E. Project: Polling Public Opinion (20 points)
For your project, you must create, administer and tabulate
the results of a public opinion poll by working through each of the
If you feel the need for some background from the experts, visit
or any nationally-ranked polling company of your choice.
Please don't try to short-cut the steps below or you'll have problems trying to write the analysis at the end. It won't take you long to work through the steps leading up to your poll but it IS important that you take the time to really think about the decisions below.
What is your topic?
The Texas Lyceum the following are currently the top ten most important issues to Texas voters. Choose
one of those issues to be the topic of your opinion poll.
Middle East unrest
As you probably noticed, each of the topics is very broad. Now that you have chosen one of the topics, decide what specifically you want to ask about that topic. In other words, narrow your focus so your questions
and results will be specific enough to be of use. What exactly do you want to find out about your chosen topic?
What questions are you asking?
minimum of 20 questions for your survey.
Decide the kind of questions you will use. Are you going to use
closed-ended questions (yes or no, for example), open-ended questions (what do you think about...) or a mix of the two? Open-ended questions usually provide more specific information but it's really difficult to compare the responses since there's no limit to how people can respond. Closed-ended questions limit responses to those responses you provide -- "should schools __ ... yes or no." Respondents can only give a "yes" response or a "no" response. That limits the responses but it also makes it easier for you to compare them ... "yes-5, no-15."
Your questions may all be about the topic or some of your questions may ask for information not related to your topic but that you believe will be interesting when compared to opinions on the topic -- for example, respondents' sex, age, ethnic group or etc.
Demographic questions allow you to compare the responses on topic questions. For example, whether men
and women respond differently to questions about unemployment or whether young
and old respond differently to questions about education. The majority of your questions must be topic questions. Whether or not you include a few demographic questions is up to you. When writing your questions, it's important to think about what kind of data you want. Are there answers you want to compare -- do men and women answer differently on question __. If you use open-ended questions, how are you going to compare those responses?
wording of your questions. Are any of your questions leading rather than neutral? Do any of your questions use emotion-laded or controversial words or phrases that might push respondents to one answer or the other? Continue rewording your questions until they are all as neutral, value-free and nonbiased as you can make them. For some help with this, look at
Six Rules for Writing Survey Questions. (The link is under Optional Links above.)
How will you administer your survey?
one method for administering your survey -- verbally by phone, verbally in person (you verbally ask respondents for their answers), written in person (you give respondents the written questions
and ask them to mark their answers) or another method. (Please do not administer your survey online or through email since such surveys are -- by definition -- not random sample surveys and so are not representative of your population.)
not administer your survey in a place where you well known such as at work or church. Since the people at such places know you, you can never be confident that their responses are honest.
Any place that provides anonymity and has a high traffic volume works well -- malls, shopping centers, stores (with prior permission from the manager), schools
and etc. If choosing a single-purpose site, such as a school, ask yourself if the likely population you’ll be polling will be appropriate to your topic and questions.
and, to a lesser extent, verbal in-person surveys should have a script you can follow so you won't inadvertently express bias to your respondents. Written surveys require no script but the survey given to respondents must be neat, simple to understand and legible. (Always check your spelling
and grammar in advance!)
Once you've decided on the method you will use to administer the survey, decide how you will choose which people to survey. You are responsible for administering the survey to a
minimum of 20 respondents.
If calling by phone, will you call 1 person whose last name starts with a different letter of the alphabet - one A name, one B name, etc? If administering the survey in person, will you approach the first 20 people you see
and continue until 20 have agreed to take the survey?
Most students find that administering a written survey in a public place is the easiest method ... even if it's a little scary at first. If you choose this method, keep a few things in mind.
Make certain your survey uses correct spelling
and grammar, is neat
and makes sense. (Test it on a friend first.)
Take extra copies of your survey
and several pens or pencils. Borrow a couple of clipboards if you can.
Not everyone will participate so don't take it personally.
Most people will participate if you tell them it's a class assignment.
Always say "thank you."
Write an analysis of your results that
addresses all of the following.
Once you have completed your survey, analyze the data ... look for patterns, surprises, similarities
and etc in the responses.
Be careful to address all of the topics below.
Your analysis should be thorough, specific, include relevant concepts from
the course material and be free of spelling and grammar errors.
the specific questions in your survey and the responses to each
THIS MUST BE AT THE
BEGINNING OF YOUR ANALYSIS.
Put the specific 20 questions
and responses in a table or a list separate from the discussion below so it is easy to see what questions you asked
and the responses you received. Responses should be raw numbers, not percents, e.g. yes 15, no 5 ...
not yes 75%, no 25%. In other words, I want to see how each of your respondents answered.
your decisions and rationales in the steps above
the kinds of questions you asked
and why, how you administered your survey
and why, etc
any problems you had administering the survey including unanticipated
problems coming from the decisions you made
an objective analysis of your data and what those data tell you about
public opinion and your topic
You are welcome to use percents in the discussion
and there's no need to discuss every question. Spend your time discussing those responses you found most interesting. However, your analysis of the data should comprise the
majority of your report. A report that consists of long discussions covering B
and C above with a few sentences of analysis tacked on at the end, won't earn many points.
specific and detailed connections to course content
This is your last chance before I assign course grades to show me how much you've learned. Make the most of it!