Hypothesis Testing Flashcards
What is a null hypothesis?
The default position (H0)
The original statement given in the question
What is the alternative hypothesis?
The alternative possibility in respect to the null hypothesis (H1)
What is a one-tailed test?
A situation in which you are testing whether a parameter is more than or less than a value
What is a two-tailed test?
A situation in which you are testing whether a parameter is not equal to a value
How do you decide whether to accept/reject the null hypothesis?
If the data is within the critical region for the value stated, accept the null hypothesis
if the data falls outside of the critical region for the value stated, reject the null hypothesis and accept the alternative hypothesis
What are the two methods for conducting a hypothesis test?
- Finding the critical region and comparing to the test statistic
- Finding the probability of being at least as extreme as the test statistic and comparing to the significance level
How do you carry out the probability comparison method of hypothesis testing?
- Write out the hypotheses
- State the distribution being used for the test
- Draw the graph and shade
- Write a number line of X values
- Look up probability
- Compare probability and significance level
- Concluding statement
How do you carry out the critical region comparison method of hypothesis testing?
- Write out the hypotheses
- State the distribution being used for the test
- Draw the graph and shade
- Write a number line of X values
- Construct a table of probabilities
- Compare critical region and test statistic
- Concluding statement
What should you do if the probability of the test statistic < significance level?
Reject the null hypothesis
(There is enough evidence to suggest…)
What should you do if the probability of the test statistic > significance level?
Accept the null hypothesis
(There is not enough evidence to suggest…)
How do you find P(X is greater than or equal to A)?
1 - P(X is less than or equal to A-1)
What do you do to the significance level if you are dealing with a two-tailed test?
Half it
10% significance level becomes 5% on each side
Where is the expected value plotted on the bell curve?
At the peak of the curve (the centre)
Where should the test statistic be plotted on the bell curve?
As the critical value
Then, test whether this is true
What does the actual significance level of a test mean?
The probability of this test statistic falling in the critical region, assuming the null hypothesis is correct.
(if P = 0.02 so therefore 0.02 < 0.05 sig level, then the actual sig level is 2%)