was used. This coefficient is simply the standard deviation divided by
the mean and is particularly useful in defining the degree of spread of
samples with widely different means (Wine, 1964).

Koch and Link (1971b) present a technique for comparing two coefficients

They suggest that
of variation for data sets with large sample sizes.
the coefficient of variation is approximately normally distributed.
They present a test statistic for determining whether two coefficients
of variation can be considered to be drawn from populations with the
This same test statistic was used, but to
same population coefficient.
accommodate the small sample size, the more general t distribution was
However, inequality in sample variances causes the test statistic
used.

to no longer follow the student t distribution; therefore, the confidence

level was adjusted following a procedure given in Snedecor and Cochran

(1976; pp. 115-116).

It is not known whether these modifications of the

Conclusions resulting from the procedure
test are appropriate or valid.
given about should be considered to be tentative and suggestive and are
The computation equation used was:
certainly not definitive.

C, - Cy
(=

2

2
(1+2C, )

(df, /N,)

2df,

+

Cc,

2

2
(142C, )

a | 1/2
2df,

Where t“ is the test statistic and the level of significance is approximated by using the t table and calculations outlined in Snedecor and

Cochran (1976; pp. 115-116).

C, and Cy are the coefficients of variation

for sample groups one and two (i.e., the top l-cm and 5-cm sampling,
respectively); df, and df are the number of degrees of freedom associated with variance calculation for sample groups one and two; and N)
and No are the number of observations in each group.
When t~ exceeds
the critical value, the significance between the two coefficients is
denoted at some preselected confidence level and the ruling hypothesis
is rejected and the altemate hypothesis is accepted.
Site Il

The sampling coefficient of variation for the l-cm samples was 0.688 as

compared to 0.471 for the 5-cm samples.
The ruling hypothesis is that
the coefficient of variation for the l-cm samples is the same as that
for the 5-cm samples at Site I.
The alternate hypothesis is that the
coefficient of variation for the l-cm samples is not the same as that
for the 5-cm samples.
The t” statistic for comparing these two coefficients of variation is 1.53 with four degrees of freedom.
The critical
value at the 95 percent confidence level is 3.18.
Therefore, the ruling

hypothesis is not rejected.
Site IT

The sampling coefficient of variation for the l-cm samples was 0.177 as
compared to 0.265 for the 5-cm samples.
The ruling hypothesis is that

677

Select target paragraph3