Finals Review

Question 1

1. The research design for a quantitative study involves decisions concerning all of
   the following except:
   A) which conceptual framework to use.
   B) whether there will be an intervention.
   C) what types of comparisons will be made.
   D) how many times data will be collected.

Answer 1

Ans: A
Feedback:
The research design is the overall plan for answering research questions. In
quantitative studies, the design incorporates decisions about whether there will be
an intervention, what type of comparisons will be made, and how many times data
will be collected. The selection of a conceptual framework is not a design-related
decision.

Question 2

2. A nurse researcher tested whether sucrose (vs. sterile water) had a beneficial
   effect on infant pain during immunizations. Neither those administering the
   intervention nor the parents of the infants knew which infants received the sucrose.
   This strategy is an example of:
   A) randomization.
   B) attrition.
   C) crossing over.
   D) blinding.

Answer 2

Ans: D
Feedback:
Blinding involves concealing information from participants (here, the parents) and
research staff (those administering the sucrose or sterile water) to enhance
objectivity and minimize the risk of bias. Randomization is a strategy designed to
control confounding variables, as is the use of a crossover design. There is no
indication that these strategies were used in this example. Attrition refers to the
loss of participants and is not a research strategy.

Question 3

3. When researchers withhold information from the people collecting research data
   (e.g., information about who received an intervention and who did not), the
   researchers are taking steps to minimize which type of bias?
   A) Selection bias
   B) Attrition bias
   C) Expectation bias
   D) History bias

Answer 3

Ans: C
Feedback:
Blinding (masking) involves concealing information from participants and staff to
Minimize the risk of biases stemming from people’s expectations about the study outcomes. Research staff’s expectations can result in changes in what they see,
what they hear, and what they say to participants. Selection bias (or selection
threat) refers to preexisting differences between groups being compared. Attrition
bias results from the differential loss of participants from study groups. A history
bias occurs when an external event co-occurs with an intervention and obscures the
intervention’s effect.

Question 4

4. When researchers decide what to use as a counterfactual, they are making a
   decision about:
   A) the type of comparison that will be made.
   B) where the study will take place.
   C) how confounding variables will be controlled.
   D) when and how often the data will be collected.

Answer 4

Ans: A
Feedback:
A counterfactual is what would happen if people are simultaneously exposed and
not exposed to an intervention. When evaluating an intervention, researchers must
select a comparison strategy to represent the “not exposed” situation. Other design
decisions concern where to conduct the study, how confounders will be controlled,
and how often data will be collected, but these are unrelated to designing a suitable
counterfactual.

Question 5

5. Which of the following is a question that researchers should ask when deciding
   how to control confounding variables?
   A) From whom should information about the design be withheld?
   B) When will the research data be collected?
   C) What factors, other than the independent variable, could affect the outcome?
   D) Where and how often will data collection occur?

Answer 5

Ans: C
Feedback:
When developing a plan to control confounding variables, researchers often need to
Answer the question “what other variables could influence the outcomes?” The
other options, which all relate to important design decisions, are not relevant to the
issue of controlling confounding variables.

Question 6

6. Which of the following is an important criterion for making a causal inference
   about the relationship between two variables?
   A) The absence of ambiguity about how outcomes will be measured
   B) The absence of ambiguity about which variable occurred first
   C) The ability to randomly assign study participants to groups
   D) The ability to blind study participants and research staff

Answer 6

Ans: B
Feedback:
A key criterion for establishing causality is ensuring that the cause (the
independent variable) preceded the effect (the outcome). A causal inference does
not depend on blinding or on how outcomes will be measured. Randomization is a
strategy for facilitating causal inference, but it is not a criterion for the inference.

Question 7

7. The causes of health-related phenomena (e.g., obesity) are usually:
   A) probabilistic.
   B) deterministic.
   C) prognostic.
   D) counterfactual.

Answer 7

Ans: A
Feedback:
Causes of health-related phenomena usually are not deterministic but rather are
probabilistic²that is, a cause increases the probability that an effect will occur
rather than determining that it absolutely will occur. Causes are not “prognostic (predictive) or “counterfactual.”

Question 8

8. An important function of research design strategies in a quantitative study is to
   exert control over:
   A) outcome variables.
   B) mediating variables.
   C) carryover variables.
   D) confounding variables.

Answer 8

Ans: D
Feedback:
Confounding variables can obscure the relationship between the independent and
dependent variables; they need to be controlled to enhance causal inferences.
Outcome variables and mediating variables are variables in which researchers are
Interested and do not need to be controlled. There is no such thing as a “carryover variable.”

