Chapter 09: Experimental and Quasi-Experimental Designs

The nurse researcher is planning a study in which subjects will not be randomly assigned to groups. Which type of design would be appropriate for this study? a. Nonequivalent control group design b. Posttest only control group design c. Solomon four-group design d. True experimental design

ANS: A Feedback A A nonequivalent control group design is a quasi-experimental design in which subjects are not randomly assigned to groups. B A posttest only control group design consists of two randomly assigned groups but neither group is pre-tested. C A Solomon four-group design is an experimental design that requires random assignment to groups. D True experimental design requires random assignment of groups.

Which factor should be the primary consideration when determining the validity of experimental study findings? a. The extent to which the researcher controlled intervening variables b. The number of subjects who dropped out of the control and experimental groups c. The credentials and previous experience of the researcher d. The number of persons involved in the data collection process

ANS: A Feedback A Control is the ultimate criterion in experimental research. B Although mortality (dropout rate) is important, it is not the primary consideration in determining the validity of the findings. C Although researcher credentials are important, they do not ensure experimental control. D The number of data collectors is not the primary criterion for determining the validity of conclusions.

Which action would increase the control of an experimental study? a. Establishing strict criteria for subject selection b. Allowing participants to decide whether or not to be in the experimental group c. Ensuring that subjects do not know the purpose of the experimental study d. Increasing the number of assistants involved in data collection

ANS: A Feedback A Strict subject selection criteria help to minimize bias and limit extraneous variables. B Self-selection for the treatment group would be likely to decrease control, increase bias, and limit the validity of study outcomes. C Subjects are entitled to know the purpose of a study via informed consent. D Increasing the number of research assistants to collect data can decrease control if they are not specially trained.

The experimental treatment initiated by a researcher can be described as what kind of variable? a. Independent b. Dependent c. Extraneous d. Outcome

ANS: A Feedback A The experimental treatment is the independent variable. B The dependent variable is the outcome variable. C An extraneous or intervening variable can interfere with establishing causality between the independent and dependent variable. D The outcome variable is the dependent variable.

Which research designs could be classified as experimental? (Select all that apply.) a. Solomon four-group design b. After-only design c. Nonequivalent control group design d. After-only nonequivalent control group design e. One-group pretest-posttest design f. True experimental pretest-posttest control group design g. Time series design

ANS: A, B, F Feedback Correct The Solomon four-group design is an experimental design. The after-only design is an experimental design. A true experimental pretest-posttest control group design is an experimental design. Incorrect A nonequivalent control group design is a quasi-experimental design. An after-only nonequivalent control group design is a quasi-experimental design. A one-group pretest-posttest design is a quasi-experimental design. A time series design is a quasi-experimental design.

A research study in a laboratory setting incorporates four groups (two control and two experimental) in which only one experimental group and one control group are pretested, while all four groups are posttested. Which descriptions of the design are accurate? (Select all that apply.) a. Experimental design b. After-only experimental design c. Real-world experimental design d. Quasi-experimental design e. Solomon four-group design

ANS: A, E Feedback Correct The example describes a type of experimental design, a Solomon four-group design. The example describes a Solomon four-group design. Incorrect An after-only experimental design does not include any pretesting. A real-world experimental design would be conducted in an actual practice setting, not a laboratory. The example describes an experimental study conducted in a laboratory. A quasi-experimental design would be conducted in a real-world setting.

An after-only nonequivalent control group design would be appropriate to study which of these research questions? a. What is the effect of the cause of spousal death on widows' self-esteem? b. What is the effect of preoperative teaching on vascular complications after orthopedic surgery? c. What is the effect of a mentorship program on improvement of students' clinical performance in a baccalaureate nursing program? d. What is the effect of a videotaped teaching method on knowledge of adolescent males about the warning signs of testicular cancer?

