English 202D

One of the best routes to take is to have your questionnaire done using desktop publishing software. It will look like it was typeset by a professional printer. The result may so impress your readers that it could improve your chances of having them respond. Also consider having the questionnaire copied on attractive paper. Undoubtedly, a questionnaire printed on a faded mimeo sheet or on a copying machine that leaves smudges will not invite quick replies, if any. Avoid reducing the size of your questionnaire on a copying machine in an attempt to save space; the small print will be hard to read and also intimidating. Use plenty of white space, which wil be pleasing to the respondents' eyes. Proofread meticulously. Also make sure that your questionnaire contains a title, a place for the respondent's name, if desired, and instructions.

Once you are sure that the questionnaire is right, deliver or mail it to your respondents. Decide which approach is more feasible. Delivering the questionnaire in person may create goodwill and my increase your chances of receiving a completed questionnaire back from these individuals. However, when you think a personal visit might influence the content of a reader's response, mail your questionnaire. In any case, make sure your questionnaire reaches respondents at times convenient for them, not just for you. Do not give questionnaires to respondents when other commitments prevent their answering you. For example, students preparing for final or state board examinations will not postpone their studies to fill out your questionnaire. Some large companies close for two weeks in the summer and give all employees and supervisors a vacation. Sending a questionnaire to anyone just before or during that shutdown would be pointless. Asking for comleted questionnaires around the Christmas holidays is equally ill advised. If you are aware that respondents will be on vacation, busy, or out of town on business, send your questionnaire at another time.

If you mail your questionnaire, send it by first-class mail and include a stamped, self-addressed envelope. Respondents are far more likely to answer when you make it as easy as possible for them. Make certain that the respondents'names and addresses are correct. People will not be inclined to respond if you spell their names incorrectly or if you address someone with a Mr. When that person is a Ms., or vice versa. Some given names (Pat, Leslie, Terry) apply to both men and women. If you are uncertain, omit the courtesy title and write "Dear Pat Hayes" or "Dear Terry Bronti."

When you mail the questionnaire, enclose a cover letter and follow these procedures.

1. Introduce yourself and tell why you are writing.
2. Explain the significance of the topic of our questionnaire.
3. Emphasize how important the respondent's answers are.
4. Tell what benefits the respondent can gain by answering your questionnaire - the most important function of the letter.
5. Discuss the kind of questionnaire you have devised.
6. Ask that the questionnaire be returned by a specified date.
7. Perhaps promise a gift for returning the questionnaire or even send a gift along with the questionnaire. A coupon, a crisp dollar bill, or a donation to the respondent's favorite charity can improve your rate of return.
8. Thank the respondent and promise, if practical to do so, to inform him or her of the outcome.

Tallying by Hand

The use of separate tally sheets for each question offers these advantages.

1. You are better able to tabulate, organize, and summarize responses for each question, since all answers to that question are on one sheet.
2. You can more easily check responses against individual questionnaires, since each response is coded by identification and question number.
3. You have the flexibility of comparing responses to various questions by simply pulling tally sheets for these questions and laying the sheets side by side.
4. You have a record of the number of questionnaires returned by any specific date.
5. You can conveniently keep a running tally and will not have to wait until all questionnaires are in to begin counting.

Coding and tabulating open-ended questions may present problems. You will have to make option columns as you go along, but still impose some type of manageable limits on these comments that most relevantly answer the questions. Those comments may be buried n useless remarks; the highlighter pen will make the most relevant remarks stand out. A different color highlighter pen can be used to identify statements you foresee quoting directly in your report, thus making them easier to find later.

Look for patterns of similar responses so you can count the number of times similar comments are made. An open-ended question completed by all of your respondents might elicit several clusters of closely related responses.
 

Tallying by Computer

For smaller projects, if you have basic keyboarding skills, access to personal computer, and even a modest knowledge of computer applications, you can use valuable software technology to your advantage.

Two common types of software you can use are spreadsheets and database systems. A spreadsheet allows you to arrange data in rows and columns much like the tally sheet used for manual tabulations. Spreadsheets also enable you to make basic arithmetic calculations from data. These calculations might include totals for such questionnaire items as the number of male and female respondents or the number (and percentage) of respondents who answered "yes" or "no" to a dichotomous question. Where averages are important in your research, spreadsheets can help you to calculate them, too. Spreadsheets do not, however, allow for sophisticated statistical analyses. Yet their "number crunching" capabilities will help you to tabulate even large data sets with ease.

Database programs are computerized filing systems in which you can record separate pieces of information about the individuals who respond to your questionnaire. For example, for each respondent a database could record such characteristics as sex, age, education, as well as responses to specific questions. The database can be useful when your questionnaire relies on open-ended questions. In those cases, each respondent's verbatim response can be stored in his or her own database entry. Database software also allows you to sort through entries. For example, if you wish to examine open-ended responses by males to a particular question, the database application could give you a printout of all the answers by male respondents.

Some software publishers offer integrated applications that include word processor, spreadsheet, and database systems. These integrated packages enable you to move information from one type of application to another. In the case of the database example above, suppose that you wanted to reproduce all of the open-ended responses given by men and women as separate appendices in your report. An integrated software package would allow you to copy information from the database to the word processor.

The software described here is relatively easy to use and can save you much time and effort. It comes in a wide range of prices as well. Some applications are even available free as shareware, or public domain software.

To how many people must you send a questionnaire in order to obtain reliable results? The answer depends on your statistics. If you are interested in finding out how technicians in a particular hospital feel about rotating shifts, then your respondents would be all the technicians in that one hospital. Similarly, if you want to determine whether your sorority chapter should have a barbeque or dance at the end of the spring term, you would question every member of your chapter. But your respondents may not always belong to such a limited and easily accessible group. Many times you will have to consider a larger audience whose members could not all be reached and questioned individually. You might want to find out how customers at a certain store feel about a warehouse branch that would sell merchandise at a reduced rate. Certainly you could not question everyone who shops at this store. Or you might be interested in surveying how a certain profession (dairy farmers, dietitians, medical secretaries) would react to a piece of impending legislation. Being unable to question everyone in a particular profession, you would use some form of sampling.
 

Systematic Random Sampling

Citing another example, let's say that you wanted to find out the customers' preferences for shopping hours at your store. You could stand outside the store and ask everyone who happens to walk by. However, you might talk to people who are visitors to your town and who never have or never will shop at the store, or you could select residents of the town who just stopped to avoid the rain and who never have shopped at the store. To obtain a cross section of those who in fact are customers of the store, you must select as your target population those who are on record as having purchased goods from the store. Perhaps their names are on recrod as credit card holders. You could then go through the store's list of credit card holders and by random selection pick out names of individuals to whom you will send your questionnaire.

Another way to get a statistically valid sample is to pick names randomly out of the residential listings of your town's telephone book. This selection technique would be a good one if you wanted to find out how a relatively small town or suburb (under 30,000 people felt about a bond proposal or welfare services. The important point to remember is that the selection process must be systematic and allow each member of the population (in this case, all those individuals listed in the telephone book) an equal chance of being selected.
 

Stratified Random Sampling

Quota Sampling

The Ten Commandments for writing good questions, click Ten.

For a discussion on writing a report based on your questionnaire, click here.