Lecture 4

Question 1

What is the difference between a null and alternate hypothesis

Answer 1

Null hypothesis; there is no association or difference bw 2 variables
Alternate hypothesis: there is an association or difference bw 2 variables

Question 2

How do we test a hypothesis

Answer 2

testing a hypothesis

1. set up the null and alternate hypothesis
2. find the value of the test stats
3. Refer test stats to know distribution follows the null hypothesis
4. Find the probability if you are wrong, if there is a difference bw 2 values
5. decide if there is a difference bw the 2 groups or not

Question 3

Discuss the possibilities of being wrong

Answer 3

There are 2 possibilities of being wrong
type 1 
usually set 5%, there is a 5% chance of being wrong 
type 2 
usually set at 2% chance of being wrong

Question 4

What is the significance of the comparing the means?

Answer 4

Comparing the means to determine if they are independent of each other other or are they related

Question 5

When is the null hypothesis rejected

Answer 5

The null hypothesis is rejected when the P values is less than 0.05%

Question 6

Explain the significance of Z values

Answer 6

For the Z values look at where most of the data sits, its the limit like the bell curve for vet, you set the lines for which people can pass and if you are on the outside you are in different population. You can make a standard curve and compare that to the population mean. There can be an overlap b/w the types of error you can move the line to decide how much error you want.

Question 7

What is the T test

Answer 7

T-test tests the hypothesis, there is an alpha level (its the critical value), which is the cut off line for the different populations, its like the decision level of if student fail or pass the semester, there is 5% chance of being wrong which is when you get a sup.

Question 8

What is the P value

Answer 8

P value is the probability of being wrong in concluding the different bw 2 groups

Question 9

What would the p value be if the null hypothesis =0

Answer 9

null hypothesis =0, there is no difference bw 2 groups

Question 10

When should you use a 2 tailed t-test?

Answer 10

2 tailed t-test is used when your not sure if the results will sit with in the range you have set with type 1 and type 2 error
-Are these means difference?

Question 11

When should you use a 1 tailed t-test?

Answer 11

1 tailed t-test;

is 1 mean larger than the other mean?

Question 12

When should you use a t-test

Answer 12

t-test is used to compare 2 means
assumes normal distributed,
if the sample size is small the T test is less reliable because distributed will not be normal

Question 13

When do you use a pool T test

Answer 13

pool t test is when then the data is independent, so its not matched

Question 14

When do you use a paired t test?

Answer 14

paired t test for 2 measurements for 1 animals (matched)

Question 15

When do you use one way ANOVA test?

Answer 15

one way ANOVA test, is used to test for difference among 2 or more independent groups, a one way is used to test differences among at least 3 groups
The critical value is the cut off point, if its inside or outside the range to see if the means differ

Question 16

when do you a ANOVA test

Answer 16

ANOVA test, testing the difference bw 2 means, you don’t which mean is different from the other
it assumes the data is distributed normally
testes b.w more than 2 means
tells the difference bw more than 2 means
F distribution is the family of distribution, depends on the number of degrees of freedom.

Question 17

When is the chi-square test used

Answer 17

Chi-square test
its for data which isn’t normally distributed
works on degree of freedom (n=1)
can compare numbers of different groups

Question 18

When is the validity of chi-square tests questionable?

Answer 18

All the values should be above 5 if not use the mann Whitney test because its too small or you could combine the categories to include the size

Question 19

Which of the following is parametric tests and non-parametric

• t test
• ANOVA
• chi-square

Answer 19

Parametric test (normally distributed data)
interval ratio
-t test
ANOVA
Non-parametric (not normally distributed data)
count
chi-square