Deck 11: Bias

Question 1

What does the term "bias" mean in research studies? Choose the best answer.

A) an estimation error that is higher or lower than the true value of a variable
B) a systematic error due to faulty study design, data collection, or analysis
C) an incorrect assumption applied to variables in a study
D) an error in measuring the values of a variable

Answer 1

B

Question 2

The difference between the estimated and actual value of a variable for three cases were as follows: -3, +9, +2. What type of error do these differences suggest?

A) bias
B) measurement error
C) random error
D) variance

Answer 2

C

Question 3

What is a cross-sectional study?

A) a study with data collected at one time point
B) a research study of data for several time periods
C) a study comparing variables with each other
D) a research study of several themes or subtopics

Answer 3

A

Question 4

In a study on changes in weather patterns in Buffalo, NY, temperatures were recorded for each month, using a sample of 10 zip codes. The process of recording temperature per month was repeated every 5 years for a different sample of 10 zip codes. The data covered a total time period of 20 years. What type of research study is this?

A) longitudinal study
B) cross-sectional study
C) repeated cross-sectional study
D) prospective study

Question 5

What does the term "cohort study" refer to?

A) study of longitudinal data of cases with a common starting event
B) study of past characteristics of a set of cases across time
C) study of cases with variable values within a fixed range
D) study of several sets of cases at one time period

Question 6

A study collects data for a set of individuals, on a common set of variables, over multiple periods of time. Which term best describes this type of study?

A) prospective
B) longitudinal
C) fixed-panel
D) cross-sectional

Question 7

Which of the following statements is true of a case-control study?

A) It identifies a control set who have a certain condition or characteristic.
B) It fixes a number of cases who have identical characteristics.
C) It defines a group of cases whose data is collected over time.
D) It sets a control group who do not have a condition of interest.

Question 8

A study identified a sample set of 200 policyholders who had either one or no motor vehicle incidents in the last 2 years. The characteristics of this set of policyholders was analyzed in the study. What type of bias does the study suffer from?

A) sample bias
B) survivor bias
C) nonresponse bias
D) prospective bias

Question 9

A researcher designs a study on political affiliations of those residing in Mercer county, New Jersey. A sample set of households were chosen who resided mainly in and around the town of Princeton. Which type of bias is present in this study design?

A) sample selection bias
B) case-control bias
C) survivor bias
D) retrospective bias

Question 10

What is a "sampling frame"?

A) the limits set for a unit to be included in a study sample
B) list of elements in a population from which sample is to be drawn
C) the general population of the domain being studied
D) the range of values of a sample for the outcome variable

Question 11

A random sample of 2,000 units is chosen from a data set of 49,000 elements. What is the probability of selection to the sample?

A) 24.50
B) 0.408
C) 0.041
D) 2.450

Question 12

A study on political affiliations of residents of Mercer county, New Jersey, created random samples of 100 residents from each census tract within the county. The samples of 100 were then combined to create a master sample for the study. What is this type of sample called?

A) simple random sampling
B) oversampling
C) cluster sampling
D) stratified random sampling

Question 13

Which primary factor distinguishes cluster sampling from stratified sampling?

A) sampling unit
B) sample size
C) population size
D) sampling method

Question 14

What does the term "matching" mean in case-control studies?

A) pairing a case with a control based on similarity of one variable
B) sampling as many controls as the number of cases of interest
C) taking a preset number of cases with presence of a risk factor
D) selecting a fixed set of control units determined by the analyst

Question 15

Which are the nonprobability methods of sampling?

A) case-control and availability
B) convenience, quota, and case-control
C) availability, stratified, and quota
D) cluster, convenience, and quota

Question 16

A research study on the quality of customer service of telecom call centers, chose 20 call centers for the study, with the condition that at least four had to be from Pakistan, two had to be selected from Ireland, and, another four were to be from the Philippines. What type of sampling method was used by this study?

A) stratified
B) cluster
C) quota
D) convenience

Question 17

What is "selective observation" in designing a research study?

A) selecting only cases that support the study's hypothesis
B) choosing the first x%, (x=fixed number), cases in a data set
C) using cases that are already available to the researcher
D) including a fixed number of cases based on a characteristic

Question 18

What does the term "differential nonresponse" refer to?

