Quantitative Designs in Nursing Research Flashcards
The purpose of research design? And some different types of research(3)
- to aid in the solution of research question; and to maintain control of how data collected
1) basic research- a research question is driven by a researchers curiosity or interest in a theoretical question(ex. Test assumption in nursing theory)
2) applied theory- designed to solve clinical problems rather than to acquire knowledge for knowledges sake(ex. To improve health of particular group of patients)
3) biomedical- lab research with animals; pharmacy
How is a research design chosen?
- usually the state of knowledge in a particular topic dictates the type of research design that is used
- the research problem, purpose, literature review, theoretical framework, and hypotheses should fit together to create a well designed research project
Another purpose of research design? Cont
-the contribution of a study’s find in to evidence based practice is related to the level of evidence provided by the study results
Objectivity of problem conceptualization?
- theoretical framework- used to guide the review or the literature
- review of the literature- provides researcher with knowledge about: what is known about the problem, methods used to study the problem, results, suggestions for future research, if replication of a study is recommended
What is accuracy?
- researcher should choose a design that is consistent with the research problem and one that offers maximum amount of control of variables
- if a problem has not previously been studied using a particular methodology, a pilot study is often conducted to test out whether the research design chosen will actually work with that particular patient population
What is feasibility?
- capability of study to be conducted successfully
- ex) available subjects, timing of research, time required for subjects to participate, costs, and analysis of data
What is control?
- Involves holding the conditions of the study constant
- helps to rule out extraneous or mediating variables that complete with the explanation of the study’s findings
- establishes strict eligibility criteria for the study participants and helps to eliminate bias( on the DV)
How can we establish control?
- use of homogeneous sample
- use of consistent data collection procedures
- manipulation of the IV consistent
- randomization- each member of the population for study has equal access of being selected
Internal and external validity?
- the results of a study must be credible and dependable. So how do we evaluate this?
- internal validity- asks whether the difference or change in DV is a result of the IV under study
- the researcher must be able to rule out other factors or threats that just explain the relationship between the IV and DV
- a number of threads may exist and must be considered by both the researcher, and consumer who is critiquing the study
Threats to internal validity? (6)
1) history- an event that occurs that may affect results or provide another explanation
- ex) researchers studying smoking behaviour of teens
- then federal government launches a big campaign aimed at teens half way through the study
Threats to internal validity?
2) Maturation- refers to a change that occurs over time and is external to the events of the investigation
- includes developmental, biological, and/or psychological processes that occur ex) changes that occur due to death of a spouse, changes in health status, depression
Threats to internal validity?
3) testing- repeated testing with the same instrument may result in a threat to internal validity
- ex. Knowledge test before an info session and then again right after, and then every week for six weeks using same test
Threats to internal validity?
- Instrumentation- calibration of equipment such as thermometer, BP machine, weight scales, glucometer
- can also include techniques/observational data ex. Measurement of waist/hip ratio, size of a bruise or wound
- making changes to a measure that has already been validated
Threats to internal validity
- Mortality- loss of patients in the study after the first measurement
- subjects who dropped out May be different from the group that stayed in study
- also includes loss due to death
Threats to internal validity?
- Selection bias
- need to representative sample for study
- occurs when subjects decide whether they want to be in the study
- ex. Smoking cessation program- only those who are interested in such a program will agree to be in study so this result is selection bias. Major bias problem in clinical research studies
External validity?
- asks: are results generalizable to additional populations and to other sites/conditions? Ex) if you repeated this study in another clinic in another provided would you get the same results?
Threats to external validity(3)?
1) selection bias
- are results generalizable to other populations
- patients at different sites may be very different
- ex. Gender of patients in a study- results may not able to be generalized to all patients if mostly one gender
- muti-site studies help to deal with selection bias
External validity? Threats
- Reactive effects: subjects response to being studied
- ex. Subjects may rate an intervention higher just b/c had more opportunity to interact with a researcher
External validity- threats
- Measurement effects
- administration of a pre-test may sensitize subjects so that answer the post-test differently
- ex- pre-test on nurses attitudes towards HIV patients
- workshop on how nurses are expected to treat HIV patients- measure post-test
- subjects learn from the session what answers are socially acceptable
Experimental designs?
