Final Flashcards
What is validity?
The best available approximation of the truth or falsity of a given inference, proposition, or conclusion
What is conclusion validity concerning?
Relationship between the study and what was reported
What is internal validity concerning?
Whether the relationship concluded was causal
When generalizing observational findings to substantiate a hypothesis, what type of validity is needed?
Construct validity
When generalizing observational findings to other persons, places, and times, what type of validity is needed?
External validity
Name four types of validity
- conclusion
- internal
- construct
- external
What are two major types of sampling methods?
- probability sampling
- nonprobability sampling
Probability sampling uses what kind of selection?
Some form of random selection
Nonprobability sampling uses what kind of selection?
Systematic or haphazard, but not random
groups in sampling
Who is the theoretical population?
Who is being generalized
groups in sampling
Who is the study population?
The population they have access to
groups in sampling
What is the sampling frame?
How/when they got patients
generalizing studies
Continuously assessing how the study group generalizes to the theoretical population is an example of what model?
Model I: Sampling
generalizing studies
What model examines the gradients of similarity to the times, places, people, and settings of the study?
Model II: Proximal Similarity
How does random error affect an average? How does it affect the variability around the average?
No effect on average, only increases variability around the average
measurement of error
The true score model includes which factors?
Study measurements equal to “true” value plus both random error and systematic error
How does systematic error affect an average? What is this effect called?
Changes the average, called bias
If there is likely another cause for patient improvement beyond the intervention, this is a threat to:
Internal validity
Name three measures of central tendency
- mean
- median
- mode
Name four descriptive statistics for central tendency
- dispersion
- range
- variance
- standard deviation
Dispersion is higher when…
scores are more spread out in distribution
What is range?
Difference between the largest score in the data and the smallest score
What is variance of a sample?
A measure of how data points differ from the mean
When is coefficient of variation useful?
Measuring relative dispersion or comparing the spread of two distributions
(i.e., to compare the distribution of two variables with vastly different mean
values)
Describe a normal distribution’s shape
Roughly bell-shaped
__% of all scores fall within 1 standard deviation of the mean
__% within 2 standard deviations
__% within 3 standard deviations
68% within 1
95% within 2
99.7% within 3
Larger standard deviation means…
more scatter about the mean, worse precision
Smaller standard deviation means…
less scatter about the mean, better precision
What is interquartile range?
Range in the middle of the data, the difference between the upper and lower quartiles in the data set
A statistical difference is a function of difference between ___ relative to the ___.
difference between means relative to the variability
A small difference between means with large variability could be due to…
chance
The P in P value stands for…
probability
Smaller P values indicate…
stronger evidence against the null hypothesis
What P value is a scientific benchmark for concluding fairly strong evidence against null hypothesis?
<0.05
A .95 confidence interval means the hypothesis is rejected at P value…
.05
A .99 confidence interval means the hypothesis is rejected at P value…
.01
What is statistical power?
The ability of a study to find a significant result (if one really exists)
The two measures of association used most often are:
- relative risk/ risk ratio
- odds ratio
measures of association
How is risk ratio expressed?
chances of something happening/ chances of all things happening
measures of association
How is odds ratio expressed?
chances of something happening/ chances of it not happening
When is risk ratio used?
When comparing outcomes of those exposed and those not exposed
When is odds ratio used?
In case-control studies or cohort studies
What is survival analysis?
Analyzes time is takes for an event to occur, such as relapse or death
What is regression analysis?
Analyzes relationship between variables, such as age and risk
What is sensitivity?
True positives from those known to have the disease/condition
What is specificity?
True negatives from those known to not have the disease/condition
What is positive predictive value?
The proportion of true positives among all positive test results
What is negative predictive value?
The proportion of true negatives among all negative test results
What does the receiver operating curve compare?
Sensitivity to specificity
If there is likely another cause for patient improvement beyond the intervention, this is a threat to…
internal validity
What is the risk of having low back pain if 10/25 men typically have it in a year?
10/25 = .4
What are the three parts of EIP?
Research, patient presentation, clinical expertise
The possibility that a healthcare outcome is due to something other than the healthcare represents what fallacy?
Fallacy of false attribution
What is used in inferential statistics to detect a significant difference between two samples?
P value
If the proposed theoretical biological method of action of intervention does not make physiological sense, this is a threat to…
construct validity
We can increase generalizability through…
varying patient characteristics
How do you formulate a searchable question?
PICO
List study types from highest to lowest level of evidence
- meta-analysis
- systematic review
- randomized control trial
- cohort
- case-control
- case series
- case study
- animal model
What is PRISMA used to evaluate?
Meta-analysis quality
Name four ways to reduce error in your study
- triangulation
- review
- pilot work
- training
When too many subjects are lost to follow-up or unequal loss of subjects between two groups happens, which type of bias could be present?
Transfer bias
A more reliable measurement will have a ___ range of data.
narrow
If you were to compare the distribution of glucose level with height, which descriptor would be best?
Coefficient of variation
If the odds ratio for men to women for lifetime obesity is 1.2, what does this mean?
Men are 1.2 times more likely to have obesity
If the heterogeneity is high for a particular meta-analysis, then…
the combination of data is not substantiated
Reading the table of contents of a few key journals is an example of…
keeping current
Randomized controlled trials must include…
an experimental intervention
If laboratory findings are not likely to be applicable to patients in a clinical setting, there is a threat to…
external validity
If the outcome measures used for a study are on non-validated scales that may have reliability or reproducibility problems, there is a threat to…
conclusion validity
When there is no blinding to treatment, it may result in…
performance bias
If you were to compare the distribution of two variables with vastly different mean values, which descriptor would be best?
Coefficient of variation
P50 in rank percentile is the same as which other term?
Median
What is used to evaluate systematic review quality?
PRISMA
Name two factors that the power of a study is influenced by
Number of patients, magnitude of effect
When historic controls are used as a comparison group for patients undergoing an intervention, it may result in…
chronology bias
Randomized clinical control trials are always…
prospective
The difference between probability and non-probability sampling is…
probability sampling uses random selection
If the proposed theoretical model of cause and effect is not supported, this is a threat to…
construct validity
After performing thorough literature search, you must…
appraise the articles
Conclusion validity asks the question:
Is there a relationship between the proposed cause and effect?
