Questions and CORRECT Answers
What is a null hypothesis? What is an alternative hypothesis? How do you report it? -
CORRECT ANSWER - H0- There is no effect
HA- There IS an effect
"The evidence supports the HA, so we reject the null hypothesis"
"We reject H0 up to the 95% interval"
What is p-value (probability values)?*** - CORRECT ANSWER - 1. Refers to the
probability that a condition or circumstance would happen just by chance without experimental
intervention.
2. Standard of acceptability is 1 out of 20 or a p-value od < or = 0.05****
What is a type I error? - CORRECT ANSWER - You believe there's a relationship when
there is none, or a false positive.
What is a type II error? - CORRECT ANSWER - When you don't believe there is a
relationship, when there is one, or a false negative.
What is the two-sample t-test used for? how do you test this in stata? - CORRECT
ANSWER - To determine whether a sample is different from an expected value, we are
assessing what's different from each other.
ttesti var1 var2
What is ANOVA? - CORRECT ANSWER - Analysis of Variance determines whether
more than two unrelated groups are different from one another by testing a difference in means.
, What does a chi-squared test do? - CORRECT ANSWER - Tests whether proportions in a
sample are different from an expectation. Called a goodness of fit test.
What is ordinary least squares regression? How do you code it in Stata? - CORRECT
ANSWER - It's a means of eliminating "noise" in a dataset so we can see the relationship
between the variables. It uses y=B0+B1 +u as the equation.
regress var1 var2
Bivariate OLS Regression - CORRECT ANSWER - Seeks to describe the relationship
between DV (y) and IV (x). Uses a linear model to line the observations to generalize the
relationship between x and y.
How do we measure correlation? - CORRECT ANSWER - Using "r". We use r^2 to
measure a line between 0 and 1. The closer to 1, the more positive a relationship (45-degree
line). The closer to -1, the more negative a relationship (45-degree line).
Gauss-Markov Assumptions - CORRECT ANSWER - Conditions must hold!
Linear relationship between y and x.
Random sample of data (to make inferences from)
No perfect co-linearity in explanatory variables x
Homoscadasticity: u has the same variance for all values of x.
What is homoscadacity? How do we test for this in stata? - CORRECT ANSWER - It
means the error term is constant across all values of x. When we don't have this, we have
heteroscadacity.
We use estat hettest