Regarding methodology
What does the "methodology" of a survey consist of?
The "survey methodology” refers to a series of practices and principles which set market research apart from merely asking questions. The goal of a good methodology is to obtain all the necessary information and to ensure that the results are reliable, accurate, and useful for the survey creator.
A good methodology involves good planning of questions, a proper selection of respondents, and for the questionnaire to be simple and comprehensible. If you need help in any one of these areas, you will find information here:
Selecting respondents,
Preparing effective questions
Is the Internet a good medium for interviewing people?
Yes, if the people you wish to interview are connected to the Internet.
Most survey creators find that the persons they wish to interview use the Internet. For example, surveys for executives, employees, consumers with high purchasing power, university students, etc.
Are Internet surveys valid?
Just as with traditional paper surveys, the results of a survey which is well designed and executed will be more "valid" than a poorly made survey.
The information contained in this "frequently asked questions" section, and the information furnished in other sections of this Website, will help you to design excellent surveys.
Who should I interview to obtain the best results?
The information provided by a survey is useful for learning about the group of persons answering the survey or about the persons similar to those who have answered. Therefore, the results of a survey made to retired persons will not be very useful for taking decisions regarding working persons.
How do I create efficient questions?
When preparing a question, the most important thing is to know what information you wish to obtain. Many interviewers find it is very useful to write the information they wish to obtain first, before writing out the questions.
Generally speaking, a question must be designed in the simplest and clearest form possible. The idea here is to make the questions as concise as possible and for the whole world to understand it the same way.
For example, a question such as “If the general elections were today, who would you vote for?” is better than “Who would you choose?” because the latter could be interpreted in different ways (who do you like best, for example). Likewise, "Assuming the general elections were today, who would you give your vote to?” is a question which is too long and complicated.
It's important not to use words which could have an influence on the respondent's response. For example, “On a scale from 1 to 10, how would you rate an omelette chewing gum flavour?” is better than “On a scale of 1 to 10, how much would you dislike the flavour of an omelette chewing-gum?”
And, lastly, be consistent: if you ask a question about Renault, you should ask about Citroen in the same way.
Remember:
- Use clear language, which could be understood by anyone
- Avoid asking personal questions
- Be brief
- Follow a logical order
What types of questions should I use?
Within the closed questions category, there are several types, namely:
Dichotomous questions. Questions which only have two possible responses: "Yes" or "No". They allow the opinion of the respondent on a certain subject to be clearly distinguished.
Multiple choice questions. Give the respondent a series of options. Allow respondents to identify preferences, consumption levels, uses and activities. For example: What brand of soft drinks do you consume? 1) Coca-Cola 2) Pepsi 3) Fanta 4) Sprite.
Opinion questions. Questions in which the respondent gives a value judgment and which are used to denote trends. For example: How would you rate the quality of product X? 1) Excellent 2) Very Good 3) Good 4) Sufficient 5) Poor.
Filter questions. The filter questions are those which, depending on the answer given, will allow other unnecessary questions to be skipped in order to make the survey more effective and efficient. For example, if you are carrying out a work environment question amongst your employees, you may be interested in entering a series of specific questions for each department for which we will place the following filter question: Which of the following departments do you belong to? And depending on the respondent's response to this question, the following question will concern the specific department to which he or she belongs.
Open questions (those in which the respondent freely expresses his opinion) are good for obtaining shades of meaning, but they are the most difficult to analyze.
How many response options should I give?
Normally, from three to six. There can be more, but if you give too many options respondents may lose interest. On the other hand, the respondents must find the option which matches their response because if they can't find an answer they might quit the survey.
Studies show that it is a good idea to include the "none of the foregoing" or "no opinion" options in order to obtain a higher response percentage.
What is a filter question? What is it for?
Filter questions are those made previously to another question, in order to eliminate the participants who are not affected by it.
In other words, they stipulate whether subsequent questions are to be offered to the participant or not.
For example: Filter question
Which company department do you belong to?
1- Marketing and sales
2- Financial
3- Production
4- (...)
If the person answers 1 he or she is asked a specific question concerning marketing and sales, if he or she answers 2 the question is asked regarding the financial department...
Representativeness of the sample
The major concern when designing the sample of respondents is to ensure that the sample correctly represents the population segment which is being studied. For example, if you wish to get general information on Internet users and you only ask programmers, you would get information about that segment, not about Internet users in general.
However, if you carry out the surveys to the entire public you wish to interview, as is the case with your customers and employees, and not only to a sample, you don't need to worry.