Survey Question Types 

This part of the assignment is designed to get you to think about what types of questions would generate the type of statistical data that you need if you were to conduct your own survey. 

1. If you could conduct a survey, what would the topic be? Try to make this a one or two word answer. Examples might be: bullying, police brutality, sports participation. It can be more than two words, but what I’m trying to get you to do here is to distill your topic down to a concept rather than writing a question here which is what people have historically had a tendency to do.

Now, let’s move on to devising questions you would ask respondents about your topic. I am going to give some examples using different topics every time, but all of your answers must be relevant to the topic you have chosen above.

The first type of question that you will likely want to ask is demographic (actually there is a fair amount of disagreement about whether demographic questions should come first or last, but we are going with first). Examples of demographic topics include gender, race, age, political affiliation, and college classification, among many others. These are important questions to ask because you’ll want to be able to compare two or more groups to see if their responses to your other questions differ from one another. The specific demographic questions you ask depend on which groups you are interested in comparing. It also will depend on who you plan to have take your survey. For example, it doesn’t make much sense to ask if someone is a student athlete or not if you only have student athletes take your survey. It also doesn’t make sense to ask college classification if you plan on asking people who aren’t in college to take it. Examples of demographic questions are: What is your gender? What is your race?

2. In this space, write down two demographic questions that you believe would be relevant to your topic and why they would be relevant. To give you an example of what I mean by why it would be relevant, if I was doing a study about bullying, I might ask, “what is your gender” because I want to know if boys have been bullied more often than girls.

For questions that relate to your topic, I want you to think about what would generate a nominal, ordinal, or interval/ratio response. As you have seen, this is called the level of measurement. You want to make questions that generate the right level of measurement so you’ll know what statistical methods you can use.

1. In the space below, write      down a question about your topic for which the answer would be a simple      yes or no response. A question that only has two options is called      dichotomous, and because it is a word response, it is also nominal. If I      were doing a study about COVID, I might ask for example: Have you ever had      a positive COVID test? 
2. In this space, write down      a question about your topic for which the answer would be a word response but that has more      than two possible responses. Remember, nominal variables are not      ranked in any way. If the possible answers have a ranking, then the      variable is ordinal. Review over the section below about ordinal questions      to make sure you don’t accidentally do one of those here. Remember, for      statistical purposes you’d want to make sure your answers were short and      not long open-ended questions. Imagine you were to ask a question and then      have a list of four or five answers to pick from like a multiple choice      question on a test. For example, if I were studying eating habits of      college students, I might ask: Where did you eat lunch from yesterday:      SAGA, the Trojan Center, an off-campus restaurant, at home, none of the      above. See how those are not ranked in anyway? One option isn’t “more” or      “better” than another.

Ordinal questions in the social sciences mostly follow one of two formats – a number scale or a Likert scale (which is words). We will start with the number scale:

1. In this space, write down      a question about your topic that could be worded like this: “On a scale      from 1 to 10, how do you feel about…” You can change the exact wording.      You can change the scale to go to something more than 10. You can say      something different than “how do you feel about”. The important part is to      generate a number on some type of scale. For example, if I were studying      the effectiveness of the city-wide recycling program, I might ask, “On a      scale from 1 to 10, with 10 being the most likely, how likely are you to recycle      the next time you have an empty plastic bottle?”

The second type of ordinal question is word scale, which usually takes the form of what’s called a “Likert” scale. Likert scales usually have five options. You are probably very familiar with this type of question, though you may not have known what they were called. They are perhaps the most common type of question asked on questionnaires. The answers to these types of questions are something like: strongly agree, agree, neither agree nor disagree, disagree, strongly disagree. The five options can be different though depending on what you ask. Another example is strongly approve to strongly disapprove. One way you can construct a question like this is to say, “Please tell me how much you agree or disagree with the following statement:” Then, you make the statement and offer the answer options. For my example, I am going to pretend I am studying the issue of gun control. I might ask, “How much do you agree with the following statement: Alabama needs stronger gun control laws. strongly agree, agree, neither agree nor disagree, disagree, strongly disagree

1. In this space, write down      a question about your topic that would use a Likert scale as the answer      possibilities. 

Finally, you could also ask questions that generate number responses that are not ranked in any way. If you ask, “what is your age,” as part of collecting demographic information, that is an interval response, but I want you to come up with a question directly related to your topic that would generate a number answer. A prompt that might help you think of something would be to complete this sentence, “How many….” Here are some examples: “In the last seven days, how many days did you exercise?”, “How many hours per week do you spend on schoolwork?”, “How many minutes does it take you to get from home to your first class?” Do be mindful that asking something like “how many times have you been bullied” isn’t a very good question because people aren’t likely to keep count. Make sure the question is something that you yourself could answer. (Actually, that’s true of all survey questions you write.)

1. In this space, write down      a question about your topic that would have an interval response.