research lecture 7 Flashcards
what 2 test are looking for differences
T-test
anova
BIVARIATE CORRELATION
INTRACLASS CORR. COEFFICIENTS
REGRESSION
INTERNAL CONSISTENCY
KAPPA
FACTOR ANALYSIS
these are looking for ____
relationships
what is the relationship between raters scored called
inter rather reliability
What is the relationship between the two sets of scores from the same rater
intra rater
What is the relationship between gait speeds captured by the Vicon in patients
following a stroke at time 1 and again at time 2….can each of the patients walk
at about the same speed
test - retest reliability
Are the items in a test or survey consistent with each other? what is this asking
internal consistency
what is it called when we see if a score on one outcome measure can predict another outcome measure
predictive validity
what is it called if we ask Can two different outcome measures get similar results
concurrent validity
Can two different outcome measures (one is a gold standard) get similar results?? what validity is this
criterion
Can sub-groups of items in an outcome measure explain factors of the same
complex construct….such as satisfaction, coordination, etc???? what validity is this
construct
what research is pearson correlation commonly used in
methodological research (investigating a toll)
what data must pearson correlation have
internal/ration data (quantitative data)
what is the statistical abbreviation for a Pearson Correlation statistic.
r
if r is +1, 0 or -1 what does this mean
+1 = perfect positive correlation
0= no relations
-1= perfect negative correlation
what is the significance of pearson correlation
shows the probability of the relationship happening by change
if the p value is .0032…the there is only a .32% chance that the relationship occurred just due to chance. In other words…this is a “-____ ” relationship.
real
is p is < ___ then there is a real relationship
.05
does pearson R have something to do with the graph slope
no
if the R is 0 how will the line look on the graph
it will be horizontal
“r” is also an __ ___ … just like cohens d but used for _____
effect size
relationships
what effect size is it for these r values
+/- .1=
+/- .3=
+/- .5=
small
med
large
what is r^2 called
coefficient of determination
what is the coefffeiceint of determination (r^2)
how much of the variability in one variable can be predicted by other variable
A correlation of r=.845 means that r2 = .714 or ____% of the variability inndevice B can be predicted by device A….and vice versa.
71.4%
does statistical significance always imply a useful relationship ?
no
Statistical significance does not necessarily imply a useful relationship.
For example….if you have 90 subjects (N = 90) with an r value of 0.205….it will be statistically significant, but only 4% (r2= 0.04) of the variance is
shared.
just see the exam0ple lol
if r=.8 so r^2= .64 what does that mean
64% of variable y can be predicated by variable x … vise versa
does pearson correlations alwaays measure teh association and the agreement ?
no it measure the association but not always the agreement
what is an example of a poor agreeement
if u have 2 raters and one rater scores it a 10 and the other rater scores it a 100
___ ____ does not mean causes and effect
correlation coefficient
if r was .952 what kind of relationship would u say it has
strong significant positive relationship
if p < .05 would u reject the null
yes
what is abbreviated as r^rb
posing biserial correlation
which correlation is run between 2 levels of categorical variables (ex/ 2 groups ) AND an interval/ratio level variable
point diserial correlation
is point biserial correlation and pearson correlation ran the same in SPSS
yes
Is there an association between bachelor’s trained nurses and master’s trained nurses and their average patient satisfaction scores at Methodist hospital?
this is an example of what kind of correlation and why
point biserial correlation and bc it has 2 groups (bachelors trains nerves and masters trained nurses) and is an interval/raion levle variable (patient satisfaction scores)
what is the non parametric test that is equivalent to the pearsons correlation
spearman coefficient
for the spearman coefficient the data in SPSS will be into what order
rank ordred data
what kind of data is spearman coefficient used with
ordinal
You are interested in childhood aggression. Specifically, you would like to know the relationship in mobility scores and social skills scores in formally
premature children. These children are now all 3 years of age. These two tools are scored using ordinal data. Higher scores are these tools indicate higher levels of mobility and social skills.
what correlation would u use for this
spearman coefficient bc it is ordinal data
The rs value is calculated from this summed D2 value. Basically…the ___ the difference the ____ the association.
spearman example
larger
smaller
since we aren’t analyzing the actual numbers for spearman’s correlation does the high or low scores matter ?
no
what is internal consistency
how closely related the items in an outcome measure are as a group
Ex/ written exams
Ex/ self-report outcome measures
Ex/ performance measures
these are examples of what consistency
internal
what statistic is calculated form averaging all the possible pairwise correlations between ideas and also takes into account the number of items and variance too
cronbach’s a
internal consistency value varies from …
0-1
the guideline values for interpretation for internal consistency is what
.7=
.8=
.9=
acceptable
good
excellent
for the coefficient of cronbachs alpha what do these values mean.
more than .9
.8-.89
.70-.79
.60-.69
.50-.59
less than .59
excellent
good
acceptable
questionable
poor
unacceptable
what does the standard error of th mean measure
the average distance between the M (sample mean) and u (population mean)
SEM = σ/√𝑛
sp as the n increases (population) what happens to the SEM q
decreases
___ provides a measure of how accurately, on average, a sample mean represents its corresponding population mean.
SEM
the standard error of measurement is related to what
test reliability
the vaule of standard error of measurement gives u an idea of what
how much error u should expect form a measurement
is the standard error of measurement (SEm) equation
SEM = SD square root of 1 − 𝑟
what does the r in the SEm equation mean
test retest reliability
68% of the time the SEM is how many SD away
1 .. so if the SEm is 2 and we get a score of 10 68% of the time the pateitns true score would be between 8 and 12
96% of the time the SEm is how many SD away
2 so if the SD is 2 and the patient gets a score of 10 there is a 96% that the true measruemtn will fall within 6 or 14
what is Commonly used reliability coefficient….like a Pearson but with more flexibility.
intra class correlation coefficient
how many scores is there for a ICC and a pearsons
ICC = many
pearsons= 2
Intraclass Correlation Coefficient (ICC) can be a measure of ___ or ____
agreement or assocaitation
the Intraclass Correlation Coefficient (ICC) varies from ___ and ___ is the most consistent
0-1 and 1
Intraclass Correlation Coefficient (ICC) is most often used to calculate what reliability
inter rater (can be more than 2 raters)
what Intraclass Correlation Coefficient (ICC) model is this… each participant is assessed by
a different set of randomly
selected raters. This is rare.
1
what Intraclass Correlation Coefficient (ICC) model is this… each participant is assessed by
each rater, and raters have been
randomly selected and represent all
similar raters. This is common.
2
what Intraclass Correlation Coefficient (ICC) model is this… – each participant is assessed by
each rater, but the raters are the
only raters of interest. This is
somewhere in the middle.
3
the form of the ICC is determined by what
how many measurements there are
what does SPSS called model 1,2,3 of ICC
one way random
two way random
two way mixed
This is a sophisticated way of analyzing reliability…can handle unlimited factors
(facets) and their contributions to the error of a measure…. what is it called
g- theory (generalizability)
the generalizability theory can estimate what 2 things
the relative contribution of identified error and the overall generalizability coefficient
(the closer to 1 the better)