ANS: B Feedback A A qualitative approach would be appropriate to study the effect of the cause of spousal death on widows' self-esteem. B After-only nonequivalent control group design is used when outcomes (e.g., vascular complications) cannot be measured before the intervention (e.g., preoperative teaching) and randomization is not used to assign subjects to groups. It makes the assumption that the two groups are equivalent and comparable before the introduction of the intervention. C Because student performance can be measured before and after the intervention, an after-only approach would not be appropriate. D Because knowledge of testicular cancer can be measured before and after the intervention, an after-only approach would not be appropriate.

In which way is the independent variable handled differently in a study with a true experimental design compared with a study with a quasi-experimental design? a. An experimental design requires an independent variable, whereas a quasi-experimental design neither identifies nor requires an independent variable. b. Although both designs identify an independent variable, the nature of the groups (randomized versus nonrandomized) means that manipulation may result in different results. c. Experimental designs demonstrate that the independent variable is related to the outcome, and quasi-experimental studies allow for independent variables that stand alone. d. Although both designs identify an independent variable, experimental designs identify it before the research is initiated, whereas quasi-experimental designs identify it after the data are analyzed.

ANS: B Feedback A A quasi-experimental design includes both dependent and independent variables. B The major difference between experimental and quasi-experimental designs is the randomization of groups. Independent variables are manipulated in both designs. C Both experimental and quasi-experimental designs examine the relationship of dependent and independent variables. D In both experimental and quasi-experimental designs, variables are identified before the study begins.

Which of these criteria is most important when evaluating the results of an experimental research study? a. The size of the experimental group compared with the size of the control group. b. The validity that the experimental treatment caused the desired effect on the outcome. c. The congruence of research questions to the purpose and hypotheses. d. The ability of the researcher to apply the findings to clinical or practice settings.

ANS: B Feedback A Although the groups should not be markedly different in size, findings of an experimental study are not ultimately judged by comparing group sizes. B The validity of experimental study results depends on how well the researcher has controlled alternative explanatory variables. C Although important, congruence of research questions, hypotheses, etc., is not the ultimate criterion for evaluating an experimental effect. D Clinical applicability does not guarantee the validity of experimental results.

Control is important in an experimental research study for which of these reasons? a. It ensures maintenance of participant confidentiality. b. It introduces one or more constants into the experimental situation. c. It prevents data collection from extending beyond the original time specified for the study. d. It ensures a cause-and-effect relationship between independent and dependent variables.

ANS: B Feedback A Participant confidentiality is protected by the researcher's coding of subjects to mask their identity. B Control is essential to minimize the effects of extraneous variables on the interaction of the identified independent and dependent variables. Control of variables makes it possible to explain a study's findings with accuracy. C Control in an experimental design is not related to the time period of the study. D Control does not ensure a cause-and-effect relationship between the treatment variable and the outcome.

In a true experimental study, the control group is the group that does what? a. Receives the experimental treatment b. Receives the placebo treatment c. Experiences the adverse results d. Experiences the reactive effects

ANS: B Feedback A The group that receives the experimental treatment is the experimental group, not the control group. B In true experimental research, the control group is the group that receives the usual or a placebo treatment, rather than the experimental one. C The occurrence of adverse results is not confined to the control group and, in fact, is more likely in the treatment group. D Reactive (Hawthorne) effects are the effects of being studied, which can occur in both experimental and control groups.

A study of a program to help people quit cigarette smoking had three groups of subjects. The first group received one supportive phone call 10 days after attending a program on strategies to resist smoking. A second group received a supportive phone call every week for 6 weeks after attending the program. A third group received no supportive phone calls after attending the program. What is the dependent variable in the research scenario? a. Program on strategies to resist smoking b. Elimination of cigarette smoking c. Weekly supportive phone call d. No supportive phone call

ANS: B Feedback A The program of strategies to resist smoking was the independent variable. B The dependent (outcome) variable is quitting cigarette smoking. C The weekly supportive phone call was part of the treatment (independent) variable. D The control group received no supportive phone call.

The nurse is critiquing a research study that used an experimental design. How can the nurse determine if the study's results are able to infer causality? a. The relationship between independent and dependent variables must be reciprocal. b. The relationship between independent and dependent variables must not be explainable by any other variable. c. Both independent and dependent variables should be characterized by reactive effects. d. Both independent and dependent variables should change as a result of the treatment.

