Question

2. Now we'll take a look at some potential correlations. It might be interesting to see if certain aspects of the house are more correlated with the sale price than others. Use Analyze > Multivariate methods > Multivariate, and drag all of the variables into the Y box. Use all 4 decimal places that JMP gives you a. What is the correlation between: Finished Area and Bedrooms: 7058 Acreage and Sale Price: 5045 b. Which variable has the strongest correlation with Sale Price? Sale Price c. Use the correlation matrix to identify the correct statements below.(Select all that apply; you have 5 attempts.) Sale Price tends to be higher in homes with more Bedrooms. Homes with a larger Finished Area tend to have more Bedrooms. Homes with more Acreage tend to have a higher Sale Price. Newer homes tend to have less Acreage. Newer homes tend to have less Finished Area. Homes with more Acreage tend to have less Finished Area. Homes with more Bedrooms tend to have less Acreage. From the same JMP output choose the red dropdown menu next to the word 'Multivariate' and select 'CI of correlation'. d. Let's perform a hypothesis test to check if Sale Price and Finished Area are correlated at the population level. The null hypothesis is Ho: p = 0. The 95% confidence interval is: ( The decision is ---Select-- . e. Let's perform a hypothesis test to check if Sale Price and Year Built are correlated at the population level. The null hypothesis is Ho: p = 0. The 95% confidence interval is: The decision i V ---Select--- Fail to Reject HO Reject HO

Answer 1

Que.2

a)

Correlation between finished area and bedrooms : 0.7058

Correlation between acreage and sale price : 0.5045

b)

Finished area variable has strongest correlation with sale price.

c)

Homes with larger finished area tend to have more bedrooms. Because correlation between finished area and bedrooms is higher (0.7058)

d)

From JMP output,

95% confidence interval for $\rho$ is ( 0.7646, 0.8620 )

Since above confidence interval does not contain zero, hence we reject null hypothesis.

e)

From JMP output,

95% confidence interval for $\rho$ is ( -0.0101, 0.2763 )

Since above confidence interval contain zero, hence we fail to reject null hypothesis.