Question

An occupational safety officer for a large company is conducting a study to investigate back problems in office workers who use a computer for most of the workday. The study will investigate the difference in back problems for workers who stand and workers who sit. A group of 68 volunteers have agreed to participate in the nine-month study. Half the group is randomly assigned to work while standing, and the other half is assigned to work while sitting. At the end of the study, the mean number of back problems between the two groups will be calculated. The officer will use the results to estimate the difference in the mean number of back problems between those who work while standing and those who work while sitting.

Which of the following is an appropriate inference procedure for the study?
A. A one-sample tt-interval for a population mean
B. A one-sample tt-interval for a sample mean
C. A matched pairs tt-interval for a mean difference
D. A two-sample tt-interval for a difference between sample means
E. A two-sample tt-interval for a difference between population means

Answer 1

C. A matched pairs tt-interval for a mean difference

Explanation:

When the observations from two samples are paired either naturally or by design we find the difference between the two observations of each pair. Treating the differences as a random sample from a normal population with mean ud= u1-u2 and unknown standard deviation , we perform one sample t- test on them . This is called paired difference or a paired or a matched t- test.

The test statistic

t= d`/sd/√n

where d`= ∑ di/n

and sd= ∑ (di- d`)²/n-1

with υ= n-1 degrees of freedom.

For example suppose 10 young recruits are given a strenuous physical training programme by Army. Their weights are recorded before they begin and after they complete they complete the training. The two observations obtained for each recruit. i.e. before and after measurements constitute natural pairing.