The AOL disks used in the study were obtained by randomly grabbing fifty disks from a box of unopened AOL disks in the media center.  The Imation floppies were obtained by purchasing two boxes of twenty-five each from a local computer store.  There were fifty disks of each type, for a total of one-hundred disks.  The population of interest is all AOL disks and all Imation disks.

Data:
 

	AOL	Imation
Good	50	49
Bad	0	1

Descriptive Statistics:
 

	AOL	Imation
Proportion Bad	0	.02

Graphs:
 

AOL Disks

Imation Disks

Blue/0 mean a good disk, red/1 mean a bad disk.

Significance Tests:

Chi-Square test:

        Ho:  the proportion of bad disks is independent of the manufacturer
        Ha:  the proportion of bad disks is not independent of the manufacturer

        assume that the proportion of bad disks of each type will be equal, so pi = .5

        each expected = .5 * 1
 

	AOL	Imation
Disks bad	0	1
Expected	.5	.5

        Since the Chi square test requires that the expected count be greater than 5, a Chi square test cannot be performed on the data.

95% Confidence interval:

        0-1 +- 1.96*sqrt(0(1-0)/50 + .02(1-.02)/50)
        (-1.0388, -.9612)
        95% of all samples will have differences of proportions which will fall within this range.

Z test:
        P1(AOL) = 0
        P2(Imation) = .02

        Pc = .01

        z = -1.005
        P-value = .1587
        Ho: No difference between AOL and Imation disks
        Ha: There is a difference between AOL and Imation disks
        Since P-value is greater than any reasonable alpha, there is no
        difference between AOL and Imation disks.
 

[Previous][Home][Next]