You decide to take a telephone survey of 40 families in the 721-exchange area. You want an excellent represen?
Answer:
The problem in telephone surveys is that the fact whether a random chosen telephone number will answer the phone is correlated to the answer that she/he will give.
Simple illustration.
Q. Are you man or women? A poll would have to give in 50-50 split, but on a working day in a residential area you will find more women than men.
Therefore you have to add some additional questions which have relevance to the question you are interested in, like gender and age group, profession. Afterwards you use these to correct your survey either by weighing the number of answers, or by deselecting potential answerers if you have already the quotum for that group. Other Questions and Answers: