EBP EXAM 2: Quantitative Research Design Lectures

control

what the researcher does to

maintain uniformity in the study conditions,

avoid/decrease the effects of extraneous, intervening, or mediating variables on the dependent/ outcome variable

avoid bias,

minimize threats the study's internal validity

how is control acomplished

by ruling out intervening, mediating, or extraneous variables that COMPETE with the independent variables (IV) as an explanation for the study’s outcome (DV)

how to control extraneous variables / nosie

Homogeneous sample, consistent data collection procedures since they equal constancy, manipulating the independent variable, randomization.

what does interval validity ask

whether the IV really made the difference/change in the DV

what does internal validity do

rule out other factors or threats as rival explanations or sources of bias

threats to internal validity

history, maturation, testing, instrumentation, mortality, selection bias

what does external validity deal with

the ability to generalize the findings outside the study to other populations and environments.

what does external validity question

under what conditions and with what types of subjects the same results can be expected to occur.

what can external validity be difficult to deal with

because the researcher is assuming that other populations are like the one being tested. when that is probably not the case

• not all trees are the same

threats to external validity

selection effects, Reactive effects, measurement effects

threats to external validity: selection effects

What method was used to select the sample

How were subjects assigned to groups?

threats to external validity: reactive effects

Subjects’ responses to being studied- respond they way they think

they are supposed to, OR t hey respond differently because they are being observed

threats to external validity: measurement effects

A pretest may act like an intervention and influence or change the subjects’ responses

what is internal and external validity important criteria for

evaluating bias, credibility, and dependability of the study results

homogeneity sampling

participants have similar extraneous variables that might affect the dependent variable.

what must homogeneity sampling be based on

the purpose and study’s theoretical basis → inclusion and exclusion criteria

homegeneity sampling is a way of

controlloung variances

when is homogeneity useful

when studying something where there are many factors/variables that could affect the outcome (e.g., weight)

intervention fidelity

intervention is administered in a consistent way to all subjects

• ensures same treatment for everyone

what does intevention fidelity minimize

internal and external threats to validity

randomization

another form of control where each subject from the population has an equal chance of being assigned to an intervention/ treatment group

what does randomization do

Eliminates bias, helps get a representative sample, and is used in a variety of research designs

randomization use

assign groups, or via questionnaires (randomly order questions to see if order makes a difference)

what types of quantitative research designs use randomization.

Experimental designs (RCT)

experimental designs use

used to test weather an independent variable treatment affects patient outcomes

what do experimental designs test

intervention effectives

what is an investigator doing in experimental designs

intervenes by manipulating the independent variable so they can measure change in the dependent variable

what is the primary focus when critiquing experimental design

“ to what extent did the experimental treatment (IV) cause the desired effect on the outcome (DV)?

properties of experimental designs

randomization, control group/usual care, manipulation: does it do something with independent variable

manipulation is the process of

doing something or comparing different types of treatment bu manipulating the IV for some OR all the experimental group

why is the effect of manipulation measured

to determine the result of the experimental treatment on the DV compared to no/partial treatment on the DV.

RCT’s levels of evidence

second

RCT’s design steps

• Sample is recruited from the accessible population

• Baseline measurements are taken pre-intervention (demographics, personal characteristics, etc.)

• Baseline measurement is taken of the DV(s)

• Subjects are randomized to either the intervention or the control group

Each group receives:

a. the experimental intervention (treatment group) OR b. usual care/standard treatment (control/comparison/placebo group)

• Both groups complete post- intervention measures to determine treatment effects ( DV)

• Tools/instruments reliability and validity data are clearly described.

pro/strengths of experimenta; designs

• Most powerful for testing cause & effect – DV was caused by IV (Level II evidence)

• Provides evidence for implementing the research findings into clinical practice (EBP)

cons/ weakness of experimental designs

• Complicated to design

• Expensive

• May be disruptive to clinical setting

• Ethical reasons limit what we can manipulate (e.g., cannot randomly assign people to eating fatty diet or to smoking group… OR to something harmful)

