Looking into existing data

Before collecting data, make sure to thoroughly check the existing data. Even if you think that there would not be any data, there are still two possibilities: 1) you do not know the existence of a report, etc. that has the data you are seeking, or 2) you can make your own data by applying different definitions or interpretations of existing data. Make sure to first consult other students, graduate students, faculty, administrative staff, and library staff.

Making the questionnaire

Now, if you still cannot find the data you need, you should prepare a questionnaire and collect the data by having people fill it out. Especially important points for making a questionnaire are how to write your questions and what response scales to use.

Write questions using wording that can get the kind of answers you want: The answers will change depending on the wording of the questions. For instance, if you ask ‘Do you want to start a business in the future?’, the respondents cannot give an accurate answer if they do not know what you mean by the word “business.” Similarly, it is unclear whether the “future” refers to right after graduation, within 10 years after graduation, or after you have finally retired from your job. In addition, questions like ‘Do you want to start a business?’ and ‘Do you want to have your own business?’ can be interpreted differently by different people. As such, you need to think carefully about what you actually want to know and what wording you should use to convey that accurately to others.

Preparing options that enable the respondents to give you the kind of answer you want: You will obtain different results for the question ‘Do you want to start a business?’ between when you have the respondents answer only using ‘Yes’ or ‘No’ and when having them answer on a 5-point scale from ‘I definitely want to’ to ‘I have no such plans at all’. If the scale has too few points, it is hard to receive any results of interest, but if it has too many points, it will increase the burden on the respondents instead and it will become difficult to understand the differences between the points on the scale. A 5-point scale is better for questions concerning opinions or thoughts. It is also important to ask questions that respondents can answer using objective numerical values as much as possible, such as ‘How many books about entrepreneurship have you read?’ and ‘How many entrepreneur seminars have you participated in?’

Knowing the structure of a questionnaire

The questionnaire should be structured with the following 5 components.

1. Explanation sheet: Start with the survey title, explanation of the survey’s aims and significance, and the formal request for the respondent’s cooperation.
2. Face sheet: Here, you ask about the respondent’s personal details, behaviors, and experiences. Personal details include age, sex, school year, birthplace, club membership, profession, annual income, years of experience, number of children, and the children’s ages. Examples of behaviors and experiences include number of visits to the library, weekly hours of self-study, number of books you read per month, and jogging distance.
3. Survey items: These items are the core of the survey – they are the questions for the data you want to collect. The following are different ways of asking questions, and among these, direct questions, yes/no questions, and rating questions are the easiest question formats to use for the analysis after you have finished collecting the data.
   • Direct questions: How many points did you score in the National Centre Test? How much did you pay for your computer?
   • Yes/No questions: Do you have a significant other? (1. Yes; 2. No)
   • Rating questions: Do you want to get married in the future? (1. Not at all; 2. Not so much; 3. Neutral; 4. Somewhat; 5. Definitely)
   • Multiple-choice questions: Mark all answers that you think are important when deciding whether to get married. (1. Personality; 2. Age; 3. Annual income; 4. Academic history; 5. Health; 6. Appearance; 7. Similar interests; 8. Similar values)
   • Ordering questions: Pick the three answers that you think are most important when deciding whether to get married, and rank them in order of priority. (1. Personality; 2. Age; 3. Annual income; 4. Academic history; 5. Health; 6. Appearance; 7. Similar interests; 8. Similar values)
4. Open-ended questions: These questions are used to ask about matters that are difficult to quantify or place on a scale.
5. Final words: At the end of the questionnaire, express your gratitude to the respondents for answering the survey.

Deciding on survey subjects and method

Once you have decided that you need to collect data, make a survey plan. An especially important step is deciding on the subjects and the survey method.

Choosing subjects that are suitable for the aims of the survey: Think about what group you want to investigate and whether you can get the respondents who represent that group. If you want to know ‘how many students at Nagoya University are thinking about starting a business in the future’, it is not realistic to include all students in the survey. This is why you will need to get a sample from the group. Moreover, your selection criteria will differ depending on the students you want to know about; for instance, students at a particular department, students in the natural sciences, or all students.

To give a concrete example, the male-to-female student ratio at Nagoya University is 7:3. Thus, if you collect data from 20 individuals, the sample should be made up of 13 men and 7 women. However, if you want to know about the students at the School of Letters, such composition would change as the male-to-female student ratio at the School of Letters is 4:6. Moreover, a third of all students at Nagoya University are enrolled in the School of Engineering. Therefore, if there are only 2 engineering students out of 20 students in the sample, it cannot be said that this sample represent the students of Nagoya University.

Choosing a data-gathering method that suits your aims: You also need to consider how to gather the data from the respondents. To be certain of receiving responses from every respondent, it would be good for you to interview them one by one. However, this will take too much time if there will be many respondents in the sample because you have to schedule meeting times with them. In most cases, you can distribute a questionnaire instead and collect it after it has been filled out. Nowadays, in addition to paper questionnaires, you can collect answers online as well.

Asking professors for advice

There are many other important insights and techniques for conducting surveys, other than those written here. Hence, it is good to read books on survey techniques thoroughly before planning surveys. In particular, the making of scales and the selection of survey subjects should be done in accordance with a rigorous process, which you should know.

Moreover, there are many things that you cannot know from just the literature. For example, when checking that the questions can be answered without being misunderstood or there are no ambiguities in the questions, it is important to have someone else actually look at it and provide feedback. It is good to ask professors actively for advice as you are preparing the survey. It is best to get the sufficient advice while you are still at this preparatory stage, so that you will not end up with data that is not what you were looking for.