Question 9

9. A hypothetical condition embodying what would have happened to people if they
   were simultaneously exposed and not exposed to a causal factor is:
   A) a cause.
   B) an effect.
   C) a confounder.
   D) a counterfactual.

Answer 9

Ans: D
Feedback:
In a research context, a counterfactual is what would have happened to the same
people exposed and not exposed simultaneously to a causal factor. A counterfactual
can never be realized, but it is a good model to keep in mind in designing a study
to address cause-probing questions. An effect is the difference between what
actually did happen with the exposure (the cause) and what would have happened
without it. A confounder is a variable or influence that is extraneous to the research
question and that needs to be controlled.

Question 10

10. There are three well-established criteria for inferring a causal relationship,
    attributed to John Stuart Mill. Which is one of those criteria?
    A) A cause and effect must occur simultaneously.
    B) A counterfactual must be known and controlled for in the research design.
    C) The relationship between the cause and the effect cannot be explained as being
    caused by a third factor.
    D) The relationship between the cause and the effect must be deterministic.

Answer 10

Ans: C
Feedback:
One of the criteria for a causal relationship is that the effect cannot be caused by a
third variable, i.e., a confounder. Confounding variables must be controlled and
ruled out. A cause must precede the effect (not occur simultaneously).
Counterfactuals are not controlled, confounders are. Causes are almost always
probabilistic, not deterministic (i.e., they increase the likelihood that the effect will
occur).

Question 11

11. For which broad category of research questions is causality not an issue?
    A) Therapy
    B) Meaning
    C) Prognosis
    D) Etiology/harm

Answer 11

Ans: B
Feedback:
Questions about meaning are addressed using qualitative approaches and are not
focused on causality. Many research questions, however, are about causes and
effects, including Therapy questions, Prognosis questions, and Etiology/harm
questions.

Question 12

12. Which of the following is not a standard feature of a true experimental design?
    A) An intervention
    B) A placebo
    C) Randomization
    D) Control

Answer 12

Ans: B
Feedback:
A true experimental design is characterized by having an intervention, control (a
control condition), and randomization. Placebos are sometimes used in
experimental design, but they are not a standard feature.

Question 13

13. In a true experimental design, what does the researcher manipulate?
    A) The independent variable
    B) The dependent variable
    C) The confounding variable
    D) The mediating variable

Answer 13

Ans: A
Feedback:
Experimenters manipulate the independent variable by administering an
intervention to some participants and withholding it from others. Experimenters
deliberately vary the independent variable (the presumed cause) and observe the
effect on the outcome. They do not manipulate dependent, confounding, or
mediating variables.

Question 14

14. The use of randomization for assigning participants to conditions eliminates:
    A) systematic bias.
    B) ethical problems.
    C) the need for a control group.
    D) the need for manipulation.

Answer 14

Ans: A
Feedback:
Random assignment is done using methods that give every participant an equal
chance of being in any group, such as by flipping a coin. In this fashion, the risk of
systematic bias is eliminated. Randomization is the most reliable method for
equating groups on all characteristics that could affect study outcomes²that is, for
approximating the ideal but impossible counterfactual. Randomization does not
eliminate ethical problems, the need for a control group, or the need to manipulate
the independent variable.

Question 15

15. Which statement is true?
    A) Random assignment is accomplished with random sampling.
    B) Grouping participants with similar features together is the best way to achieve
    random assignment.
    C) The assignment of participants to different conditions at random is a signature of
    a true experiment.
    D) Recruiting participants from different neighborhoods results in random
    assignment.

Answer 15

Ans: C
Feedback:
A true experiment requires randomly assigning participants to different groups or
conditions; randomization is a signature of a true experimental design. Random
sampling is a method of selecting people for a study and is not used to accomplish
randomization to groups. Grouping participants with similar features and recruiting
participants from different neighborhoods are not random methods; they are flawed
methods for achieving comparable groups.

Question 16

16. A pretest in a pretest±posttest design corresponds to:
    A) an outcome measure.
    B) a baseline measure.
    C) a measure of attitudes toward the intervention.
    D) a measure of the independent variable.

Answer 16

Ans: B
Feedback:
In pretest-posttest designs, data on the dependent variable are collected both
before the intervention (a baseline measure) and after it (an outcome measure). In
an experimental study, the independent variable typically is not “measured”; it is
the variable manipulated by the researcher (e.g., receipt vs. nonreceipt of an
intervention). Researchers do not often measure attitudes toward an intervention,
and in any event, they would not be measured at the pretest stage of the project.

Question 17

17. Which statement is true?
    A) A crossover design allows participants to choose the intervention they will
    receive.
    B) A crossover design is useful for eliminating crossover effects.
    C) A crossover design is not a true experimental design.
    D) A crossover design achieves the highest possible equivalence among participants
    exposed to different conditions.