ANS: B Feedback A The relationship between independent and dependent variables is not necessarily reciprocal. B Critical to causality is the control of extraneous variables that may influence changes in the dependent variable. C Reactive effects are the subjects' responses to being studied, which interfere with establishing a cause-and-effect relationship. D In a cause-and-effect relationship, the independent variable changes the dependent variable.

Which are characteristics of quasi-experimental research designs that are considered weaknesses, compared with experimental designs? (Select all that apply.) a. Cost of implementation b. Confidence in making causal assertions c. Subject to contamination of internal validity d. Disruptive to the usual routine of the setting e. Difficult to perform in a particular clinical setting f. Findings due to interference of extraneous variables

ANS: B, C, F Feedback Correct The primary weakness of quasi-experimental designs, compared with experimental designs, is the weakened confidence in making causal assertions. Quasi-experimental designs are particularly subject to contamination of internal validity by extraneous variables. Findings of quasi-experimental studies may be due to the influence of extraneous variables, rather than the experimental treatment (independent variable). Incorrect The costs associated with quasi-experimental designs are generally much lower than the costs of experimental designs. Quasi-experimental designs are much less disruptive to the usual routine of the setting than experimental designs. Quasi-experimental designs are easier to perform in a particular clinical setting than experimental designs.

A nurse researcher is planning a quantitative (experimental) research study to test the effect of a new wound product on wound healing. Which of these factors is necessary to implement an experimental pretest-posttest design? (Select all that apply.) a. Inclusion of a second control group b. Ability to collect pretest data c. Use of two posttreatment data collection times d. Subjects randomized to the control group only e. Random assignment of subjects to treatment and control groups

ANS: B, E Feedback Correct To implement a pretest-posttest design, it must be possible to collect pretest data. Random assignment must be possible for both treatment and control groups. Incorrect There is no need for a second control group for a pretest-posttest design. There is no need for two posttreatment data collection times; only one is needed. Randomization must be possible for both treatment and control groups.

The nurse is critiquing a research study that involves subjects undergoing chemotherapy for breast cancer. Which finding would meet a criterion for inferring causality between performance of exercise and adaptive behavior of the subjects receiving chemotherapy? a. Adaptive behavior began before the exercise program was initiated. b. Adaptive behavior occurred in the control group over time. c. The level of adaptive behavior increased in direct proportion to an increase in the intensity of the exercise. d. No difference in adaptive behavior was observed between subjects in the experimental group and subjects in the control group.

ANS: C Feedback A A change in the dependent variable before treatment began would interfere with a cause-and-effect explanation of change. B A change in adaptive behavior in the control (no treatment) group would interfere with a cause-and-effect explanation of change. C Causality is supported by a clear proportional change in the dependent variable (adaptive behavior) following a clear proportional change in the independent variable (exercise). D A lack of difference between groups would indicate that the independent variable (treatment) was not causally related to change in the dependent variable.

Random assignment of subjects to experimental and control groups has which of these effects? a. Minimizes mortality within groups b. Maximizes variance between groups c. Decreases systematic bias d. Increases selection bias

ANS: C Feedback A Randomization has no effect on mortality (dropout rate). B Randomization minimizes pretreatment variance between groups, rather than maximizing variance. This helps to ensure that posttreatment differences are related to the independent (treatment) variable alone. C Random assignment to groups helps to ensure that other (extraneous) variables that could affect changes in the dependent variable are equally distributed within groups, thus reducing systematic bias. D Randomization reduces selection bias, rather than increasing it.

How are threats to validity affected when studies are conducted in laboratory settings? a. Threats to both internal and external validity are increased. b. Threats to both internal and external validity are decreased. c. Threats to internal validity are decreased, whereas threats to external validity are increased. d. Threats to internal validity are increased, whereas threats to external validity are decreased.

