Bivariate Analysis

Question 1

What is the difference between the independent and the dependent variable? What are some conditions of causality?

Answer 1

The independent variable is the cause, and the dependent variable is an effect.
Conditions of causality:
1. The cause has to come before effect
2. There has to be a correlation - they need to be linked!
3. It needs to be non spurious, or genuine

Question 2

What is the difference between the research hypothesis and the null hypothesis?

Answer 2

Research hypothesis: A statement about what we expect to find vis a vis the relationship between the independent and dependent variable. These can be directional (they state which variable they expect to be higher than the other) or non-directional (they just say that there’s an effect.
Null hypothesis: A statement that the two variables do not have an effect on each other

Question 3

What is the logic of hypothesis testing?

Answer 3

1. Test the null hypothesis: We can’t prove that there is a relationship, but we can disprove that there is not a relationship
2. Calculate the probability of observing the test statistic if the null hypothesis was true: This is done through test statistics based off of sampling distributions! The higher the test statistic, the more likely it is that you can reject the null hypothesis
3. Decide whether you reject or fail to reject the null hypothesis

Question 4

What is the critical value?

Answer 4

This is the minimum number of standard errors you need to exceed to reject the null hypothesis, that will limit our error rate. Critical values are determined from:

1. Significance levels: Our error rate, which is kind of the flip side of confidence intervals. If we are 95% confident, that means that 5% of the time, we will be wrong. That means 5 out of 100 times, we will find a relationship even if the null hypothesis is true
2. Degrees of Freedom: For t-test, this is the sample size minus the parameter estimates. For chi-square, this is rows and columns, minus their parameter estimates, multiplied together

Question 5

What are the different types of errors?

Answer 5

Type 1: Finding a relationship where there isn’t one

Type 2: Finding no relationship where there is one

Question 6

What are the two ways of describing bivariate relationships?

Answer 6

A bivariate relationship is a relationship between two variables.

1. Comparing percentages between 2 or more groups. This is what we do for crosstabs
2. Comparing means. 2 groups is a t-test, more groups is regression and correlation