Research validity Flashcards
Assignment/allocation bias, selection
pretest selection that may effect the outcome
repeatability or test retest reliability
measure someone now and later to get the same score (same person or two different people)
Attrition
Those that dropped out from the study overtime
Compensatory or rivalry
one group working and motivated to perform better than the other group.
Resentful demoralization
Opposite of rivalry, resentful control group not working as hard as the intervention group.
Diffusion
treatment effect spread from treatment group to control group.
Maturation
subjects change during the course of the experiment
Testing
subjects getting better at the test
Regression to mean
a person who scores high or low the first time will not score as high the second time.
History
event outside of the study has affected the subject
Instrumentation
Change/error in equipment that altered the outcome.
Similar to a paired t test, multiple measures overtime taken on groups
repeated measures annova
Parametric test
assumptions about the distribution
robust to some violations of the assumptions
Non parametric tests
distribution free tests
fewer assumptions
less efficient
Yes or no, ratio outcome
logistical regression
Case Study
retrospective research on single subject
Case series
retrospective research on a few subjects
Cross sectional
snap shot in time
Case controls
retrospective research examining risk factors between 2 groups
Cohorts
usually prospective research following people over time with specific risk factors; bias present
Randomized control trial
at least 2 groups with control group and intervention group; all subjects had equal chance of being in either group
Non randomized control trial
at least 2 groups with control group and intervention group; subjects not randomly assigned to group
Randomized cross over trial
at least 2 groups with control group and intervention group; subjects participate in treatment period, take a break during the washout period to return to baseline, then the subjects switch groups
All participants provide one data point for both the control group and intervention group.
Pre post design
Measurements taken at beginning and end; no intervention
Meta analysis
a systematic review that makes a quantitative conclusion from pooled data
Single subject
prospective research on single subject
