showing that people opposed marijuana
legalization way back when dinosaurs
roam the Earth in 1970 isn't relevant to
policy discussions today a tracking poll
is a continuous poll used to chart
changes in opinion over time it asks the
same question every time such as about
presidential approval or perhaps who you
plan to vote for and it may even contact
the same people and then it just tracks
the changes in the response to that
question over time when you hear that
the president's approval rating is up or
down four points since last month this
is the kind of poll being discussed
benchmark poll are typically conducted
by a candidate before they've officially
announced their candidacy this is one of
the first things that they might do to
find out where they stand with the
public before any campaigning they can
gather information such as the strengths
and weaknesses of a candidate find out
if people have even heard of them what
they associate with that person which
demographic groups have more favorable
attitudes towards them that kind of
stuff whereas the first three are most
likely contacting people via the phone
or Internet exit polls are done in
person interviewing people as they exit
the polling place on Election Day
they're trying to gain insight into
voting behavior to help campaigns and
media organizations predict the outcome
of the election before the votes have
been counted they're trying to figure
out which demographic groups showed up
and voted and what the key factors
affecting voter Choice were in the
election and these insights helped them
project the result before all the votes
have been counted in contrast to these
types of polls or mass surveys campaign
sometimes use focus groups to gather a
small number of Voters and Lead an
in-depth discussion about a candidate to
determine how people feel about her it's
like a brainstorming session on steroids
where participants openly discuss their
thoughts and feelings about a candidate
or an issue super detailed insights
absolutely scientifically valid not so
much topic 4.6 is called evaluation of
public opinion data and to be honest
this section is less about learning
content and more about you feeling
comfortable analyzing polling data in
charts graphs and infographic form check
out the ultimate review packet for great
practice questions the other main idea
of this section is asking can we trust
polls are they reliable polling methods
have changed significantly since 2016
pollsters are far less reliant on live
phone polling since many households only
have cell phones and either block or
simply don't answer calls from pollsters
online opin probability based paneling
and even Tech texting have grown
significantly and most pollsters now use
more than one contact method lastly
let's talk about reliability and
veracity reliability ensures we get
consistent repeatable results if the
poll was properly done we should get
very similar results if we immediately
redid the poll all other things equal
and veracity refers to the accuracy of
the data it's not enough for the results
to be consistent we want them to be
accurate for example do they accurately
predict election outcomes or not all
right well it's a wrap for this one
until next time this has been a money
production thanks again for watching I
appreciate you so much you the real MVP
if you can help me out that like button
or tell friends about my channel I
appreciate it check out the ultimate
review packet and I will see you in the
next video
Chrome Extension