EPI exam 3 - Week 2! Flashcards
The statistical ability of a study to detect a true difference, IF one truly exists between group-comparisions is what?
Power (1-Betta)
Power (1-B) gives us the level of accuracy of what?
Correctly accepting/not accepting the Null Hypothesis.
How is Sample Size relevant in detecting if a difference between groups truly exists?
The larger the sample size, the greater the likelihood.
What are the three factors that go into sample size determination?
- Minimum difference between groups deemed significant
- Expected variation of measurement
- Type 1 and Type 2 error rates & Confidence Interval.
The smaller the difference between groups necessary to be considered significant, the greater the ________ is needed?
Sample Size
What must you do with Error rates and Confidence intervals when determining sample size?
Add in anticipated drop-outs or loss to follow up.
Statistical tests used to determine possible error-rate or likelihood of chance in comparing difference or relationship between variables is reported as what?
P-value
How do we obtain a P-value?
A statistical test critical value is calculated and compared to the appropriate table of probabilities for that test, then a Probability value (P-value) is obtained.
The probability value is based on what?
Probability of observing, due to chance alone, a test statistic value as extreme or more extreme than actually observed if groups were similar.
When is the probability value selected?
Before the study starts (a priori)
If the p value is lower than the pre-selected alpha value (5%) then we say it is _____ and _____
Statistically significant
reject null hypothesis.
If we reject a null hypothesis the risk of what type of error is low?
Type 1
At baseline you want your p values to be what?
Above (5%) because you want to accept the null hypothesis.
Why do you want to accept the null hypothesis at baseline?
Because you do not want your groups to be significantly different until after you have administered your treatment options.
Lavene’s Test
Asses whether variances between groups are different.
What are the three required assumptions of Interval/Ratio Data for proper selection of parametric test?
- Normally-distributed
- Equal variances
- Randomly-derived & Independent
A 95% confidence interval shows what?
We are 95% sure that the true difference or relationship falls within a certain range.
If the confidence interval crosses 1.0 for ratios or 0.0 for absolute differences it is significant/not significant?
Not significant
What is a forest plot in relation to this lecture?
Another way to show CI information to show whether or not there is a statistical significance to the data.
Just because something is statistically significant does not mean it is what?
Clinically equivalent
What is the first question you must ask when deciding what kind of statistical test to run?
What data level is being recorded? Followed by does the data have order/magnitude and does the data have a equal consistent distance along the entire scale?
What does the first question of selecting the correct statistical test give you?
The type of data we are working with.
What is the second question you must ask yourself when selecting the appropriate statistical test?
What type of comparison/assessment is desired.
What is a correlation test?
Provides a quantitative measure of the strength and direction of a relationship between variables. It is given with a -1 to +1 range.
What is a partial correlation?
A correlation that controls for confounding variables.
What is the nominal correlation test?
Contigency Coefficient
What is the Ordinal Correlation test?
Spearman Correlation
Pearson Correlation is for which data type?
Interval
If the P value is greater than a 0.05 in the pearson correlation it means what?
There are no linear correlations present, but there may be non-linear correlations present.
All correlations can be run as what to control for confounders?
Partial correlations