Answer 17

Ans: D
Feedback:
In crossover designs, people are exposed to more than one condition and serve as
their own controls²resulting in the highest possible equivalence among those being
compared. Crossover designs, which are true experimental designs, do not allow
participants to choose interventions they will receive. Crossover designs can be
susceptible to carryover effects²that is, being influenced in the second condition by
what participants experienced in the first.

Question 18

18. A nurse researcher tested two different head-of-bed elevations for the
    prevention of ventilator-associated pneumonia. In this 2-day trial, patients were
    randomly assigned to receive either the 30-degree elevation or the 45-degree
    elevation on the first day, and the other elevation on the second day. What type of
    design was used?
    A) Crossover
    B) Pretest-posttest design
    C) Delay of treatment
    D) Time series

Answer 18

Ans: A
Feedback:
In this crossover design, all patients received both treatment conditions, in a
randomized order. In a delay of treatment design, control group members are
offered the intervention after being put on a wait-list, which did not occur in this
example. Pretest±posttest designs involve multiple groups of participants
randomized to different treatment conditions, which is not the case in this example.
A time-series design is quasi-experimental and does not involve randomization.

Question 19

19. Which statement is true about RCTs?
    A) The control condition in an experiment corresponds to the “I” in Pico-format
    questions.
    B) The control group condition is a proxy for an ideal counterfactual.
    C) Control groups never receive the experimental intervention.
    D) The most common control group condition in nursing research is a “no treatment” control group

Answer 19

Ans: B
Feedback:
In a true experiment or RCT, researchers use the control group as a proxy for an
ideal (but physically impossible) counterfactual. The control condition corresponds
to the “C” in PICO-format questions, not the “I”. Control groups do sometimes get
the experimental intervention, for example, in a wait-list design. A “no treatment”
control group is rare in nursing research because care is seldom withheld
altogether- “usual care” is a frequently used counterfactual.

Question 20

20. An RCT could use any of the following designs, except a:
    A) crossover design.
    B) nonequivalent control group design.
    C) randomized posttest-only design.
    D) randomized pretest-posttest design.

Answer 20

Ans: B
Feedback:
An RCT (the acronym for a randomized controlled trial) always involves
randomization, which is a signature of a true experiment. The only answer option
that does not involve randomization is the nonequivalent control group design,
which is quasi-experimental.

Question 21

21. A researcher randomized nursing home residents to a fall prevention
    Intervention or to “usual care.” After outcome data were collected, control group
    members were offered the chance to receive the intervention. This control group
    strategy is called:
    A) a “usual care” group
    B) an attention control group.
    C) a wait-listed control group.
    D) an alternative treatment control group.

Answer 21

Ans: C
Feedback:
It is true that the control group received “usual care” during the study, but the
most prominent feature of this delayed treatment (wait-list) design is that people in
the control group were able to receive the intervention at a later date. There was
no alternative treatment in this example, and there is no indication that control
group members got special attention during the intervention.

Question 22

22. To test the effect of a Hatha yoga intervention on blood pressure, 100
    prehypertensive adolescents were randomly assigned to either Hatha yoga classes
    or to an art class. This control group strategy is called:
    A) a “usual care” control group
    B) an attention control group.
    C) a wait-listed control group.
    D) an alternative treatment control group.

Answer 22

Ans: B
Feedback:
In this example, the art class was offered to students as a way of giving
adolescents the same amount of special attention as those in the treatment group,
without any of the hypothesized active ingredients of Hatha yoga. The art class was
not “care as usual”, and there is no indication that the control group eventually
received the yoga intervention. The art class is not especially likely to affect the
adolescents’ blood pressure and so is not an alternative treatment.

Question 23

23. Which of the following is a major strength of true experimental designs?
    A) They are less artificial than other designs.
    B) They are undertaken with a strong theoretical underpinning.
    C) They tend to be less costly than other designs.
    D) They permit greater confidence in making causal inferences than other designs.

Answer 23

Ans: D
Feedback:
The great strength of experimental studies lies in the confidence with which cause and-effect relationships can be inferred. Through the controls imposed by
manipulation, control, and randomization, alternative explanations for group
differences on the outcomes can usually be discredited. Limitations of experimental
designs include that they may be more (not less) artificial than other designs, and
they are often quite costly. Not all experimental studies are undertaken with a
guiding theoretical framework.

Question 24

24. Which statement is true?
    A) RCTs are in the middle of most evidence hierarchies for Therapy questions.
    B) RCTs are more susceptible to confounding variables than quasi-experiments.
    C) Many variables of interest to nurse researchers cannot be experimentally
    manipulated.
    D) RCTs are the most often used design for studying the effect of cigarette smoking
    on lung cancer.

