Project 3 and 4 (Spring 2008)

Stat 201 – Section 4
Spring 2008
Ramón V. León

Project 3 and 4
Due March 7

The data for this project is being emailed to you or can be found in the Stat 201 – Leon home page There are 3364 observations, which is the entire population of the percent of product bought in the retail stores for one major retail chain. Stokely, Nabisco, Tyson, or other such major companies are frequently interested in the percent of their product bought in a store. For instance, most such companies might have anywhere from 50 to 100 products to place in a store. The right placement combination maximizes the manufacturer or supplier profits as well as the profits of the retail store. There are ten products in this data set with some auxiliary variables for later analysis. The variable value for the product is the percent of the product sold in that store. Some products might have a value of zero because of either no sales or no placement of that product at that retail store.

(1) You are to select two products. The first product is products 3 and the second is product 8.

(2) Select random sample A of size 21 from the 3364 observations of product 3 (prod3). For this random sample find the median, mode, mean, Q1, Q3, IQR, range, variance, standard deviation, and the coefficient of variation. Construct the stem-and-leaf and box-plot. Construct the 95% confidence interval for the mean percent product sold. Interpret your findings. All of this is to done with calculator. Show your work.

(3) Select random sample B of size 62 from the 3364 observations of product 8 (prod8). For these sample do all the above in (2) but also include the histogram and the normal probability plot. This is to be done with either JMP or NCSS. Interpret your findings.

(4) Using NCSS or the JMP bootstrapping macros that I supply with detailed instructions and construct the 95% bootstrap confidence interval for sample B. Compare your answer here and in (3). Which interval estimate is more appropriate [(3) or (4)] and why?

(5) Use NCSS’s bootstrapping macro or JMP’s bootstrapping macro to do a two sample confidence interval for the difference between mean percent of sales (resample=5000 with samples A and B). Use a 95% confidence. Interpret your answer.