Research Flashcards
Non-experimental research
A “No experiement” design, usually includes two broad categories of research:
1.) Descriptive research
2.) Ex post facto/correlational research (in the past)
–Cross sectional
–Cohort
–Longitudinal
Cross sectional study
An observational study done on a population at one point in time and it can help estimate prevelance.
Examines populations with similar attributes that differ in a specific variable (such as age), however the aim is to find relationships between variables at a specific point in time.
Designed to find relationships between variables at a specific point in time or “surveys.”
Cohort study
Compares particular outcomes in groups of individuals who are alike in many respects but differ according to a particular characteristic.
Centered on determining whether a particular outcome will develop over time as a result of a key characteristic.
Can be retrospective or prospective
Longitudinal study
Takes multiple measures of a group overtime to find relationships between variable.
Descriptive research
Aims to describe situations, experiences, and phenomena as these values exist in nature.
Instruments of measure
- Objectivity: The degree to which perceptive variability is eliminated from a unit of measure. No matter who does the measurement, it will be correct.
- Sensitivity: refers to the percentage of true positives correctly identified as a true positive.
- Validity: Measures the degree to which a variable measures what it is intended to measure.
- Reliability: The extend to which a measurement is repeatable under the same conditions.
PICOT
A mneumonic derived from the elements of a clinical research question:
* Patient (population?)
* Intervention (Treatment?)
* Comparison (Control?)
* Outcome (results, effectiveness of what is being measured)
* Time (duration of the measure/data collection?)
Useful framework to answer a clinical based question
Major steps in the research process
1.) Formulating the research problem
2.) Reviewing related literature
3.) Formulating the hypotheses
4.) Selecting the research design
5.) Identifying the population to be studied
6.) Specifying methods of data collection
7.) Designing the study
8.) Conducting the study
9.) Analyzing the data
10.) Interpreting the results
11.) Communicating the findings
Ex post facto/correlational research (in the past)
Examines relationships among variables.
Experimental research
A true research design that includes experimental manipulation of variables utilizing 1.) randomization and 2.) a control group to test the effects of an intervention of experiment.
Randomized control trials (RCTs), these are quantitative designs.
Quasi experimental
An experimental research design that involves manipulation of variables but lacks a comparison group or randomization.
When you’re unable to randomize, or have control group.
Qualitative research
Includes case studies, open ended questions, field studies, participant observation and ethnographic studies where observations and interview techniques are used to explore phenomena through detailed descriptions of people, events, situations, or observed behavior.
“The lived experience.”
Pros/cons:
-Research bias
-Small study participants can call into question the generalizability of findings.
-Produces very rich data through no means of research.
Type 1 and 2 errors in reasearch
Type 1 error: False postitive +, incorrectly rejecting the true null hypothesis.
Type 2 error: False negative -, failing to reject a null hypthesis which is false.
Meta-analysis vs Meta-synthesis
- Meta-analysis: Tests hypotheses by using numerous quantitative studies to systematically assess the results of previous research.
- Meta-synthesis: Analyzes data across qualitative studies in order to build new theories.
Levels of evidence hierarchy
Ranked from top to bottom, pyramid.
1.) Meta-analysis of RCTs
2.) RCTs (quantitatie designs)
3.) Quasi-experimental (lack control or randomization)
4.) Qualitative cohort studies
5.) Case control studies
6.) Editorials/expert opinions
Confidence interval
Measures the degree of certainty or uncertainty in a sampling method.
A small confidence interval implies a very precise range of values.
Standard deviation
Indicated the average amount of deviation of values from the mean.
* 68% of the sample fall within in 1 SD of the mean
* 95% of the sample fall within 2 SD of the mean
* 99% of the sample fall within 3 SD of the mean
Low or small SD indicated data are clustered tightly around the mean
HIgh or large SD indicated data are more spread out.
Level of significance
The probability that the results reported happened by chance.
Example: a p<.05 means that there is a 5% chance that the result is insignificant, or that it happened by chance alone.
Perfect correlation
A measure of the interdependence of two random variables that ranges in value from +1 to -1
-1 indicates a perfect negative correlation
0 indicates an absence of correlation
+1 indicates a perfect positive correlation
t- test:
Statistical test to evaluate the differendes in means between two groups.
Reliability
The consistency of a measurement, or the degree to which an instrument measures the same way over time with the same subjects.
Think practice questions osn exams.
Reflect the estimated repeatability of a measurement
This is estimated in two ways:
–Test/retest: The more conservative method to estimate reliability. One should get the same score on exam 1 as one does on exam 2. Involves two administration of the measurement instrument.
–Internal consistency: Estimates reliability by grouping questions in a questionnaire that measure the same concept. Involves only one administration of the measured instrument.
Cronbach’s alpha
A common way of computing correlation values among the questions on instruments.
As with a correlation coefficient, the closer it is to 1 (optimal >.7), the higher the reliability estimate of the instrument.
Validity
The degree to which a variable measures what it is intended to measure overtime.
