FOM 5.2.3 Flashcards
What is the independent variable vs the dependent?
Independent - predictor variable/risk factor
Dependent - Response variable
What is the null hypothesis?
There is no association between exposure and disease
What is the alternative hypothesis?
The exposure is associated with disease
What are the two results from hypothesis testing?
Fail to reject the null hypothesis
Reject the null hypothesis
What is the Chi-square test?
Used to compare proportions between two (or more) different groups with percentages or proportions. Only using categorical data only (no mean values).
Comparing the percentage of members of 3 different ethnic groups who have essential hypertension
What is the t-test?
Used to compare mean values between TWO different groups. Numerical data only
Ex: Comparing mean blood pressure between mean and women
What is ANOVA?
Analysis of variance
Used for data that includes multiple variables
More than two groups or more than two variables
Comparison of mean values (numerical)
Ex: Comparing mean blood pressure between members of 3 difference ethnic groups
What is the “magic number” associated with p-value?
.05
What is the P-Value related to?
A measure of relative consistency between the null hypothesis and the data collected
What are the two types of error and how are they different?
What is a type 1 error?
You say groups are different yet they are the same
What is a type 2 error?
You conclude the groups are the same but in reality they are different
What are the alpha and beta in regards to type 1 and 2 errors?
Alpha - Probability of making a type I error
Beta - probability of making a type II error
If Beta = .20 then what is the power?
80%
What are the factors in calculating sample size?
Because if there really is a big difference, you don’t need many people to see it (conversely, if you want to be able to detect a really small difference, you need a larger sample size)
If you’re not willing to take much of a risk in making a type I error, you need more people (the more people you have, the less likely a random error becomes)
If study power becomes smaller, beta gets bigger; this means that you’re willing to take a bigger type II error risk, so you need fewer people (a smaller sample)
