Question

In conducting a survey, what are the factors that you must consider to reduce measurement errors in the design? Provide an example of both a good and bad survey question to support your response.

Answer 1

Measurement errors in the design of a survey occur when the survey statistic differs from its true value due to imperfections in the way the statistic is collected: such as improper wording of the survey questionnaire.

Examples of a bad Survey question & their correction:

(1) Avoid Leading Questions:

How small was Lal Bahadur Shastri, India's past Prime Minister?

The mistake in the question: "small". This immediately makes the survey respondent to give a small value for the height of Lal Bahadur Shastri, India's past Prime Minister.

The correct form of this question is:

What is the height of Lal Bahadur Shastri, India's past Prime Minister?

(2) Avoid loaded question:

In your opinion, should the Government take action to close health spoiling liquor shops in the country?

The question is loaded with the words, "health spoiling", which will make the respondent biased against liquor shops.

The correct form of this question is:

Should the Government close liquor shops - What is your opinion?

(3) Avoid double question

The questionnaire should orient towards only the objective of the survey,.It should not divert the attention to other areas.

Bad question:

For meeting the increased demand of the company's product, should the company pay more for the existing staff and make them work more or recruit additional staff.?

Here the double question iis: should the company pay more for the existing staff and make them work more or recruit additional staff.?.

The correct question is with the objective of the survey: increased work load or additional staff.

(4) Avoid binomial questions (yes/ no)

The questionnaire should provide for flexibility on the part of the respondent.

Bad question: Do you drink alcohol? (Yes/No)

This binomial question cannot be answered correctly because depending upon the circumstances even a non-drinker may take alcohol just to keep company.

(5) Avoid embarrassing questions

For example, if a surveyor openly asks a student, "Have you cheated in the examination?".

The answer will be naturally "No", since the student will be afraid of after effects.

The correct method to make the survey anonymous.