Data Analysis Essay

[pic 1]

[pic 2]

Solution for Q1:

1. For this question, one-way ANOVA in combination with Tukey-Kramer Procedure is used to answer from part A to C in order to evaluate the difference in the average price of the baskets of 34 common grocery items at different supermarkets. The independent variable is price of 34 items from there different supermarket chains: (1) Coles, (2) Woolsworth and (3) Others.

ANOVA hypothesis testing: First, we key in the data for each of the stores in the excel sheet as follows:

[pic 3]

We clicked “Data”, followed by “Data analysis”, reaching “Data Analysis” as shown below,

[pic 4]

We highlighted the data for “Input Range” and we ticked “Labels in First Row”. Clicked “OK”

[pic 5]

Result shown as follows,

[pic 6]

Analysis

Null Hypothesis: HO = There is significant difference in the average price of basket of 34 items at different supermarkets?

Alternate Hypothesis: H1 = There is no significant difference in the average price of basket of 34 items at different supermarkets

Reject Ho if p-value < 0.05

Accept Ho if p-value >= 0.05

As seen in (a), p-value = 0.00106. Reject Ho.

There is no significant difference of grocery prices among supermarkets.

No. Significant difference in the average price of the basket at different stores does not mean that larger buying power of those chains makes them cheaper than other supermarket.  It just means that there is a difference in the average price of the basket at different stores.

C) To understand the comparison better, we have done the Tukey Kramer Multiple Comparisons. The output is as follows:

Conclusion: There is significant difference in the average price at different stores. But the other supermarkets have the same price as Coles and Woolworths supermarkets since there is no difference between group 1 (Coles) and 2 (Woolworths) with group 3 (other). Hence we can’t agree with first group of analysts, who believe that the largest buying power of Coles and Woolworths supermarket make them cheaper than other supermarkets.