Final Review by Alex :) Flashcards
What sampling is this: Everyone in population has equal chance of being selected
Simple Random
What kind of sampling is this: we will select every 5th student
systematic sampling
What kind of sampling do you divided the population into parts or sections and take people from each section
stratified
What kind of sampling: You divide the population into groups or section out only select from certain parts
cluster sampling
What kind of sampling is it when you take volunteers
Convinence samping
What level of review: anything with kids
Full
What level of review: more than minimal risk involved
full
What level of review: blood draw, physical exam
expedited
What level of review: survey of nonsensitive nature that are not anonymous
expedited
what level of review: no more risk than daily life
expedited
What level of review: educational setting using normal curriculum
exempt
what level of review: anonymous survey
exempt
What level of evidence: RCT
1b
What level of evidence: All or none
1C
What level of evidence: Cohort or low quality RCT
2B
What level of evidence: Outcome research
2c
What level of evidence: Indvidual case control study
3b
What level of evidence: case report, small case series
4
What does PEDRO assess?
What is a good score?
Assesses RCT
6+ is good
What does Quadas assess?
What is a good score?
diagnostic accuracy of test (sensitivty and specificity)
7+ is good
Who is the primary author in a study?
Who is the senior author in the study?
Who is the major contributor?
First author
Last author
Second author
Where do you find the consort diagram in research
Methods section
What does alpha level represent?
What does P- value represent?
alpha: pre-set significance that the result is due to chance
P value: actual probability that the result is due to chance
What is a type 1 error?
What is a type 2 error?
Type 1: Reject the null hypothesis when u shouldnt
Type 2: Fail to reject the null when you should have
What is the definition of power?
The chance to find a significant difference if there’s one to be found
What does Cohens D Measure?
Effect size
If effect size goes up, power goes _______
Up
What is considered a small Cohens d?
medium?
large?
0.2
0.5
0.8
Experiment vs Quasi-experimental
Experiment - randomly assigned
Quasi - cohorts
What is a methologicial study
Deals w/ reliability and validity of clinical research measures
What is it called when you divide a test into two and give each form to one group of individuals to measure reliability
split half reliability
Two equivelant sets of questions are given to the same sample to measure reliability
parallel forms reliability
What is content validity?
Does the outcome measure have every domain of the thing it’s suppose to measure
What is face validity?
Simple observation of whether or not the instrument measures what its suppose to measure
Concurrent vs critereon based validity
concurrent- comparing 2 outcomes which aren’t gold standard
Criterion-based validity - does one test coorelate with the gold standard test
What is predicitive validity?
What is construct validity?
predictive- can it predict the future preformance outcome
construct- does it measure what it’s suppose to measure
How do you calculate sensitivity
True positive / (True positive + False negative)
How do you calculate specificity
true negative / (true negative + false positive)
How do you calculate positive predictive value
How do you calculate negative predictive value
true positive / (true + false pos)
true neg / (true neg + false neg)
As the prevalance of a disease goes up, what happens to positive predicitve value?
Negative predictive value?
PPV goes up
NPV goes down
How do you calculate Positive likelihood ratio?
Sensitity / (100-specificty)
How do you calculate negative likelihood value
(1-sensitive) / specificity
What is considered a good Negative likelihood ratio
.1 is best, closer to 1 is worse
What is considered a good positive likelihood ratio
Closer to 10 is good, Closer to 0.1 is useless
How much influence does a LR+ of 2 have on the probability
15%
1= 0%
2 = 15%
5= 30%
10= 45%
How much influence does a -LR of .5 have on probability?
15%
1 = 0%
0.5 = 15%
0.2 = 30%
0.1 = 45%
Name the lines from Left to Right
Left: Mode
Middle: Median
Right: Mean
What is considered a small effect size?
Medium?
Large?
Small: R = .1 d = .2
Medium: r = .3 , d= .5
Large: r = .5 d = .8
How do you increase power?
Increase or decrease effect size
increase or decrease sample
alpha level 0.01 or 0.05
1 or 2 tail test
Increase effect size
Increase sample size
Alpha = .05
Use 1 tail test
What is the nonparametric independent T test called?
What is the repeated measure nonparametric T test
Mann Whitney U
Wilcoxon signed ranks test
What is a bonferonni correction?