ANS: C Feedback A Studies conducted in laboratory settings have fewer threats to internal validity. B Studies conducted in laboratory settings have more threats to external validity. C Studies conducted in laboratory settings have fewer threats to internal validity (control issues) and more threats to external validity (generalizability to other populations). D Studies conducted in laboratory settings have fewer threats to internal validity and more threats to external validity.

A study of a program to help people quit cigarette smoking had three groups of subjects. The first group received one supportive phone call 10 days after attending a program on strategies to resist smoking. A second group received a supportive phone call every week for 6 weeks after attending the program. A third group received no supportive phone calls after attending the program. In the above scenario, which of these properties of experimental research is evident? a. Quasi-experimental design with no control group b. Random assignment to research groups c. Manipulation of the treatment dose d. Control of extraneous variables

ANS: C Feedback A The group receiving no phone call is the control group. B The scenario does not describe whether random assignment was used. C The scenario describes manipulation of the independent variable by varying the interventions (doses) of phone calls. D The scenario does not describe control of extraneous variables.

A study was conducted to determine if videotaped instruction was more effective than written instruction (a pamphlet) to teach menopausal women the benefits of regular mammograms. Which condition would be considered an antecedent variable? a. There were 38 subjects in the pamphlet group and 34 subjects in the videotape group. b. During the study, 50% of the videotape group received flu shots, whereas 25% of those in the pamphlet group received flu shots. c. Four members of the pamphlet group had mothers who had been diagnosed with breast cancer, but no members of the videotape group had mothers who had been diagnosed with breast cancer. d. While participating in the study, five members of the videotape group began performing breast self-examination (BSE), while two members of the pamphlet group began performing BSE.

ANS: C Feedback A The sample size is a factor occurring during the study, not before it. B Because influenza immunization occurred during the study, it is not an antecedent variable. C An antecedent variable is something that occurred before the study but can affect the outcome and confound results. D Because BSE occurred during the study, it is not an antecedent variable.

A planned quasi-experimental study using a nonequivalent control group design has several extraneous variables that threaten its internal validity. The nurse researcher's actions should be based on which understanding? a. The results of the study will not be statistically significant. b. The random assignment of subjects to groups will nullify the intervening variables. c. The study should be strengthened by controlling intervening variables. d. Because observation rather than causation is the purpose of the study, no action is necessary.

ANS: C Feedback A The statistical significance of the study will not necessarily be affected by the extraneous variables. B By definition, a nonequivalent control group design indicates no random assignment was done. When extraneous variables are identified as threats to the study, they should be controlled using other approaches. For example, if experimental and control groups differ systematically on an important intervening variable (e.g., socioeconomic status), a better match of groups should be attempted. C Identifying and controlling threats to internal validity will strengthen the study's design. D Causation is being examined because this is a quasi-experimental study.

What type of threat to validity is minimized by the use of a Solomon four-group design? a. Mortality b. Selection bias c. Testing d. Antecedent variable

ANS: C Feedback A Use of four groups does not control subject mortality (attrition rate). B The Solomon four-group design will not by itself minimize selection bias. Strict selection criteria and random assignment to groups are necessary to minimize selection bias. C Testing threats to internal validity are minimized by the Solomon four-group design. D Because antecedent variables such as motivation to learn existed prior to the study, they should be identified and measured before the study; differences can be statistically controlled. The Solomon four-group design, however, does not control antecedent variables.

Which of these research designs could be classified as quasi-experimental? (Select all that apply.) a. Solomon four-group design b. After-only design c. Nonequivalent control group design d. After-only nonequivalent control-group design e. One-group pretest-posttest design f. True experimental pretest-posttest group design g. Time series design

ANS: C, D, E, G Feedback Correct A nonequivalent control group design is a quasi-experimental design. After-only nonequivalent control-group design is a quasi-experimental design. A one-group pretest-posttest design is a quasi-experimental design. A times-series design is a quasi-experimental design. Incorrect A Solomon four-group design is an experimental design. An after-only design is an experimental design. A true experimental pretest-posttest design is an experimental design.