Answer 24

Ans:C
Feedback:
A large number of human attributes cannot be randomly assigned to people (e.g.,
blood pressure, age, fatigue), and so experiments are not always feasible.
Nevertheless, RCTs are near the top of evidence hierarchies for Therapy questions,
only one rung down from systematic reviews of RCTs. Quasi-experiments are lower
down on evidence hierarchies for Therapy questions because they are more
susceptible to confounding influences than RCTs. The link between lung cancer and
smoking has never been studied in an RCT because it would be unethical to
randomize people to a smoking or nonsmoking group.

Question 25

25. Which of the following is invariably a feature of quasi-experimental research?
    A) A control group
    B) An intervention
    C) Matching
    D) Randomization

Answer 25

Ans: B
Feedback:
Quasi-experiments (trials without randomization) involve an intervention but lack
randomization and, sometimes, lack a control group. Matching of subjects is a
method of controlling confounding variables by consciously forming comparable
groups; this strategy is used in some, but not all, quasi-experiments.

Question 26

26. Which of the following statements is true?
    A) A signature of quasi-experiments is that they involve randomization.
    B) Quasi-experiments sometimes²but not always²involve an intervention.
    C) A hallmark of quasi-experiments is that they involve a comparison group.
    D) In a nonequivalent control group design, it cannot be assumed that any groups
    being compared are equivalent at the outset.

Answer 26

Ans: D
Feedback:
One weakness of a nonequivalent control group design is that researchers cannot
assume that groups being compared in terms of outcomes at the end of the study
are comparable at the outset²there could potentially be many confounding
variables. Quasi-experiments never involve randomization, and some do not involve
a comparison group, but they do always involve an intervention.

Question 27

27. A researcher initiated an innovative cigarette cessation intervention. Seventy
    participants received the intervention, and their cigarette smoking 3 months later
    was compared to smoking at baseline. What type of design is this?
    A) A crossover design
    B) A nonequivalent control group pretest±posttest design
    C) A one-group pretest±posttest design
    D) A time-series design

Answer 27

Ans: C
Feedback:
The researcher in this example used a very weak one-group pretest±posttest
design. There was no repeated collection of data (so, not a time-series design), and
no comparison group (thus, not a nonequivalent control group design). A crossover
design is an experimental design that involves randomly assigning people to
orderings of alternative treatments, which did not occur in this example.

Question 28

28. A team of researchers evaluated a protocol designed to reduce medication
    errors. Data on medication errors were obtained for each month in the 12 months
    preceding the intervention, and for each month in the 12 months following the
    intervention. What type of design is this?
    A) A crossover design
    B) A nonequivalent control group pretest±posttest design
    C) A one-group pretest±posttest design
    D) A time-series design

Answer 28

Ans: D
Feedback:
A time-series design involves the collection of outcome data over an extended
period, during which time an intervention is introduced, as in this example. In a
one-group pretest±posttest design, data are collected twice, whereas in this
example, data were collected 24 times. There is no comparison group, so this is not
an example of a nonequivalent control group design. A crossover design is an
experimental design that involves randomly assigning people to orderings of
alternative treatments, which did not occur in this example.

Question 29

29. Compared to true experiments, quasi-experiments are more susceptible to:
    A) rival hypotheses about causality.
    B) ethical transgressions.
    C) feasibility problems.
    D) recruitment difficulties.

Answer 29

Ans: A
Feedback:
A major disadvantage of quasi-experiments is that they are susceptible to rival
hypotheses²that is, alternative plausible explanations for what caused the
outcomes, rather than the intervention of interest. Quasi-experimental research is
not more prone to ethical transgressions than true experiments, and in most cases,
quasi-experiments are more feasible (not less feasible) than RCTs. Some people are
reluctant to be randomized, and so it is often easier to recruit participants for a
quasi-experimental study than for an RCT.

Question 30

30. What is a central feature of nonexperimental (observational) studies?
    A) Researchers do not manipulate the independent variable.
    B) Researchers do not manipulate the dependent variable.
    C) Data are collected through the observation of participants’ behaviors.
    D) Their ability to elucidate cause-and-effect relationships is unparalleled.

Answer 30

Ans: A
Feedback:
In nonexperimental (observational) research, the researchers do not intervene²
that is, they do not manipulate the independent variable to test its effect on the
dependent variable. Data for nonexperimental studies may come from the direct
observation of participants’ behaviors, but this is not always true and is not a
central feature. Nonexperimental studies are relatively weak in terms of their ability
to support causal inferences.