Divides the alpha level by # of t tests
example 0.05 alpha level becomes 0.025 with 2 tests
What kind of anova do you use with multiple independent variables and repeated measures?
Split plot ANOVA
How many independent variables does this ANOVA have? 2x3x4x5 ANOVA
4 IVs
What is more powerful with large numbers of comparisons, bonferoni or tukey
tukey
What kind of alpha adjustment can you use if the variance is not homogenious
Games-Howell
How many pairwise comparisons in this 3x2 ANOVA?
9
What is the nonparametric version of a one way anova?
Kruskal-Wallis test
Mauchley’s test is testing what assumption?
Sphericity- equal variance between all possible pairs
For what kind of test is sphericity an assumption?
Repeated measure ANOVAs
What is the non parametric equivelant to a repeated measures ANOVA
friedman’s anova
-uses wilcoxon signed ranks
What is a factorial anova?
An ANOVA with more than 1 independent variable
A 2x2 factorial anova tests what things?
- Interaction
- Main effect of first IV
- Main effect of second IV
What kind of ANOVA takes into account factors that can influence the result
Example: Patient compliance
ANCOVA
Takes into account co-variants
What kind of ANOVA can have multiple dependent variables?
MANOVA
Rater 1:
1
10
100
1000
Rater 2:
.1
1
10
100
Do these have a weak or strong pearson coorelation
Pearson coorelation of 1 = strong
A pearson coorelation of what is considered strong?
over .5
note:
.1 small
.3 medium
.5 large
What is R^2 (R squared)?
Coefficient of determination
determines how much one variable can predict another
T or F: Pearson coorelation must be high to have a useful relationship
F
What test would you use to compare 2 levels of categorical variable and one interval/ratio variable based on how much they coorelate
Point-biserial coorelation
What is the non-parametric equivalent to the pearson coorelation
spearman coefficient
What static measures internal consistency?
Cronbach’s alpha
What is considered an acceptable cronbach’s alpha?
Good?
Excellent?
Acceptable: .7
Good: .8
Excellent: .9
How do you calculate standard error of the mean?
SD / Square root of N
What percent of patient’s are within 2 SD of the mean?
95%
What test is for reliability but can be used for more than one set of scores
commonly used to test inter-rater reliability
ICC
Intraclass correlation coefficient
What is model 1 ICC?
Each participant is assessed by a different set of random raters
What is model 2 ICC?
Each participant is assessed by each rater, raters represent all simular raters
What is model 3 ICC?
Each participant is asssed by each rater, but the raters are the only raters of interest
What is the best resource to search for predatory journals?
Where is a good place to find legitimate journals?
Beall’s list
doaj.org
Who has to pay for open access to an article?
Author pays
How is H-Index calculated for a specific author
Lists the authors papers and the number of citations they receive
What reference software does Dr. Mitchell want us to use?
Refworks
Is this acceptable writing in research:
“Patient did 20 reps of quad sets”
No
You have to write the full explanation of what they did
What are the 3 levels of categorization for unversities doing research
Which one is TWU?
Doctoral Level
R1- this one is the best
R2-
Doctoral/professional- Texas Womens (3rd best)
Masters Level
Bachelor
What statistic measures inter-rater reliability ONLY with nominal data
Kappa
Kappa requires what kind of data?
Nominal
What Kappa score indicates perfect agreement?
1 = perfect agreement
0.5 moderate agreement
0= no agreement
What test measures whether or not 2 independent variables are independent from one another
Chi Square
(measures independence or relationship)
Chi square can test for what two things
Independence OR relationships
just depends on how you write your hypothesis
What kind of test is used to find different groups of factors or constructs in an outcome questionarre
Exploratory factor analysis
What factor replaces reliability in qualitative research
dependability
What factor replaces external validity in qualitative research
transferability
What factor replaces internal validity in qualitative research
credibility
What kind of regression used one I/R predictor and one I/R outcome?
What uses multiple predictors and one I/R outcome?
Linear
Multiple Linear
What kind of regression uses predictors of any level of data and an outcome of 2 categorical variables
what kind of regression uses predictors of any level and multiple categorical outcomes.
logistic
multinominal logistic
What kind of regression has one or more predictors at any level of data and one ordinal outcome with multiple levels
ordinal logistic
What is a greenhouse geiser correction?
Used when theres an issue with sphericity in RM ANOVAs
