In: Statistics and Probability
Business sensor technology provides a way for companies to learn about their customers, employees, and operations: data captured from sensors can be used to improve engagement, Sales, productivity, safety, and much more. A PwC survey of Global business and IT executives found that 25% of automotive executives: 27% of energy, utilities, and mining executives, 30% of hospitality and leisure executives; 33% of industrial products executives; and 52% of retail and consumer executives say their companies are currently investing in business sensor technology.
Suppose these results were based on 500 business and IT executives in each of the five industries: Automotive, Energy, Utilities, and Mining; Hospitality and Leisure; industrial products; and Retail and consumer.
A) at the 0.05 level of significance is there evidence of a difference among the industries with respect to the proportion of executives that say their companies are currently investing in business sensor technology?
B) Compute the p-value and interpret its meaning.
C) If appropriate, use the Marascuilo procedure and a= 0.05 to determine which companies differ in their currently investing in business sensor technology.
* Solve manually
A)
Hypotheses are:
H0: There is no evidence of a difference among the industries with respect to the proportion of executives that say their companies are currently investing in business sensor technology.
Ha: There is a evidence of a difference among the industries with respect to the proportion of executives that say their companies are currently investing in business sensor technology.
Here we need to use chi square test. Following is the 2 by 5 contingency table from the the given information :
Automotive | Energy, utility and minning | Hospitality and leisure | Industrial products | Retail and consumer | Total | ||
Used Sensor technology | Yes | 500*.25=125 | 135 | 150 | 165 | 260 | 835 |
No | 500-125=375 | 365 | 350 | 335 | 240 | 1665 | |
Total | 500 | 500 | 500 | 500 | 500 | 2500 |
Expected frequencies will be calculated as follows:
Following table shows the expected frequencies:
Automotive | Energy, utility and minning | Hospitality and leisure | Industrial products | Retail and consumer | Total | ||
Used Sensor technology | Yes | 167 | 167 | 167 | 167 | 167 | 835 |
No | 333 | 333 | 333 | 333 | 333 | 1665 | |
Total | 500 | 500 | 500 | 500 | 500 | 2500 | |
Following table shows the calculations for chi square test statistics:
O | E | (O-E)^2/E |
125 | 167 | 10.56287425 |
135 | 167 | 6.131736527 |
150 | 167 | 1.730538922 |
165 | 167 | 0.023952096 |
260 | 167 | 51.79041916 |
375 | 333 | 5.297297297 |
365 | 333 | 3.075075075 |
350 | 333 | 0.867867868 |
335 | 333 | 0.012012012 |
240 | 333 | 25.97297297 |
Total | 105.4647462 |
Following is the test statistics:
Degree of freedom: df =( number of rows -1)*(number of columns-1) = (2-1)*(5-1)=4
The critical value is 9.488
Since test statistics is greater than critical value so we reject the null hypothesis.
B)
The p-value is: 0.0000
Excel function used for p-value: "=CHIDIST(11.119,2)"
C)
The formula for calculating critical range Marascuilo procedure is:
Here we have 5 population so degree of freedom is df=5-1=4
The critical value is 9.488
Following table shows the absolute difference and CR for each of the (5* (5-1)) /2 = 10 pairs:
pi | ni | pj | nj | D=Absolute(pi-pj) | CR | D> CR | Significant difference |
0.25 | 500 | 0.27 | 500 | 0.02 | 0.085429325 | No | No |
0.25 | 500 | 0.3 | 500 | 0.05 | 0.086850216 | No | No |
0.25 | 500 | 0.33 | 500 | 0.08 | 0.088054492 | No | No |
0.25 | 500 | 0.52 | 500 | 0.27 | 0.091073649 | Yes | Yes |
0.27 | 500 | 0.3 | 500 | 0.03 | 0.087892716 | No | No |
0.27 | 500 | 0.33 | 500 | 0.06 | 0.089082901 | No | No |
0.27 | 500 | 0.52 | 500 | 0.25 | 0.09206834 | Yes | Yes |
0.3 | 500 | 0.33 | 500 | 0.03 | 0.090446413 | No | No |
0.3 | 500 | 0.52 | 500 | 0.22 | 0.093388273 | Yes | Yes |
0.33 | 500 | 0.52 | 500 | 0.19 | 0.094509276 | Yes | Yes |