Question 31

31. A sample of both users and nonusers of electronic cigarettes was followed over
    a 10-year period to assess whether the e-cigarettes had any long-term adverse
    effects. What type of design was this?
    A) Time-series
    B) Case-control
    C) Prospective
    D) Crossover

Answer 31

Ans: C
Feedback:
Prospective (cohort) designs begin with a presumed cause (here, electronic
cigarettes) and look forward in time for its effects. Retrospective designs, such as
case-control designs, begin with the outcome and look back in time for antecedent
causes. Both time-series and crossover designs involve an intervention, and there
is no intervention in this example.

Question 32

32. Women with and without lymphedema following breast cancer surgery were
    compared in terms of their presurgical weight, as measured by their body mass
    index (BMI). Which design did the researcher use?
    A) Time-series
    B) Case-control (retrospective)
    C) Prospective
    D) Crossover

Answer 32

Ans: B
Feedback:
In this example, the researcher began with the outcome, the presence or absence
of lymphedema, and then they looked back in time to see if being overweight or
obese was a potential risk factor. This is a retrospective case-control design.
Prospective design studies begin with a presumed cause (e.g., weight) and look
forward in time for its effects on outcomes of interest. Both time-series and
crossover designs involve an intervention, and there is no intervention in this
example.

Question 33

33. A researcher collected cross-sectional data to estimate the percentage of
    community-dwelling older people who were at risk of malnutrition. What type of
    study is this?
    A) A descriptive correlational study
    B) A prospective study
    C) A retrospective study
    D) A prevalence study

Answer 33

Ans: D
Feedback:
Prevalence studies document the prevalence rate of a condition or problem at a
single point in time. There is only one variable in this example (risk of
malnutrition), and so it is neither a descriptive correlational study, a retrospective
study, nor a prospective study, all of which involve examining relationships among
two or more variables.

Question 34

34. A nurse researcher plans a study to compare the occurrence of anxiety
    disorders in military personnel deployed overseas with those who were never
    deployed overseas. What research design would the researcher use?
    A) A crossover design
    B) A time-series design
    C) A nonequivalent control group design
    D) A nonexperimental design

Answer 34

Ans: D
Feedback:
The researcher would need to use a nonexperimental design because there is no
intervention in this example. Existing groups (those who did or did not deploy
overseas) are being compared in terms of anxiety disorders. All of the other answer
options are research designs for studies that involve an intervention.

Question 35

35. If a researcher studied the relationship between women’s age and their
    frequency of performing breast self-examination, what type of design would be
    used?
    A) Descriptive correlational
    B) Quasi-experimental
    C) Longitudinal
    D) Experimental

Answer 35

Ans: A
Feedback:
Nonexperimental (observational) research includes descriptive correlational studies
that are aimed at describing relationships among variables without a goal of
elucidating causal connections, as in this example. Experimental and quasi experimental studies involve interventions, and both are aimed at facilitating causal
inferences. In this example, data on age and breast self-examination would be
gathered once, and so this study would not be longitudinal.

Question 36

36. What is a major weakness of correlational studies?
    A) They tend to be expensive.
    B) They tend to be artificial.
    C) They are weak in their ability to support causal inferences.
    D) They are impractical.

Answer 36

Ans: C
Feedback:
A weakness of correlational studies is that they can harbor biases due to self-selection into groups being compared, which in turn hinders causal inferences.
Correlational studies can be costly, but often they are not. They are less artificial
than studies involving interventions and are often practical to implement.

Question 37

37. Researchers collect data at a single point in time in which type of study?
    A) Time-series studies
    B) Cross-sectional studies
    C) Longitudinal studies
    D) Crossover studies

Answer 37

Ans: B
Feedback:
Cross-sectional designs involve data collection at one point in time, whereas
longitudinal designs involve data collection at two or more times over an extended
period. In a time-series design, outcome data are collected over a period of time
before and after introducing an intervention²often from different people. In
crossover designs, people are exposed to more than one experimental condition in
random order, and in such designs, data are collected from the same people
multiple times.

Question 38

38. A nurse researcher is studying the attitudes of high school students toward
    bullying at Metro High School during the past school year. What type of study is
    this?
    A) A time-series study
    B) A cross-sectional study
    C) A longitudinal study
    D) A follow-up study

Answer 38

Ans: B
Feedback:
In this study, data focused on the students’ opinions about bullying would be
collected once, and so this is a cross-sectional study. If the study examined changes in students’ attitudes over time, the study would be longitudinal- data
would have to be collected two or more times. Follow-up studies are also
longitudinal, typically focused on the long-term effects of an intervention or a
health condition. Time-series designs involve an intervention, and this study did not
involve an intervention.