Quasi-experimental design use

used to test weather an independent variable treatment affects patient outcomes

what do quasi-experimental design test

intervention effectiveness-> the investigator actively intervenes by manipulating the independent variable so they can measure change in the dependent variable

what is the primary focus of criquing a quasi experimental design

“ to what extent did the experimental treatment (IV) cause the desired effect on the outcome (DV)?”

what does quasi experimental design as test

cause and affect relationships

why are quasi experimental designs called that

either randomization may not be possible or there may not be a control group

quasi experimenal designs and control correlation

Lack of control makes evidence provided by quasi-experimental designs less convincing

quasi experimenal designs LOE

level 3

what does quasi experimenal designs weakens

the confidence in making causal assertions that the results occurred because of the intervention

quasi- experimental design: Nonequivalent

there is an experimental and control group BUT NO randomization.

what is the most used type of quasi-experimental study in nursing

quasi- experimental design: Nonequivalent

threats to internal validity in quasi- experimental design: Nonequivalent

selection, maturation, testing, and mortality

quasi- experimental design: Nonequivalent results

Results of the study still relatively strong because the researcher can compare the two groups on antecedent variables before the intervention

quasi experimental: After-Only Nonequivalent Control Group Design

-Posttest only control group design

-Randomization is NOT used to assign subjects to groups.

in quasi experimental: After-Only Nonequivalent Control Group Design what does the researcher assume

that the 2 groups are comparable before the introduction of the IV

when are quasi experimental: After-Only Nonequivalent Control Group Design particularly useful

when testing effects are suspected to be a potential major threat to internal validity

QUASI: One Group (Pretest- Posttest Design)

- When only one group is available for study

- Everyone is in experimental group - data are collected before and after treatment is introduced

there is not what in One Group (Pretest- Posttest Design)

no control group or randomization

what must you interpret

findings with careful consideration of the design limitations

Quasi experimental design strengths/pro

• Practical

• Less expensive

• Generalizable

• Sometimes the only feasible alternative because these designs are more adaptable to real-world settings

• Replication of a study can strengthen evidence

Quasi experimental design weakness/ cons

Unable to demonstrate clear cause-and-effect relationships

what do non experimental designs not do

manipulate the independent and do not have a control group 

when are non-experimental designs used

when researchers wish to construct a picture of a phenomenon at one point or over time, test relationships and differences among variables at one point in time or over time, explore events, people, or situations as they naturally occur

are independent variables manipulated in non experimental designs

no

why arent independent variables not manipulated in non experimental designs

because they naturally occurred

subjects and control group in non experimental

subjects are not randomized and there is no control group

what does non experimental explore

relationships or differences between variables

nonexperimental designs are either

surveys or relationship/ differences studies

what LOE is nonexperimental designs

level 4

nonexperimental survey studies may be

descriptive, exploratory, comparative

what do nonexperimental design: survey studies collects

detailed descriptions of existing variables

NE: survey studies use the data to

 Justify and assess current conditions and practices

 To improve health care practices

 Provide basis for program development and interventions

NO surveys can be used

alone, interchangeably or together to describe the survey type.

variables in survey studies may be classifed as

opinions, attitudes, or facts (gender, income level).

what do survey studies provide the basis for

experimental studies

when are studies called comarative

when used to study differences between variables

how is data for survey studies collected

through questionnaires or interviews

survey studies samples may be

 small or large

 broad or narrow

survey studies seek to

RELATE variables to each other NOT CAUSATION

survey studies advantages

• Can obtain a large amount of information from a big population economically

• Sample can be surprisingly accurate of population

survey studies disadvantages

• Information obtained is often broad but superficial (breadth over depth).

• Conducting surveys still requires expertise in several research areas.

• Large-scale surveys can be time- consuming and costly.

NE: correlational studies

Examines the association between two or more variables

correlational studies are interested in

• Whether the variables co-vary?

