T711 – Fall 2018 Assignment 2. The data contains the information about the direc
ID: 3074914 • Letter: T
Question
T711 – Fall 2018
Assignment 2.
The data contains the information about the direct marketer discussed In class; in particular, the data contains information about the age (coded as young, mid or old) and gender of a customer, whether s/he owns a home or rents, whether or not s/he is married, whether or not s/he lives far from a brick-and-mortar store selling similar items, the customer’s salary, number of children, and purchase history (coded as low, medium, high and NA if there is no history available) and the number of catalogs we have sent that customer in the past. And finally, the data records the amount of money the customer has spent. The following questions refer to this data.
• (Model 1) Let’s investigate the joint impact of salary and the number of children on the amount spent; to that end, run a regression model with “AmountSpent” as the response variable and “Salary” and “Children” as the only 2 predictors; then, answer the following questions
Is “salary” an important predictor for amount spent? Why?
Is “number of children” an important predictor for amount spent? Why?
Does the number of children have a positive or negative impact on the amount of money a customer spends?
Do salary and number of children both have the same relationship with amount spent?
Is this a good model? Why?
Predict the response variable “AmountSpent” for Salary = 30000 and Children = 2 with the option interval="prediction".
• (Model 2) Next, let’s investigate the effect of purchase history on the amount a customer spends.
To that end, first create a “plot of means” (in Rcmdr’s Graphs menu) with History as the factor and AmountSpent as the response variable; attach this plot as Figure 1. What can you learn from this plot?
Now, run a regression model with AmountSpent as the response variable and History as the (only) predictor. Provide a precise interpretation of the resulting coefficient labeled as “History [T.Low]”
• ( Model 3) Next, we want to investigate the joint effect of purchase history, salary and amount spent. To that end, create a XY conditioning plot with salary as the explanatory variable, amount spent as the response variable and history as the condition. Attach the resulting plot as Figure 2. What can you learn from this plot?
Explanation / Answer
Model 1
Is “salary” an important predictor for amount spent? Why?
Yes, because it has a very small p-value; it is very significant.
Is “number of children” an important predictor for amount spent? Why?
Yes, because it has a very small p-value; it is very significant.
Does the number of children have a positive or negative impact on the amount of money a customer spends?
Its coefficient is negative; hence it has a negative impact (the more children a customer has, the fewer s/he spends at the store).
Do salary and number of children both have the same relationship with amount spent?
No, while salary has a positive effect on amount spent, number of children has a negative effect.
Is this a good model? Why?
This looks like a good model as it explain almost 56% of the total variation in amount spent, using only two variables
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.