Question 39

39. In a study examining the quality of life of patients who had undergone coronary
    Artery bypass surgery, the research team collected data on patients’ quality of life 2
    months after the surgery and then 2 years later. What type of study is this?
    A) Cross-sectional
    B) Crossover
    C) Follow-up
    D) Retrospective

Answer 39

Ans: C
Feedback:
In a follow-up study, data are gathered at two or more points in time from a well defined group, such as people who have a certain illness or experience. This study
involves two periods of data collection, and so it is not cross-sectional. There is no
intervention in this study, so it could not be a crossover study. Retrospective
studies involve looking for a cause of an outcome by looking back in time for
possible antecedents, which is not the case in this study.

Question 40

41. Which strategy can promote constancy of conditions in implementing an
    intervention?
    A) Using standardized research protocols
    B) Using a prospective rather than retrospective design
    C) Maximizing statistical power
    D) Using a homogeneous sample

Answer 40

Ans: A
Feedback:
The careful development of standard protocols is an important strategy for
enhancing constancy of conditions in implementing an intervention. In carefully
controlled quantitative research, steps are taken to achieve constancy in the study
conditions so that researchers can be confident that outcomes reflect the influence
of the independent variable and not the study context. All intervention research is
prospective and not retrospective. The use of a homogeneous sample is used to
control participant factors, not the study context, and maximizing statistical power
enhances statistical conclusion validity, not constancy of conditions.

Question 41

40. Which of the following is a major challenge specific to longitudinal designs?
    A) Loss of research control
    B) Deterioration of participants’ health
    C) Inferential challenges
    D) Attrition

Answer 41

Ans: D
Feedback:
A major problem in longitudinal designs is attrition²the loss of participants from
The study sample over time. Participants’ health may deteriorate over the course of
a longitudinal study, but it may also improve over time. Research control is no
more likely to be threatened in longitudinal than in cross-sectional studies.
Inferential challenges face all researchers, regardless of design.

Question 42

42. There are several alternative strategies for controlling confounding participant
    characteristics. Which method is the most effective?
    A) Using a homogeneous sample
    B) Statistically controlling confounders
    C) Matching participants on confounding variables
    D) Randomizing participants to conditions

Answer 42

Ans: D
Feedback:
Randomization is the most effective method of controlling participant characteristics
because it controls all possible confounding variables. Other control techniques,
which are less effective, include homogeneity (minimizing variation on the
confounding variables), matching (deliberately making groups comparable on one
or more confounding variable), and statistical control procedures.

Question 43

43. Randomization within a crossover design is a powerful strategy for controlling
    confounding participant characteristics. What problem can potentially arise in such
    a design?
    A) Intervention effects
    B) Carryover effects
    C) Rival hypothesis effects
    D) Causal effects

Answer 43

Ans: B
Feedback:
Crossover designs are often susceptible to the risk of carryover effects, which occur
when the people who are exposed to two different conditions are influenced in the
second condition by their experience in the first condition. Intervention effects are
what researchers hope to find, they are not a problem. The terms “rival hypothesis effects” and “causal effects” are not research terms.

Question 44

44. Researchers do not need to know in advance which confounding variables
    should be controlled for which control strategy?
    A) Matching
    B) Randomization
    C) Statistical control
    D) Homogeneity

Answer 44

Ans: B
Feedback:
Randomization is the most effective method of research control because it controls
all possible confounding variables without researchers having to identify them.
Other control strategies, such as matching, statistical control, and homogeneity,
require researchers to identify in advance the confounders that need to be
controlled.

Question 45

45. Homogeneity is a principle that can be used to control extraneous variation
    from participant characteristics. What does homogeneity entail?
    A) Randomly assigning participants to homogeneous groups
    B) Not allowing important confounders to vary in the research sample
    C) Making the treatment conditions as homogeneous as possible
    D) Creating homogeneous groups of participants through matching

Answer 45

Ans: B
Feedback:
One way to control confounding participant characteristics is to not allow them to
vary²that is, to select participants who are homogeneous with respect to
confounders. Participants are not randomized to homogeneous groups, and
homogeneity of treatment conditions is the opposite of what needs to occur to
make the design powerful. The use of matching is unrelated to the strategy of
minimizing variation in the sample with respect to confounders.

Question 46

6. In a retrospective case-control study of the relationship between (1) having
   Versus not having a herniated disc and (2) patients’ prior history of trauma in a
   sample of 200 men and women, the researcher would most likely use which method
   to control the gender distribution in the two groups?
   A) Having patients serve as their own controls
   B) Randomizing patients to groups
   C) Matching cases and controls on gender
   D) Maintaining homogeneity of the sample

Answer 46

Ans: C
Feedback:
Matching is a common control procedure in case-control studies. In this example,
the researcher would match patients with and without a herniated disc with respect
to gender. Randomization is not used in case-control studies, and participants are
used as their own controls only in crossover designs. Homogeneity could have been
used to control gender by restricting the sample to either men or to women, but
this was not done in this example.

