Research: Process, Data, & Sampling Flashcards
Key steps in research process
- Identify the problem/issue
- Form hypothesis
- Operationalization - create measurable variables that address the hypothesis.
- Select study design that properly analyzes the data.
Subjective Data
Qualitative data - described verbally, depends on opinions of observer or subject, interviews are an example.
Objective Data
Quantitative data - observable and can be tested and verified. Described in numbers. Examples are surveys, questionnaires, etc.
Key points in data collection
- Collect it close to the time of intervention.
- Frequently collect data
- Keep data collection process short.
- Standardize recording procedures.
- Collection method to fit the study well.
Key considerations in selecting a study design
- Standardization - can the data be collected in an identical way?
- Level of certainty: The study size neede to achieve statistical significance.
- Resources - the availability of funding and other resources needed.
- Time frame required.
- Capacity of subjects to provide informed consent and receive ethics approval from HSRC and IRB.
Three common study designs
- Exploratory
- Descriptive survey design
- Experimental sutdies
Explain exploratory research design
- Common when little is known about a problem/issue.
- Key feature if flexibility.
- Results comprised of detailed observations made.
- Conclusions include educated guesses or hypotheses.
Explain descriptive survey design
- Variable have been already been studies and more research needed.
- Variables are partly controlled by the situation and by investigator.
- Causality cannot be proven, but evidence may support causality.
Ethical concerns in selecting study design
- Research must not lead to harming clients.
- Denying an intervention may amount ot harm.
- Informed consent is essential.
- Confidentiality is required.
What is a single system study approach?
- One client selected per system (n=1)
- Observations are made prior to, during, and following an intervention.
Describe the most basic single system design A-B design
- Baseline phase (A) no intervention.
- Followed by intervention phase (B) with data collection.
Define a case study/predesign DESIGN A
Observational design with no interevention.
Define a case study/predesign DESIGN B
An intervention-only deisgn without any baseline
Define design B-C
“Changes case study” design
No baseline recorded - first intervention (B) applied and then changed (C) and data are recorded.
List the common single system experimental designs
- A-B-A design
- A-B-A-B design
- B-A-B design
Describe the A-B-A design
- A - data collected without any intervention.
- B - intervention applies and data collected.
- A - intervention removed and data collected
What are the benefits of the A-B-A design?
- Inferences regarding causality can be stated
- Two points of comparison are achieved.
What is a drawback of the A-B-A design
It’s usually unethical to remove the intervention, thus it is poorly recommended.
Described the A-B-A-B design
A - data collected prior to intervention.
B- intervention applied, data collected.
A - interevention removed, data collected.
B - intervention reinstated at close of study
Benefits of A-B-A-B design
- Greater causality inferences can be made.
- Temporary removal of intervention is problematic (especially if the subject drops out of the study prior to re-introduction).
- time consuming study
Describe the B-A-B design
- Referred to as the “intervention repeat design”
- no baseline phase
- The intervention is implemented immediately (before establishing a baseline). This is followed by a measurement without the intervention and then a repeat of the intervention.
Population
the total set of subjects sought for measurement by a researcher (example: all women over 18 in the United States)
Sample
A subset of the population
A subject
single unit of the population
Generalizability
The degree to which specific findings obtained can be applied to the total population
Simple random sampling
each subjected selected from a poopulation has n equal chance of being selected.
Stratified random sampling
Populations are divided into groups (i.e., age, income, gender, etc.) and then using a simple random sample from each stratified group
Cluster sampling
Used with natural gruops are readily evident in a population (i.e., residents in a county, students at a high school). These natural groups are then subjected to random sampling. Error and bias are reduced when clusters are internally heterogenous and externally homogenous.
Systematic sampling
Systematic method of random sampling. For example, choosing a number between 1 to 10, and then selecting every nth name of a randomly generated or already existing list to obtain the study sample.