Deck 17: Analysis and Interpretation: Individual Variables Independently

Question 1

What is response error

Answer 1

Response error is a type of error that occurs when a respondent gives response to a particular question but that response is not correct due to some reason. The response error can occur due to factors such as the subconscious character of the respondent which might pressurize himself to give wrong response, misinterpretation of the response by the researcher.For example: If a study is conducted to determine the number of children in a particular state and few households do not respond correctly and provide incorrect numerical value for the total number of children in his/her house, then the study will not be able to give accurate results.

Question 2

Hoffman was the owner of a medium-sized supermarket located in St. Cloud, Minnesota. He was considering altering die layout of the store so that the frozen food section would be near the section with fresh fruit and vegetables. These changes were designed to better accommodate customer shopping patterns and thereby increase customer patronage. Prior to making the alterations, he decided to administer a short questionnaire in the store to a random sample of customers. For a period of two weeks; three of the More cashiers were instructed to stand at the end of selected aisles and con-duct personal interviews with every fifth customer. Hoffman gave specific instructions that on no account were customers to be harassed or offended. Identify the major sources of noncoverage and nonresponse errors. Explain.

Answer 2

Non-coverage errors occur when a part of the population is not considered to be included in the sampling procedure.Nonresponse errors are the errors occur due to unresponsive nature of the respondent even when the respondent is a part of a sample.Non-coverage errors in the given problem are written below:
1. J. Hoffman decided to administer a short questionnaire in the store to a random sample of customers. Choosing a random sample of customers can lead to non-coverage error because it might fails to include those daily customers who come in the store. Random customers might be new to the store and does not meet the criteria to be included in the list of daily customers and thus, should not be included in the sample rather, only those who shop in the store daily have a chance of being included in the study.2. Personal interviews with every fifth customer: It is also a sampling frame problem when personal interview with every fifth customer is used to collect data. There is no list of customers and can lead to non-coverage error.Nonresponse errors in the given problem are written below:
1. Hoffman decided three of the store cashiers and conduct personal interview. Interviewers have a significant impact on the number of refusals they obtain. Hence, cashiers must be well trained and quickly convince potential customers and importance of their participation.2. The script to be used when approaching possible customers must be carefully developed and cashiers must be trained to follow the script rather giving instructions that customer to be harassed or offended. The purpose of the study should be clear.These two errors are two main categories of non-sampling errors and sometimes considered as non-observation errors because either some part of the population was either not included or not responded. Hence, the analysis obtained from the data having these errors is not as much reliable and does not represent a true characteristic of a population.

Question 3

What factors lead to response error How might the respondent, the interviewer, and the situation lead to response error

Answer 3

Response errors are the non-sampling errors that occur when an individual provides an inaccurate response to a survey. The factors that lead to response error are as follows:
1. Respondents misinterpret poorly written items.2. Researcher misinterpretation of an individual's responses.3. Inaccurate and inappropriate instructions to the respondents.4. Lack of accurate responses.5. Characteristics of the respondent that subconsciously influence his or her responses.Respondents, interviewer and the situation eventually lead to response errors in the following ways:
Respondents:
1. Respondents don't understand a question - If respondents don't understand a question, they will either skip the question which leads to potential non-response error or they will answer the question based on their interpretation which may not match the intended interpretation of the item.2. Respondents don't know the answer to the question - Sometimes, respondents do not know the actual answer to the question but many respondents will answer the question anyway which ultimately leads to response error.3. Respondents don't provide a truthful answer - Respondents may consciously lie because they want to make themselves look better or to avoid appearing "dumb" when they don't really know the answer to a question.4. Respondents, sometimes, may be in a bad mood or angry or don't care about their responses or don't want to say something negative about a product, store etc. and they knowingly provide inaccurate answers.Interviewer and the situation:
While asking questions during interview, the problem is acute with open ended question. No two interviewers are likely to use the same probes, creating clear situational differences between respondents. The content, timing of the probes may differ.The two sources of response error that might be caused by interviewer are as follows:
1. Interviewer try to record what the respondent is saying by carefully writing down the person's answers which results in data recording errors.2. Interviewer cheating is the another reason for the response error where the commercial research firms validate through telephone calls or by sending postcards to respondents to verify that they have been contacted.

Question 4

Distinguish between sampling error and nonsampling error. Why is the distinction important

Question 5

How might the different types of response error be reduced

Question 6

Tough-Grip Tires was a large manufacturer of radial tires located in New Orleans, Louisiana, and it was experiencing a problem common to tire manufacturers. The poor performance of the auto industry was having a severe negative impact on the tire industry. To try to maintain sales and competitive positions, various manufacturers were offering wholesalers additional credit and discount opportunities. Tough-Grip's management was particularly concerned about wholesaler reaction to a new discount policy it was considering. The first survey the company conducted to explore this reaction was unsatisfactory. to management. Management felt it was conducted in a haphazard manner and contained numerous nonsampling errors. Tough-Grip's management decided to conduct another study, containing the following, changes:
• The sampling frame was defined as a list of 1,000 of the largest wholesalers that stocked Tough-Grip tires, and the sample elements were to be randomly selected from this list.
• A callback technique was to he used, with the callbacks being made at different times than the original attempted contact.
• The sample size was to be doubled, from 200 to 400 respondents.
• The sample elements that were ineligible or refused to cooperate were to be substituted by the next element from the list.
• An incentive of $1.00 was to be offered to respondents.
Critically evaluate the steps that were being considered to prevent the occurrence of non-sampling errors. Do you think they were appropriate Be specific.

