What are proportions?
Where the numerator is included in the denominator (e.g. Prevalence of acid reflux in GI-related illnesses / GI-related illnesses in the population)
What are means?
The average of a group of given numbers (e.g. age, weight, etc.)
Null vs. Alternative Hypothesis?
Null hypotheses assume no differences in the groups being compared, while alternative hypotheses claim the opposite
What is the p-value and how is it interpreted?
P-values are the probability of getting the result collected in the experiment (i.e. HA) assuming the null is true in relation to the alpha (usually 0.05)
If greater than 0.05, the null hypothesis is accepted as no correlation is observed
If less than 0.05, the null hypothesis is rejected as correlation is observed
How do you interpret CI?
The wider the CI is, the more uncertainty there is. If the CI value crosses 1 (the null value), the estimate is not statistically significant.
When do you use a t-test?
When calculating means (numeric data).
When do you use a Chi-square test?
When testing proportions (categorical data)
What are regression models in epidemiology?
Describe the relationship between exposures and outcomes
How are regression models useful?
- Determines which E are predictors of O (gets a MOA for E and O)
- Can explore the relationship between one or more exposures and one outcome (uni- and multivariable)
- Controls for confounding
What are the two types of regression models and when do you use them?
Linear - Used for numeric outcomes (i.e. averages / t-test numbers)
Logistic - Used for dichotomous outcomes (i.e. proportions / Chi-square categories)
