Chapter 1 Regression, Correlation and Hypothesis Testing - Year 2

Question 1

What do you need to be very careful about when working out PMCC?

Answer 1

Weather it is a + or -

Question 2

If you are given data and told that it is encoded using x = logR and y = logP what do you need to do before calculating the PMCC?

Answer 2

You need to log both sides

Question 3

What is the explanation of the relationship between 2 sets of data (both in the form of a logarithm) when it has a PMCC of high positive or high negative?

Answer 3

There is a strong linear correlation with the data in the form x=logM and y = logP which suggests that a model in the form y = kx^n is a good fit

Question 4

When doing a hypothesis test final line explanation what do you need to make sure to include?

Answer 4

Remember to include the significance rate

“At a 1% significance rate”

Question 5

When you have an equation in the form y=ax^n

What is the y intercept?
What is the gradient?
What would the graph look like if you took logs of both sides?

Answer 5

Log a is the y intercept
n is the gradient

Question 6

When you have an equation in the form y=ax^n

What is the y intercept?
What is the gradient?
What would the graph look like if you took logs of both sides?

Answer 6

Log a is the y intercept
Log b is the gradient

Question 7

See FC on exponential modelling in pure

Question 8

What is a PMCC value

Answer 8

This is given the letter r

r ranges between 1 and -1
r=-1 means it is perfectly negative correlated
r=1 means it is perfectly positive correlated
r=0 means no correlation

Question 9

How do you calculate the PMCC on a calculator?

Answer 9

Menu
6- stats
2- y=a+bx
Input data
Options
4

Question 10

What is a common mistake with PMCC?

Answer 10

Be careful weather it is a + or -

Question 11

What is a little trick that is used in PMCC?

Answer 11

It may say that the data is encoded using logs.
If so then if the logged data produces a extreme PMCC value then there is a strong exponential relationship between the 2 variables (not a strong linear relationship)

Question 12

What will the first make i a hypothesis test always be for?

Answer 12

Stating the null and alternate hypothesis

Question 13

Where are the critical values for correlation coefficients?

Answer 13

Page 37 of the formula booklet

Question 14

What is the notation for a null and alternate hypothesis?

Answer 14

Question 15

How do you get the product moment co efficient level

Answer 15

You divide by 100

Eg1% goes to 0.01

Question 16

Explain how you would do a 1 tailed (directional) hypothesis test?

Answer 16

State the null hypothesis
State the alternate hypothesis which is p is greater than 0 or it could be p is less than 0 (as it is only a 1 tailed test
You then state the sample size given
You then state the level of significance given
You then go to the formula booklet and read of the critical value for your test and state that
You then work out the PMCC value for your data

Then show wether your PMCC value is greater or less than the critical value. If the PMCC value os on the extreme end of the critical value then you reject your null hypothesis and say that there is insufficient evidence to indicate a positive/negative correlation between the 2 variables

Question 17

Explain how you would do a 2 tailed (non directional) hypothesis test

Answer 17

State the null hypotheis which is p=0
State the alternate hypothesis which is p is not equal to 0 (because it is 2 tailed)
Then you state the sample size
Then you state the level of significance but you then half it as it is a 2 tailed test
Then you state the critical value but you must add + or -
You then state your PMCC value

You then show that your PMCC value is on the extreme side od the PMCC value

You then accept/reject your null hypothesis and write a statement saying that there is insufficient evidence to suggest … between the 2 variables