Stats Flashcards
What is the lowest level of evidence?
Level 5: expert opinion
What is the highest level of evidence?
Level 1: RCT
What is better evidence? Retrospective case studies or case series
Retrospective case studies- level III
What is a dependent variable?
It is the result, what the researcher is measuring to see if it changes. It is dependent on the interventions
What is an independent variable?
It is the interventions that the researchers are comparing. What the researchers are manipulating or using
Researchers are comparing US to exercise on the effect on neck pain. What are the independent and dependent variables?
Independent variables: US and exercise
Dependent variable: Pain
What is a P-Value?
The probability that the outcome happened by chance or not. Has to have two or more groups
What is an alpha level?
Number of the P-value that researchers set to say the results are significant. (Less than the number is significant)
If the alpha level is set at 0.3 but the P-value is 0.4. Are the results statistically significant?
No because the alpha level was set to 0.3 and the P-value has to be lower
What is type one error?
When researchers say there was a difference but there was not. “Backing a loser”
Type II error
No significant difference between interventions when there was
How do you correct type II errors?
Add more subjects
What is effect size
How much better an intervention is to another.
What are effect size scores? Values for Large, moderate, small, trivial
Large 0.8 and greater
Moderate 0.5-0.799
Small 0.2-0.499
Trivial <0.2
Define intra rater reliability
The probability that the same clinician will get the same results on a test
Define inter rater reliability
The probability that two clinicians will get the same results
Cohen’s kappa scores: or K
0-1
Less than 0.4 is poor reliability
0.4 to 0.6 is fair
0.6 to 0.75 is good reliability
Greater than 0.75 is excellent reliability
1 is perfect reliability
What is a positive likelihood ratio?
Increases the suspicion of a condition based on positive test results
What are values of positive likelihood ratios?
Greater than 10 then large shift toward dx
5-10 moderate shift toward dx
Less than 5 small shift towards dx
1 = no chance
What is a negative likelihood ratio
How much to decrease suspicion of a dx due to negative test results
Name the negative likelihood ratio values
Less than 0.1 means large shift away from dx
0.1-0.2 moderate shift away
Greater than 0.2 small shift away
1 is no shift away from dx
Subjects that know they’re in a research group work harder than they normally would
Hawthorne effect
What is the solution to the hawthorn effect?
Use control group
People work harder when they are watched is called?
Observer effect
What’s the solution to the observer effect?
Make sure all Tx groups get equal Tx time
When the control group works harder because they perceive they’re at an unfair advantage
John Henry effect
How do you correct for the John Henry effect?
Blind the participants
The is the effect where the researchers give extra attention to the experimental group
Pygmalion or rosenthal effect
How do you correct for the pygmalion effect
Blind researchers and examiners
This is the degree to which your study tests what it says it’s testing and isn’t influenced by other factors or variables
Internal validity - examples an event like a car crash or testing a hop test as everyone gets better with practice, people fall out of the study
This is the degree to which the results of the study can be applied to real life
External validity- all the effects and unrealistic design are examples
What test would you use to answer di PT students from my sample have lower testosterone than the known national average?
One sample T-test
What would you use to answer due my Tx groups function improve more than my control?
Independent samples T-test
What test to answer: did my groups function improve after Tx compared to before Tx?
Paired samples T test
What test would you use to answer: did manual therapy, exercise, or a combined Tx group improve the most?
Anova
What stats test would you use to answer: what happened to bone mineral density as age increases?
Pearson r or linear regression