Data Analysis with Microsoft Excel: Updated for Office 2007

254 Fundamentals of Statistics

5 Click the OK button to generate the output from the Sign test. See

Figure 6-17.

median

difference

Figure 6-17
Sign test
analysis of
the Diff data

95% confidence

interval

parameter values of the

hypothesis test

number of values value

<, =, and > 0

p

Even under the Sign test, we reject the null hypothesis and accept the

alternative, concluding that the percentage of women in the workforce has

signifi cantly increased. The p value for the Sign test is 0.007. We can also

construct a 95% confi dence interval with the Sign test, but because of the

nature of the test, we can only approximate the confi dence interval at this

level. The output in Figure 6-17 shows the approximate 95% confi dence in-

terval of the change in the percentage of women in the workforce to be (0.00,

0.66). We can also fi nd exact confi dence intervals under the Sign test that

are closest to 95% without going over or going under 95%.

To fi nd the exact confi dence intervals under the Sign test:

1 Click cell D5 and type 21.

The output changes to give you the exact confi dence interval that is

at most 95%. In this case that is a 93.6% confi dence interval, and it

ranges from 0.00 to 0.060. Note that you cannot change the output if

you are creating static output.

2 Click cell D5 and type 1.

The output changes and displays the exact confi dence interval that is at

least 95%. Excel displays the 98.1% confi dence interval: (0.00, 0.80).

You’ve completed your research with the Labor Analysis workbook.

You’ve found that there is suffi cient evidence in this sample of 19 cities to