Question 47

47. A researcher used a two-group quasi-experimental (nonequivalent control
    group) design to test the effect of a Tai Chi intervention on the quality of life in
    residents of two different nursing homes, one of which got the intervention. The
    researchers used analysis of covariance to control residents’ age. What strategy for
    controlling participant characteristics did the researcher use?
    A) Statistical control
    B) Matching
    C) Randomization
    D) Homogeneity

Answer 47

Ans: A
Feedback:
Analysis of covariance is a method of statistical control that could be used to remove the effect of age on the residents’ quality of life- which would be desirable
if those getting versus not getting the intervention were of different ages. In this
quasi-experimental study, the groups were not formed through randomization.
Nothing in the description suggests that participants were restricted to a certain
narrow age range (homogeneity) or that they were matched on age.

Question 48

48. Four types of validity affect the rigor of a quantitative study. Which type
    concerns the validity of inferences that an empirical relationship between variables
    exists?
    A) Statistical conclusion validity
    B) Internal validity
    C) Construct validity
    D) External validity

Answer 48

Ans: A
Feedback:
Statistical conclusion validity concerns the validity of inferences that there is a true
relationship (a correlation) between variables²for example, between a presumed
cause and its effect. Internal validity concerns inferences that outcomes were
caused by the independent variable rather than by confounding factors. Construct
validity concerns inferences from the particular exemplars of a study to the higher
order constructs they are intended to represent. External validity concerns
inferences about the generalizability of study results.

Question 49

49. There are four types of validity that affect the rigor of a quantitative study.
    Which type of validity does a researcher specifically address in choosing to use a
    true experimental design?
    A) Statistical conclusion validity
    B) Internal validity
    C) Construct validity
    D) External validity

Answer 49

Ans: B
Feedback:
Internal validity concerns inferences about causality²that outcomes were caused
by the independent variable, rather than by other factors. A true experimental
design offers the strongest evidence about cause-and-effect relationships. Other
forms of validity (construct, external, and statistical conclusion validity) are not
specifically about causal relationships.

Question 50

50. Which of the following can result in a threat to statistical conclusion validity?
    A) Temporal ambiguity
    B) Maturation
    C) Small sample size
    D) Selection bias

Answer 50

Ans: C
Feedback:
Inadequate statistical power is a major threat to statistical conclusion validity, and
small samples tend to yield insufficient power to give research hypotheses a fair
test. Temporal ambiguity, maturation, and selection bias are threats to internal
validity.

Question 51

51. A nurse researcher tested the effectiveness of a telephone intervention to
    reduce anxiety among family caregivers of patients with Alzheimer’s disease.
    During the trial, the researcher performed an assessment and found that staff were
    not following the established intervention procedures. The assessment focused on:
    A) statistical power.
    B) intervention fidelity.
    C) risks for participant attrition.
    D) maturation threats.

Answer 51

Ans: B
Feedback:
Intervention fidelity concerns the extent to which the implementation of an
intervention is faithful to its plan. Intervention infidelity can be a threat to statistical
conclusion validity, as can low statistical power. The researchers’ assessment did
not focus on participant attrition or maturation threats.

Question 52

52. In their cross-sectional correlational study, researchers hypothesized that stress
    was a risk factor for depression among college students. The researchers found a
    correlation between stress and depression, consistent with the hypothesis. Which is
    a possible threat to the internal validity of this study?
    A) Selection
    B) Mortality
    C) Maturation
    D) Temporal ambiguity

Answer 52

Ans: D
Feedback:
In this cross-sectional study, it is impossible to tell if being stressed “caused” higher levels of depression, or if being depressed “caused” higher stress levers- or if a
third factor contributed to (or caused) both. Without knowing which came first
(temporal ambiguity), alternative explanations are possible. Selection is not
relevant in this study, as groups of students are not being compared²there is only
one group. There is no indication that mortality (attrition) was a problem, and
maturation is not relevant in this example.

Question 53

53. Using a one-group pretest-posttest design, nurse researchers studied the effect
    of quiet music on sleep efficiency in low-birth-weight infants. Which threat to
    internal validity would be especially salient in this study?
    A) Selection
    B) Mortality
    C) Maturation
    D) Temporal ambiguity

Answer 53

Ans: C
Feedback:
In this example, the intervention (quiet music) is competing with normal processes
of development in infants (maturation) as the cause of changes in sleep efficiency.
With one-group designs, selection is not a threat. Mortality (attrition) might be a
threat, but loss of participants was not noted and would be less of a threat than
maturation. In intervention studies such as this one, there is no temporal
ambiguity: The intervention was delivered and then sleep efficiency was measured.