- test cause and effect relationships because they can help to eliminate potential alternative explanations(threats to validity) for the findings
- the validity of conclusions of study depend on how well the researcher have controlled other variables that may further explain the relationships
Types is experimental designs. Three properties**
1) randomization
2) control
3) manipulation
Manipulation of variables? Types
- antecedent variables- factors such as age, stage of disease, gender, health status etc may affect DV
- try to control by patient selection( before study starts)
- intervening variables- occur during course of the study(not part of the study) however, they can effect DV and results of study ex) change in physical, biological or psychological processes that occur as part of life
Types is experimental design(3)
1) true or classic experimental design
2) Solomon four-group experimental design
3) after-only experimental design
1) true or classic experimental design
2) Solomon four-group experimental design
3) after-only experimental design
key to design models- p. 221
R- random assignment O-observation via data collection on DV O1- specifies pretest data collection O2- post test data collection X- intervention
True or classic experiment: steps
- Randomization to control or experimental group
- Pre-test to both groups
- Experimental treatment to one group
- Post test to both groups
- there may be more than one experimental group in some studies when testing more than one intervention
Advantages of experimental design?
- most appropriate design for testing cause and effect
- provides a basis for changing nursing practice
- findings of experimental research judged on the validity that the experimental treatment caused the desired effect on the outcome
Why not used more often? Some reasons? ( experimental design)
- Experimentation assumes that all relevant variables involved are identified ahead of time. Often impossible when dealing with human subjects
- Contamination- subjects talking to each other after randomly assigned. Ex) taking Vit D
- Intervention and assignment disruptive to usual care
- Therefore, turn to quasi- experimental design
Solomon four- group design
Refer to p. 221- this design minimizes the threat to selection bias
Why used or not used? ( Solomon four group design)
- helps to rule out threats to internal validity that before- after group may experience in measures
- need a large patient populations to recruit to ensure have adequate power for sample size
- not used very often b/c of size of sample required
After only experimental design?
- used less frequently
- R to control or intervention groups; only posttest for both groups
- used when pre testing effects could create significant differences between groups
Quasi- experimental designs?
- also used to measure cause and effect relationships, but results not considered to have as high level of evidence as RCT
- used when full experimental control not possible
- randomization is lacking and there may be no control group p.225
- b/c of lack of control, results subject to contamination by many id not all treats to internal validity so hard to determine if intervention actually caused intended impact on outcome DV
Types of quasi experimental designs (4)
- Non-equivalent control group design- two groups where there is a pretest and posttest, but only one group receives the intervention.
- After-only non equivalent control group design- two groups, one group receives intervention, only posttest a used
3, one group(pre test) design - Time series design
Why use quasi experimental design?
- more adaptable to real world situations
- major difference between experimental and quasi experimental design is the R of groups; IV are manipulated in both designs
Evaluation research
Utilization of scientific research methods and procedures to evaluate a program, treatment, practice, or policy
Two types of evaluation research
- Formative evaluation: assessment of a program as it is being implemented
- Summative evaluation- assessment of the outcomes of a program that is conducted after completion of the initial program
Non-experimental designs, when to use them?
Should first ask what is the nature I the problem being studied
- Construct a description of a phenomenon; explore events, people, or situations as they naturally occur
- To test relationships and differences among variables
- To describe a phenomenon at one point or over a period of time
Types of non experimental designs
Survey: descriptive, exploratory, comparative
Relationship or difference studies: correlational, developmental- cross-sectional, longitudinal, retrospective or ex post facto
Survey studies
- collect detailed descriptive of existing variables
- data used to justify and assess current conditions and practices, or to provide direction for improving health care services
- common type of study design in nursing
Descriptive/ exploratory design
Type of non experimental research design that collects descriptions of existing phenomena to justify or assess current conditions, or to make plans for improvement of conditions
- appropriate to use when little is known about the research topic
Comparative
Used to determine if there are differences between variables
Relationship or difference studies- correlational studies
- looking for the strength of the relationship that exists between or among two or more variables
- testing where the variable covary or as one variable changes, does a related change occur in the other variable
Developmental studies
- non- experimental designs that use a time perspective
- interested in how a phenomenon changes over time
Cross- sectional studies
- examine data at into one point in time
- data collected only once from subjects
- subjects may be at different time point in whatever is being measured
- more Than one group be compared( time since diagnosis, time since treatment)
Longitudinal or perspective studies
- collects data from the same group at different time points so data focused and rich
- patients serve as their own controls allowing early trends in data to emerge
- mortality( internal validity) major threat to these studies
- problem: very expensive to do, time intensive, funding agencies usually only fund for 3 years
Retrospective or ex post facto studies
- main characteristics- the DV has already been affected by the IV
- good design if experimental design cannot be used
- aim is to link present events to events that have occurred in the past
Other types of research- methodological research
Development of research took/instruments
Other types of research- meta- analysis
Review of literature using specific critique, framework. Synthesis of findings and statistically summarizes results
Other types of research- secondary analysis
Form of research that looks at previously collected data and reanaulzes with new research question
Other types- epidemiological studies:
Examines factors affecting the health and illness if a population
Other times of research- mixed methods
Uses both quantitate and qualitative approaches
- may use surveys questionaries, interviews
- plus focus groups, patient interviews, journals, diaries
