Research Flashcards
Descriptive research
Observe and describe behavior without manipulating or controlling the situation, observations lead to hypotheses that can be tested at a later date
Correlational research
Examines the relationship between two or more behaviors or treats, determine if two behaviors vary in someway, does not show cause and effect
Experimental research
Establish if the independent variable causes the effect on the dependent variable
Quasi experimental research
When a research design does not meet other experimental requirements, most communicative disorders are quasi-experimental
Between subject research design
Performances of separate groups of participants are measured and compared between the groups, different groups are exposed to different treatments or levels of the independent variable e.g. Cross-sectional
Within subject research design
Performance of the same participants is compared and different conditions, everyone exposed to multiple conditions e.g. longitudinal
Mixed research design
Between groups and within subjects, one independent variable compares between subjects and the other independent variable compares within subjects
Single subject design
Participants evaluated as individuals rather than members of a larger group, criterion-referenced, AB design, ABA design, ABAB design
Multiple baseline design
Avoids disadvantages of treatment withdrawal, to prove cause-and-effect baselines must be independent of each other
Single factor studies
One independent variable
Multi factor studies
Many independent variables can be interaction effects
Cross sectional
Groups of people at some moment in time i.e. different age groups
Longitudinal
Measures the same group of people every five years to eliminate different histories
Internal validity
An experiment has internal validity if there is no other reason for the measured change in performance other than the change in the independent variable
External validity
The extent to which the results of the study can be generalized to other people, settings, measures, etc., study may be internally valid but results may not extend outside the parameters of the study
Threats to internal validity
History, maturation, reactive pretest, instrumentation, statistical regression, differential subject selection, mortality
Threats to external validity
Subject selection, interactive pretest, reactive arrangements, multiple treatment interference
History internal threat
Some external event happens that researcher was not aware of and could not have controlled for i.e. winning the lottery, 9/11
Maturation internal threat
Internal event that happens in the subjects themselves i.e. cognitive development, motor abilities
Reactive pretest internal threat
Taking a pretest can influence performance on post test, they learn the test
Instrumentation internal threat
Instruments need to be calibrated, but test ever is part of instrumentation as well and can influence the results
Statistical regression internal threat
Students to perform at extremes are likely to regress to the mean upon retest
Differential subject selection internal threat
If subjects are different before the experiment and after, can’t assume the differences are due to the treatment
Mortality internal threat
When you have more dropouts from one group than another
Subject selection external threat
Make sure subjects are an appropriate sample of the population you want to generalize
Interactive pretest external threat
Subjects will be affected by the pretest and some will not i.e. weighing someone for the pretest might make them aware of their weight and motivate them to diet
Reactive arrangements external threat
Subject may be reacting to their environment i.e. people know they’re being watched
Multiple treatment interference external threat
If treatments are combined, there could be an interaction between the treatments
Independent variable
Manipulated by researcher, goes on X axis, values are chosen by researcher, minimum of two level/values, no maximum
Nominal measurement
Using numbers instead of names i.e. sports players jersey number
Ordinal measurement
Numbers have an order or rank, we don’t know the distance between them i.e. first second third place
Interval measurement
Equal units i.e. temperature, no absolute zero, can have negative numbers
Ratio measurement
Intervals between numbers are the same and there is an absolute zero i.e. height, weight, age, no negative numbers
Dependent variable
Affected by the manipulation of the independent variable, outcome measure, goes on Y axis
Simple random sampling
Equal chance of being selected
Convenience sampling
Accidental sampling, selects participants from individuals who are available because of close geographical proximity or other reasons of convenience
Stratified sampling
Divides the target population into a number of non-overlapping subpopulations, such as geographical regions, then draws a random sample from each subpopulation
Participant selection criteria
Should develop a set of criteria for admission/exclusion in a research study
Systematic review
Qualitative way of looking at studies, analyzing threats
Meta-analysis
Combining data mathematically and quantifying results
Statistically significant
Not likely to have occurred by chance
Treatment efficacy
Works under ideal conditions
Treatment effectiveness
Will it work in the real world environment
PICO
Population, intervention, comparison treatment, outcome measurement
Test-retest Reliability
Tests reliability over time
Same test is administered multiple times to the same group and scores are compared
Scores should be stable
Split-half reliability
Scores from one half of the test correlate with results from the other half of the test
Often done by splitting the test into odd and even questions
Alternate form reliability
Or parallel form reliability
Scores on test A and test B correlate
Face validity
The test looks like it assess the skill it claims to assess
Not a valuable measure because it does not look at content or outcomes
Content validity
Test’s contents represent the content domain being assessed
Similar to face validity but judges the actual content of the test and not the appearance
Construct validity
A test measures a predetermined theoretical construct
Criterion validity
Established by use of external criterion, two types: concurrent and predictive
Concurrent validity
Test’s validity in comparison to a widely accepted test
Predictive validity
Test’s ability to predict performance in another situation or at a later time