A quasi-experimental study design would be a better choice for a research study than an experimental study design in which of these situations? a. The study is to be conducted in a laboratory setting. b. An experimental treatment is not part of the study. c. The researcher has not conducted research before. d. Full experimental control is not possible.

ANS: D Feedback A A laboratory setting would not preclude and, in fact, would facilitate an experimental design because it is conducive to control of variables. B If no experimental treatment is used, the study is not quasi-experimental. C Researcher inexperience should not influence the choice of design. D When full experimental control is not possible, it is appropriate for the researcher to select a quasi-experimental design.

Compared with a quasi-experimental design, which of these characteristics should a nurse expect to see only in a true experimental design? a. Experimental treatment b. Informed consent c. Hypotheses d. Random assignment

ANS: D Feedback A Experimental treatment is present in both quasi-experimental and true experimental designs. B Informed consent is required for all types of research involving human subjects. C Hypotheses are included in both quasi-experimental and true experimental designs. D Random assignment to groups is not present in a quasi-experimental study.

In a study using a true experimental design, which of these methods would control for antecedent variables? a. Manipulation of the independent variable b. After-only data collection c. Intervention fidelity d. Random assignment to groups

ANS: D Feedback A Manipulation of the independent variable cannot control a factor that occurred before the study. B After-only data collection would not control preexisting variables. C Systematically delivering the intervention would not control preexisting variables. D Random assignment to groups helps ensure that antecedent (preexisting) variables will be similarly distributed between groups.

Randomization strengthens an experimental study in which of these ways? a. It reduces the risk that a subject will be harmed by participation in the study. b. It ensures that the experimental group and the control group have the same number of participants. c. It allows the study to proceed without the researcher knowing which group is the experimental treatment group. d. It increases the likelihood that the differences in outcomes between the groups are not related to chance alone.

ANS: D Feedback A Randomization does not protect a research subject from harm. B Sampling techniques such as quota sampling ensure equal group size, not randomization. C Randomization does not blind the researcher to treatment group versus control group. D Randomization helps to ensure that any group differences that emerge in a study are related to the treatment condition and not to chance alone.

A quasi-experimental research design would be preferable to an experimental design in which of these situations? a. A smaller sample size is sufficient to obtain significant results. b. Demonstration of a cause-and-effect relationship is the desired outcome. c. Descriptive statistics will be used to analyze the data. d. Data will need to be collected in a real-world setting.

ANS: D Feedback A Sample size is not a determinant of research design. B An experimental design is stronger than a quasi-experimental design for establishing a cause-and-effect relationship. C Descriptive statistics are used to analyze data in both quasi-experimental and experimental studies, although additional, inferential statistics are used in the analysis of experimental studies. D Quasi-experimental studies are more appropriate in real-world settings, where randomization is difficult or impossible.

How are threats to validity affected when studies are conducted in real-world settings? a. Threats to both internal and external validity are increased. b. Threats to both internal and external validity are decreased. c. Threats to internal validity are decreased, whereas threats to external validity are increased. d. Threats to internal validity are increased, while threats to external validity are decreased.

ANS: D Feedback A Studies conducted in real-world settings have fewer threats to external validity. B Studies conducted in real-world settings have more threats to internal validity. C Studies conducted in real-world settings have more threats to internal validity and fewer threats to external validity. D Studies conducted in real-world settings have fewer threats to external validity (generalizability) but have more threats to internal validity (control issues).

A research study is planned to determine the effect of nutritional supplementation and type of support surface on the rate of wound healing in sacral myocutaneous flap failures. Which is an independent variable in this study? (Select all that apply.) a. Rate of wound healing b. Sacral myocutaneous flap c. Retailer source of protein supplement d. Type of support surface e. Nutritional supplementation

ANS: D, E Feedback Correct The type of support surface is one of the specified interventions or treatments (independent variables). Nutritional supplementation is one of the specified interventions or treatments (independent variables). Incorrect Rate of wound healing is the dependent variable. Sacral myocutaneous flap failure is the type of wound being studied. The retailer of the protein supplement is not an independent variable in this study. It would be an independent variable if two different retailers' supplements were being compared.