Exam 1 Flashcards
objective
measurable, quantifiable
subjective
rated, ranked, can’t be linked to a quantifiable degree or number
criterion referenced
comparitive;
superior-good-bad-poor etc
norm -referenced
scored against what is considered to be normal
discrete measuremtns
pass or fail (no actual score)
continuous measuremtns
infinite number of score (height or weight)
-dependent on accuracy of measuring tool
formative eval
takes place during the event
summative eval
takes place after the event
all research..
determines a relationship or difference
measuremtns are
true value +/- some error
X=T+E
validity
how well a device measures what it is supposed to measure w/ some acceptable error
valid X = T+ acceptable error
how can you know what is acceptable error?
increase in internal validity
more accurate - takes more time and more cost
reliability
dependability , consistent, repeatable
reliable score = true score +/- consistent error
reliability determined by
validity, repetition of measurement, stability of measured characteristics
objectivity: subset of what? determined by what?
subset of reliability
determined by validity, clear directions, ability to understand instructions
inter-rater reliability
between different judges
maximize objectivity
select measures who are not biased screening clear directions train the measurers monitor scoring
construct validity
evaluation made up of multiple factors to determine an ultimate idea
threats to construct validity
hypothesis guessing (participants guessing the outcome)
bias
expectations of researcher
criterion validity
requires criterion for comparison
measure of how well on variable or set of variables predicts outcome based on information from other variables
very precise standards need criterion validity
calibration
action taken to improve the criterion validity of a device
must be exact
content validity
simplest form, relies on logic and comparison
requires definition of measure
Ecological Validity
methods, materials, and setter of a research experiment
similar to real life situation
external validity
ability of the results to be generalize to other situations and groups
threats to external validity
internal validity, placebo, participant understanding, test group, hawthorne effect, lighting, etc
internal validity
extent to which a causal conclusion based on a study is warranted; minimizes error
threats to internal validity
bias, subjects don’t change, consistency
test wise effect, sequencing, instrumentation, contamination
central tendency
describe best of data, descriptive stats, mean;median;mode
mean
average
median
exact middle
mode
occuring the most
frequency distribution
table of every score
measure of dispersion
standard deviation
standard deviation
bell shaped curve, shows high, lows and average
measure that is used to quantify the amount of variation or dispersion of a set of data values
measures variability of scores/ shows amount of error
ability to make predictions about population sample
data set
population, sample, inferential stats
t-test
determines if 2 set s of data are significantly different from each other
p-value - probability of error
independent vs paired
paired: pre test and post test
independent: more than one measure that are independent of each other
one tailed and two tailed
one tail: predicted value is higher or lower
two tailed: tests both groups against one another
t-test limitations
one for similar populations, sample should have normal distribution, about same # of data points, data should be independent, date should be interval level or higher
correlation test
pearson product moment correlations
measure of strength between 2 variables
eval the correlation
R-value
(-1 0 +1)
stronger at the ends, weaker closer to zero
R =
correlation coefficiatient
R^2
coefficient of determination
amount of variability in one variable
value of 80% or higher is good
Bland Altman Analysis
error analysis
used in place of Pearson, use with large and diverse groups
measure agreement between 2 methods of repeated measures
test - retest
utilizes same measure, observer, instrument, location, time, conditions.
test - rest - retest
strong reliability
time consuming
Spearman correlation
ranked data
one set is ranked, the other is not
doesn’t explain why or performance