A) difference in response rate over two time periods
B) response rate differing between sample strata
C) groups having unequal response counts in a sample
D) cases with certain characteristics less likely to respond than others

Question 19

A survey questionnaire was sent to 1,000 individuals. 456 of those individuals filled out and returned the questionnaire. What is the nonresponse rate of this survey?

A) 45.6%
B) 54.4%
C) 45.0%
D) 55.1%

Question 20

What are the consequences of attrition to a longitudinal study?

A) decreasing sample size
B) loss of data on a variable
C) increment in sample size
D) oversampling of a group of cases

Question 21

Cable R has observed the number of customers, residing in Bergen county, fall from 25,315 in June, 2019 to 23,643 in June, 2020. What was the attrition rate for Bergen county?

A) 93.4%
B) 6.6%
C) 1.07%
D) 98.93%

Question 22

How would you calculate the item nonresponse rate of a survey?

A) cases with missing values divided by total cases
B) total number of respondents divided by total values missing
C) respondents with missing values on a variable divided by total respondents
D) cases in the survey divided by cases with variable value missing

Question 23

Which of the following is a "double negative" question? Choose the best answer.

A) Don't you agree that the defective rate should not be above 5%?
B) Don't you think the current defective rate is high?
C) Do you or do you not concur on the efficacy of quality control standards of the company?
D) Do you think the defective rate should not be above 5%, nor below 1%?

Question 24

Peter was unable to respond to a survey due to a health condition. John, his colleague, filled-out the survey in his place, in the best way that he could. What is this phenomenon called?

A) biased response
B) proxy reporting
C) reporting error
D) absentee response

Question 25

What type of bias is present when some values of a variable are systematically wrong in the same direction?

A) social desirability bias
B) nonresponse bias
C) question response bias
D) sample selection bias

Question 26

Social desirability bias could lead to a higher rate of item nonresponse.

Question 27

A leading question is one that uses normative wording to minimize bias.

Question 28

A loaded question should be replaced by a filter question and an associated contingent question.

Question 29

The magnitude of recall bias generally decreases with increase in the recall period.

Question 30

A measurement bias could occur regardless of the type of research study.

Question 31

Bias due to attrition occurs only in cross-sectional research studies.

Question 32

A survey of high school students on their smoking habits was conducted by a district school board. A sampling bias is likely to occur in the data of this study.

Question 33

Surveillance is the continuous and systematic collection of data from a defined sample of a population.

Question 34

Data collected by the Department of Motor Vehicles (DMV) is an example of data collected for administrative purposes

Question 35

Usage of 'big data' for research may lead to greater measurement error or bias.

Question 36

What is the difference between bias and random error? Explain with an example.

Question 37

Describe the three types of bias that could occur in research studies, with the help of an example each.

Question 38

A private hospital conducted a study to predict average number of patients, by department, for the next 12 months. The predictions were based on the hospital's patient data for the last 24 months. What type of study would you call this? Explain your answer.

Question 39

What is a repeated cross-sectional study? Why is it considered a type of longitudinal study?

Question 40

With the help of an example, explain the difference between a prospective and a retrospective study.

Question 41

In your own words, describe the characteristics of a case-control study. Illustrate with an example.

Question 42

What is survivor bias in the context of retrospective studies? Give an example of this type of bias.

Question 43

What are the primary methods for probability sampling? Which type of bias could be associated with these methods?

Question 44

What kind of adjustment is made in disproportionate sampling to restore the pattern in the population to the sample? Illustrate with an example.

Question 45

Differentiate between bias due to study nonresponse and item nonresponse. Provide an example of each.

Question 46

Design a cross-sectional study on a topic of your choice. Make sure that there are at least three independent variables in the study, with at least one of them being categorical. Discuss the sampling method you would use to prepare data for the study. Explain the reason(s) for your choice.

Question 47

For the research study in question 1, discuss the possible types of bias that could be present in the study. What steps would you take to correct this bias?

Question 48

Design a repeated cross-sectional study on a topic of your choice. Ensure that there are at least three independent variables in the study, with at least one of them being categorical. Explain the steps needed to design this type of study. Discuss possible bias due to study and item nonresponse. What steps would you take to correct this bias?