Question

(PLEASE ANSWER BOTH QUESTIONS)

Think about the statistic you hear on TV... 4 out of 5 dentist prefer Trident gum over other brands. Is this true? Where did they come up with this? What if they only had a sample size of 200 dentists and all of those dentists in their sample were interviewed immediately after leaving a free promotional event sponsored by Trident? Would that seem legitimate to you?

This is a good example of how selecting a good, probability sample is needed to have valid results. You can use a convenient sample as was used in the example above. Give an example of how to choose 200 dentists using a good, probability sample.

Answer 1

The statement would be an example to how sampling can be rigged to manipulate the results. If the sameple size of 200 dentists were interviewed immidiately after leaving a promotional event sponsored by Trident then majority of them are likely to recommend Trident. It is not a legitimate statistics survey since only a special group ro which was Trident was suggested is involved. So the result cannnot be deemed valid since the sample was too convenient.

A legitimate way of samping that saves time as well as is valid would be if there was a large scale event by the dentists association which can be considered a true proportion of the dentists.

If the event wasnt sponsored by and dental brands and participation is 1000 then we can select a systematic sample of every 5th member to get a sample of 200 dentists that is valid. This sampling also saves time and is valid.

Another way would be cluster sampling. If there are 50 cities concerned . We can choose 10 cities randomly and interview 20 dentists randomly from each city. This sampling process would require more time and expense but is valid.

Hope you understand . Happy learning.