Question 54

54. Nurse researchers randomly assigned 50 participants to a 3-month smoking
    cessation intervention and 50 to a no-intervention control group. At the end of the
    study, there were 40 in the intervention group and 32 in the control group. Which
    threat to internal validity would be especially salient in this study?
    A) Selection
    B) Mortality
    C) Maturation
    D) Temporal ambiguity

Answer 54

Ans: B
Feedback:
In this example, attrition from the study (the mortality threat) could lead to an
analysis of a biased subset of participants from the original sample of 100. Attrition
could potentially nullify the initial equivalence of the randomized groups. In
experimental studies using randomized design, none of the other threats (selection,
maturation, temporal ambiguity) is relevant.

Question 55

55. A nurse researcher implemented a physical fitness program for nurses in one
    hospital and used nurses in another hospital as the comparison group. Which threat
    would be the most serious to internal validity in this study?
    A) Selection
    B) Mortality
    C) Maturation
    D) Temporal ambiguity

Answer 55

Ans: A
Feedback:
The selection threat reflects biases stemming from preexisting differences between
groups. Selection bias is the most serious threat to internal validity in studies not
using an experimental design, such as in the nonequivalent control group design
used in this example. There would be no risk of temporal ambiguity²outcomes
would be measured after initiating the intervention. Mortality (differential attrition)
could be a problem, but it is not as serious a threat as selection. Maturation is not
relevant in this example.

Question 56

56. Which threat to internal validity can emerge when external events co-occurring
    with the independent variable affect outcomes?
    A) Maturation
    B) Selection
    C) Mortality
    D) History

Answer 56

Ans: D
Feedback:
The history threat involves the occurrence of events concurrent with the
independent variable that can affect the outcome. The maturation threat arises
when outcomes are affected by the passage of time (e.g., growth, fatigue),
competing with the independent variable as a potential cause. The selection threat
reflects biases stemming from preexisting differences between groups being
compared. The mortality threat involves effects on outcomes that result from
attrition from the study.

Question 57

57. Which of the following is usually the most serious threat to the internal validity
    of RCTs?
    A) Mortality
    B) Selection
    C) Maturation
    D) History

Answer 57

Ans: A
Feedback:
Mortality is the threat that arises from differential attrition in groups being
compared. RCTs avoid the selection threat through random assignment to groups,
but people drop out of the study in a nonrandom manner, which can result in
groups that are no longer equivalent. RCTs are not typically susceptible to history
threats (co-occurring events) or maturation threats (the effects of the passage of
time) because both treatment and control group subjects are equally exposed to
both.

Question 58

58. Using homogeneity as a strategy for controlling confounding variables can
    reduce which type of validity?
    A) Statistical conclusion validity
    B) Internal validity
    C) External validity
    D) Construct validity

Answer 58

Ans: C
Feedback:
External validity concerns inferences about whether relationships found for study
participants might hold true for different people, conditions, and settings²in other
words, generalizability. Using a homogeneous sample is easy, but one problem is
that this strategy limits the generalizability of the results. Homogeneity usually
helps to increase, not decrease, internal validity because confounders are
controlled. Using a homogeneous sample is unlikely to affect statistical conclusion
validity or construct validity.

Question 59

59. Construct validity concerns inferences from the particular exemplars in a study
    to the higher order constructs that they are intended to represent. The absence of
    which strategy in an RCT could undermine construct validity?
    A) Randomization
    B) Statistical control
    C) Matching
    D) Blinding

Answer 59

Ans: D
Feedback:
The absence of blinding could affect construct validity because it may be unclear
Whether outcomes are affected by the intervention itself or by participant’s
expectations about the intervention; the outcomes might not be replicated outside
the context of the research. The other answers refer to strategies to reduce threats
to internal validity, not construct validity.

Question 60

60. Which of the following statements is true?
    A) External validity is reduced in multisite studies.
    B) The goals of internal validity and external validity can sometimes conflict.
    C) External validity is higher in quasi-experimental than in experimental studies.
    D) Internal validity is more important to EBP than external validity.

Answer 60

Ans: B
Feedback:
There is often tension between the demands required to achieve internal validity
(where tight control enhance causal inferences) and external validity (where tight
controls might introduce artificial circumstances that can limit generalizability).
External validity is often enhanced in multisite studies if results are replicated in
diverse settings. For EBP, external validity²the ability to generalize evidence to
new settings-is very important. External validity is not necessarily higher in quasi-experimental studies.