Discussion Question 2: Bias in survey design - The Literary Digest “As goes Main

Question

Discussion Question 2: Bias in survey design - The Literary Digest
“As goes Maine, so goes Vermont”
In 1936, The Literary Digest published the results of a straw poll predicting the victory of presidential candidate Alf Landon. In the November election Landon carried only two states, Maine and Vermont. Franklin D. Roosevelt carried all the other states. Even though the organizers polled 10 million Americans in an attempt to get a representative sample, they made some fatal errors in their selection process. Read about this historic poll and comment on one of the faulty aspects of the design of the survey.

Give another example of a survey (real or invented) that is poorly designed because it contains one or more type of bias. State clearly what kinds of bias the survey has. Then someone else should reply with an indication of how the survey could be corrected. Give your post a descriptive title. You may contribute to this Discussion Question either by giving an example of a poorly designed survey.
• voluntary response sample
• convenience sample
• nonresponse bias
• bias in wording of questions
• confounding
• undercoverage
• response bias

Use one or more of the terms above when describing the faults with the survey you describe. Please make sure that all questions throughout the discussion question are answered in the response. There should be two parts to the answer.

Explanation / Answer

The faulty aspects of the poll were selection bais and non-response bias. The selection process of names on the mailing list comprised names from telephone directory, magazine subscribers, and club membership lists. But at the year 1936, the population of telephone holders, magazine subscriber and club members belog to a high income zone. Thus, the sampling process didnot include a large section of unemployed, low income group people (sans any membership or subscribership) which was a significant proportion of voters. Thus, the sampling process sufferred from selection-response bias. It was further reported that only 2.4 million responded out of 10 million people in the survey, which indicates clear non-response bias.

In order to find the opinion of freshmen regarding the food served on campus, food service at canteen has put volunteers outside the school cafeteria and the lunch time and asked people few questions. This is supposedly a convininece sample, as one would miss people who uses the cafeteria only for dinner, and one would also miss the freshmen who hates the food and never bothered to return back to cafeteria.

Related Questions

Navigate

• Browse (All)
• Browse D
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.