Creating Student Surveys

If you are considering having your students create surveys, here are a few terms with which to be familiar.

• Internal validity. This requires questions testing the hypothesis be asked two or three ways to make sure the person taking the survey is consistent in his/her beliefs. For example, a survey on attitudes about homosexuality asked these two questions to check for internal validity: "Should anyone be able to marry anyone they want?" and "Is it okay for someone to marry someone of the same gender?" The same answer to these questions would show that the person had a well thought out position on this aspect of homosexuality.

• Foils. These questions would be set apart by foils -- other questions related to the topic but not testing the hypothesis. An example of a foils would be: "If two lovers are responsible and caring, yet do not have the opportunity to have children, should they be allowed to adopt a child?" The two lovers may or may not be of the same gender and therefore the answer does not help prove the hypothesis.

• Types of questions. In an open-ended question people are asked to write a response to a question. "What would you do if your brother/sister came home and said he/she was gay?" is an example of an open-ended question. Students soon discovered many people were not willing to write answers to this type of question. They also saw there were going to be problems tabulating results. The students thus wanted to know about closed-ended questions, which require people to circle a provided answer. Although tabulation of results was relatively easy, students discovered they wanted more information. Many wanted to know why students answered the way they did.

• Participant observation. Participant observation seems easier than doing surveys: All a person has to do is watch what is going on around them. Again, students realized that this too had its problems. How does someone look objectively at a situation he/she is in on a daily basis? Jane insisted that every time a teacher called on a boy for an answer, the teacher was discriminating against girls. Jane disliked this teacher and yet she couldn't recognize how this might bias her perceptions.

• Random sampling. Each survey was given to 15% of the students in the middle school. Students drew names from a hat with the names of all the students. Students were not allowed to exchanged names. Some students, because of their hypothesis, had to select only females or needed an equal number of males and females. In these cases, names would be pulled and "extra" male or female names would be put back in the hat until the desired mix was reached.

• Anonymity. All surveys were conducted anonymously. Students were given an envelope in which the survey-takers put their surveys, and warned not to open the envelopes until all the surveys were collected.

Summer 1999