Question

What potential sources of error might be associated with the following situations?

b. A telephone survey of big city voters finds that most respondents do not like negative political ads—that is, advertising by one political candidate that criticizes or exposes secrets about the opponent’s “dirty laundry.” Researchers conclude that negative advertising should not be used.

c. A survey accessed through Instagram produces results ranking Apple Macbooks as far superior to other PCs for business applications. A retailer decides to reduce inventory of PCs other than Macbooks.

d. Researchers who must conduct a 45-minute personal interview offer $175 to each respondent because they believe that people who will sell their opinions are more typical than someone who will talk to a stranger for free for 45 minutes. Management uses the results to adjust their services offering.

e. A company’s sales representatives are asked what percentage of the time they spend traveling, talking on the telephone, participating in meetings, working on the computer, in training, and filling out reports for management. The survey is conducted via the company’s email network. Management concludes that sales reps are not spending enough time performing selling in the field.

Answer 1

Surveys or Interviews are data gathering tools wherein a small group of respondents are used for statistical collections and the inference from the data gathering is implied for the target population. In the above four examples for surveys conducted aprticular finding is drawn about a certaing target population from a small group. Errors are hindrances that the whole data colection method may not be coorect. Potential sources of error are:-

(b) It may be so that the group of people from whom telephonic survey had been conducted were real doubters of advertising through haughty politics wherein people are influenced by digging dirt of each other's political activites. Thus this SAMPLING ERROR (incorrect gathering of survey group) of selecting a survey grouop which is completely different from the target population and the survey conducted does not give the true picture of the population.

Also there may be the error of BIAS wherein the selected survey group suffers from a particular aversion to negative political gimmicks and thus the survey would indicate bias towards political strategy for negative publicity of one another.

This survey might also suffer from NON RESPONSE error since telephonic surveys are not much entertained by people answering the phone call who might immediately disconnect the phone on finding that a survey is being conducted. Thus the survey sample may be reduced to very few responses.

(c) The survey pulled from Instagram might be of only such tech savvy youngsters who own a Macbook or want to own a Macbook thus might be biased towards Apple brand for laptops and thus would not give a true picture of the target population as many people cannot afford Macbook and thus would never buy one- SAMPLING ERROR

Also just picking the survey group from instagram may result in sampling error since it may so happen that instagram at the time of survey had only few youngsters who were attracted towards Macbook due to the apple brand-BIAS

Since the survey is extracted only rom instagram hence it may happen that proper sample is not gathered and thus COVERAGE ERROR might crop up making the survey inference erroneous. There may also be measurement error i.s the survey question might not be appropriate to extract proper opinion from the survey group. Also the method of evalutaion of the survey response may be erroneous-MEASUREMENT ERROR.

(d) Prime source of errors in Paid Interview method of data collection is SAMPLING ERROR i.e the group of people called for an interview might not have any knowledge of the matter for interview and they are just there for the money. Also coverage of the group of people in the interview might not be sufficient to get a decent trend for the target population.

Then paid interviews might suffer from MEASUREMENT ERROR wherein the questions framed might not give a better picture of the dat collection at hand. Also the interview responses might not be quantifiable enough from getting a quantitative response after the completion of the process.

(e) Email survey may not be an effective data collection technique since the employees responding to emails may be BIASED to respond since a negative opinion may curtail their benefits. Thus SAMPLING ERROR happens. Also such conceptual survey on email might lead to MEASUREMENT ERROR since quantifying the repsonse to make a decision may be erroneous.