Question 7

What is office error

Question 8

"What are noncoverage errors Explain how they might be a problem with personal inter-view, telephone interviews, mail surveys, and Internet-based surveys.

Question 9

Explain the statement, "Total error is key."

Question 10

The placement office at a university has hired you to assist it in determining the size of starting salaries and the range of salary offers received by graduating seniors. The placement office has always gathered some information in this regard in that historically some seniors have come in to report the name of the company for which they are going to work and the amount of their starting salary. The office feels that these statistics may be biased, and thus it wishes to approach the whole task more systematically. This is why it has hired your expertise to determine what the situation was with respect to last year's graduating seniors.
a. Describe how you would select a sample of respondents to answer the question of starting salaries. Why would you use this particular sample
b. What types of nonsampling errors might you expect to encounter with your approach, and how would you control for them

Question 11

What is the response rate

Question 12

How can noncoverage bias be assessed What can be done to reduce it

Question 13

How should response rates be calculated for (a) mail or e-mail surveys, (b) telephone surveys with no eligibility requirements, and (c) telephone surveys with eligibility requirements

Question 14

An executive recruitment firm used a lengthy mail survey to gather information on the job mobility of mid-level managers. A sample of 500 eligible middle managers was selected, using a simple random sampling procedure. The firm used three waves of mailings. Sixteen of the questionnaires were returned due to bad addresses; all but two of the returned surveys were usable. After the third mailing, each of the nonresponding (NR) sample units was contacted by phone and asked to answer only four questions regarding variables that the recruitment firm thought were particularly important, given the objective of the study. The table below gives mean values for these variables.

a. What was the response rate for the completed questionnaire
b. Which variables, if any, seem to be most affected by potential nonresponse bias Does this tell you anything about the sample selection procedure

Question 15

Why does increasing response rate reduce the possibility of nonresponse error

Question 16

What is nonresponse error

Question 17

The local outlet of a large home and garden chain store wants to better understand the needs and perceptions of its customers. The general manager is attracted to the option of using a Web-based survey (with e-mail recruitment) because it won't cost her much to implement. She prepared a 57-item questionnaire, posted it online using one of the many online survey providers, and then had several employees stay late after work one Friday night to send the following e-mail message to 100 recent customers whose e-mail addresses were in the store's customer database:

The manager was disappointed when she found that only three surveys had been completed by the middle of the following week. Evaluate her approach to data collection and offer suggestions for how to improve the response rate the next time that she chooses to do a Web-based survey.

Question 18

What are the basic types of nonresponse error Are they equally serious for the different methods of administration

Question 19

A large furniture store located in a city in the southwestern United States was interested in determining how households living in and around the city viewed the image of the store. Store managers hired a local market research company to collect data. The researchers drew a systematic sample of 2,500 names from the local telephone directory and set out to conduct telephone interviews. At the conclusion of the data collection phase of the project, the researchers recorded 1,223 completed interviews, 598 not-at-homes, 427 refusals, and 252 nonworking telephone numbers. Calculate the response rate on the project.

Question 20

What can be done to reduce the incidence of not-at-homes in the final sample

Question 21

Suppose that the managers in the previous problem had decided prior to data collection to only include households that planned to make a furniture purchase in the next six months. A screening question was used to identify households that qualified to complete the survey. Further, the 2,500 elements drawn for the sample were accounted for as follows:
473 completed interviews, 612 not-at-homes, 431 refusals, 222 nonworking telephone numbers, and 762 ineligibles.
a. What was the response rate on this project
b. What would the response rate have been without adjusting for ineligible households

Question 22

What are the typical reasons designated respondents refuse to participate in a study What can be done to reduce the incidence of refusals

Question 23

A remodeling contractor located in a small community had developed a relatively inexpensive type of storage shed and wanted to determine the likely level of demand for the sheds among local homeowners before he began producing and marketing them. Because the population of homeowners was not easily identifiable in the telephone directory, and because it would be very inefficient to screen for homeowners due to the fact that it was a college town with the associated high proportion of renters, he didn't want to use a tele-phone survey. Although he could purchase a mailing list of homeowners in the relevant zip code areas, the cost of mail surveys was prohibitive. Because of these constraints, the contractor decided to collect data via an area sample and residential drop-off survey. That is, for randomly selected neighborhoods, local college students working for the contractor would approach each household in the neighborhood, determine whether or not the resident was the homeowner, and if so ask him or her to complete a short written survey that the student would pick up later in the evening. If no one answered the door, the worker was to leave the survey (with instructions) at the door, and to return later to pick it up.
At the conclusion of the project, 438 homes had been approached, with the following outcomes: 212 completed surveys, 31 refusals, 78 residents that rented the residence, and 117 not-at-homes. What was the response rate on this project