Research Flashcards
ICF domains
body, individual, and societal perspectives
one list of body functions and one list of domains of activity and participation
variance
-measure variability from the mean
SD
-square root of the variance
type 1 error
alpha
when you reject a true null
type 2 error
beta
when you retain a false null
critical T value
better to have a higher critical t value meaning you are reaching significance so you reject the null hypothesis
independent samples
participants in 2+ groups are unrelated and are observed once and only once. Random
dependent samples
compare the same sample of people multiple times (before and after tx.)
repeated measures
measures same group of people multiple times (pre-post design)
matched pairs study
related people matched in groups because of common characteristics
Face validity
is it measuring what’s supposed to measure
content validity
does it measure all the components of the variable
ex: surveys use this, pain questionnaires (all components of the pain -quality, etc.
criterion validity
how does it compare to the gold standard
ex: homans (poor criterion validity) vs. ultrasound Doppler (gold standard)
construct validity
how well can inferences be made from the results.
ex: results of EMG scores and MMT
confounding variable
associated with the exposure, a risk factor for the outcome, and not intermediate on the causal pathway between E and D
ex: age …Does DM lead to Dementia?
hawthorne effect
individuals perform better because they are in the study
healthy worker bias
occurs when compare workers to non-workers, workers are healthier than non-workers
control selection bias
when controls don’t represent the base population
volunteer bias
when volunteers differ from non-volunteers in study
information bias
errors in procedure for data collection
recall bias
case groups more likely to recall their exposure
interviewer bias
difference in interpretation of responses by interviewer
sensitivity
how well a test detects those with a disease, TRUE positive
SnOUT
specificity
how well a test detects those without a disease, TRUE negative
SpIN
positive predictive value
proportion of people with a + test who actually have a dx
–> TRUE POSITIVE
negative predictive value
proportion of people with a (-) test who do not have a dx
–>TRUE NEGATIVE
correlation
used to describe relationship between 2 levels of independent variables.
DOES NOT INDICATE CAUSATION
r=(-)1.0 to 1.0
0=no correlation
regression
used to predict/used to explain changed in dependent variables
R^2
coefficient of determination
portion of total variance in one measure that can be explained by the variance in the other measure
ex: r=50, r^2=.25, then 25% of variance in Y is accounted for by X.
T test
used to compare 2 levels of 1 independent variable
ANOVA
used to compare 2+ levels/means
uses F statistic
one-way anova
2+ levels of 1 independent variable on 1 dependent variable
-an extension of independent test
-F stat > critical value
IV:age (young, middle, old) DV: BP
Repeated measures ANOVA
used when subjects are tested more than once, extension of dependent t-test
two-way ANOVA
compare 2+ levels of 2 IV ON 1 dv.
two-way ANOVA
-factorial
both IVs are independent (age and sex)
two-way ANOVA
-mixed model
one IV is independent, the other is dependent
two-way ANOVA
-repeated measures
both IVS are dependent
Chi square
analyzes frequency of responses that are nominal
ANCOVA
1+ IVs, 1+covariate –> DV
aka a regression ANOVA
discriminant analysis
1 IV (2+ levels) —> 2+ DVs
MANOVA
2+ IVs–> 2+ DVs (decreases type 1 error)
MANCOVA
1+ IVs, 1+ covariates –> 2+ DVs
covariate
variable possibly involved in study that explains error
a good variate increases statistical power and adjusts groups so it makes comparison fairer
TRUE
LR >1.0
suggests disease is present
LR <1.0
suggest disease is NOT present
kruskall wallis
compares if 3 or more independent samples come from the same population
non-parametric version of an ANOVA
mann whitney test
compares 2 independent samples with ordinal data
nonparametric version of independent t-test
Wilcoxon signed rank test
used to compare 2 dependent samples with orfinal level data/
nonparametric alternative to dependent or paired t -test