• Quantifying the strength and the direction of the relationship (+ or -)

types if correlational studies

 Descriptive correlational

 Predictive correlational

NE: correlational study advantages:

 Efficient and effective for collecting a large amount of data

 Potential for:

• evidence-based application in the clinical setting

• foundation for future experimental studies

correlational studies is a framework for

exploring the relationship between 2 or more variables that cannot be manipulated

correlational studies disadvantages

 Variables are not manipulated

 Randomization is not used because groups are pre-existing, so decreases generalizability

 The researcher is unable to determine a casual relationship between variables due to

• Lack of manipulation

• Lack of control

• Lack of randomization

 Strength and quality of evidence is limited by associative nature of the relationship between the variables

NE: developmental design studies subsets

 Cross-sectional studies

 Cohort, longitudinal, prospective (or “repeated measures” studies)

 Case control, retrospective, and ex post facto studies

NE: Developmental Design: Cross- Sectional Studies

a subset of the population

NE: Developmental Design: Cross- Sectional Studies collect what data

only one point in time (often called prevalence studies)

NE: Developmental Design: Cross- Sectional Studies explores what

relationships and correlations, or differences and comparisons, or both

NE: Developmental Design: Cross- Sectional Studies variables

are related to NOT causal outcomes of the study

NE: Developmental Design: Cross- Sectional Studies LOE

lV

NE: Developmental Design: Cross- Sectional Studies advantages

-Less time consuming, less expensive, and more manageable THAN longitudinal/cohort/prospective studies

-Can collect large amounts of data at one point so results available “faster”

-Avoid effects of maturation – a confounding variable

NE: Developmental Design: Cross- Sectional Studies disadvantages

 Ability for an in-depth developmental assessment of the

interrelationships of the variables/phenomenon is reduced

 Researcher is unable to determine whether the change that occurred is related to the change that was predicted because the same subjects were NOT followed over time (in order words, subjects are unable to serve as their own controls)

NE: Developmental Design: Cohort/Prospective/Longitudinal/ Repeated Measures Studies

Cohort studies are collected from the same subjects

overtime (longitudinally)

NE: Developmental Design: Cohort/Prospective/Longitudinal/ Repeated Measures Studies explores

differences and relationships among variables

NE: Developmental Design: Cohort/Prospective/Longitudinal/ Repeated Measures Studies advantages

 Each subject is followed separately, so serves as his/her own control.

 Can obtain increased depth of responses.

 Early data trends can be analyzed.

 Assesses changes in the variables over time

 Both relationships and differences between variables can be explored

 Stronger than case control/retrospective studies→ Because the degree of control that can be imposed on extraneous variables that might confound the data and lead to bias.

NE: Developmental Design: Cohort/Prospective/Longitudinal/ Repeated Measures Studies disadvantages

 Data collection may take a long time, increasing costs in time, effort, and money.

 Testing effects may be a threat from confounding variables.

(internal or external threat?)

 Mortality is a significant threat due to the longitudinal nature

 Increased potential for attrition

 Hawthorne effect

NE: Developmental Design: Case Control/Retrospective/Ex Post Facto Studies

Look backward in time and usually examine exposure to the independent “Case” variable to the DV which has already occurred. (e.g. cancer diagnosis)

what does NE: Developmental Design: Case Control/Retrospective/Ex Post Facto Studies attempt to link

present events to past events

what is case control no

a case study

when would Developmental Design: Case Control/Retrospective/Ex Post Facto Studies be used

when experimental designs cannot be done.

what can NE Developmental Design: Case Control/Retrospective/Ex Post Facto Studies not do

manipulate the IV or do random assignment

NE Developmental Design: Case Control/Retrospective/Ex Post Facto Studies advantages

 SAME as correlational design

extra pro: Offers a higher level of control possible than with correlational studies, therefore; more confident findings.

extra pro: THESE types of studies offer a higher level of control

SOOOO…….one can be more confident in the evidence of the findings.

what is a MAJOR LIMITATIOM of NE Developmental Design: Case Control/Retrospective/Ex Post Facto Studies

Causality cannot be inferred—only relationships/associations.