28.12.2013 Views

preface to fifteenth edition

preface to fifteenth edition

preface to fifteenth edition

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

GENERAL INFORMATION, CONVERSION TABLES, AND MATHEMATICS 2.125<br />

Example 7 A new method for the analysis of iron using pure FeO was replicated with five<br />

samples giving these results (in % Fe): 76.95, 77.02, 76.90, 77.20, and 77.50. Does a systematic<br />

error exist?<br />

From Equation 4, x is 77.11; and from Equation 5, s is 0.24 for 4 degrees of freedom. Because<br />

is not known, the Student t 0.975 (2.78 for 4 degrees of freedom) is used <strong>to</strong> calculate the confidence<br />

interval at the 95% probability level.<br />

ts (2.78) (0.24)<br />

x 77.11 77.11 0.30<br />

pN p5<br />

We used a two-tailed test. Upon rereading the problem, we realize that this was pure FeO whose<br />

iron content was 77.60% so that 77.60 and the confidence interval does not include the known<br />

value. Since the FeO was a standard, a one-tailed test should have been used since only random<br />

values would be expected <strong>to</strong> exceed 77.60%. Now the Student t value of 2.13 (for t 0.05 ) should<br />

have been used, and now the confidence interval becomes 77.11 0.23. A systematic error is<br />

presumed <strong>to</strong>exist.<br />

The t test can be applied <strong>to</strong> differences between pairs of observations. Perhaps only a single pair<br />

can be performed at one time, or possibly one wishes <strong>to</strong> compare two methods using samples of<br />

differing analytical content. It is still necessary that the two methods possess the same inherent<br />

standard deviation. An average difference d calculated, and individual deviations from d are used <strong>to</strong><br />

evaluate the variance of the differences.<br />

Example 8 From the following data do the two methods actually give concordant results?<br />

Sample Method A Method B Difference<br />

1 33.27 33.04 d 1 0.23<br />

2 51.34 50.96 d 2 0.38<br />

3 23.91 23.77 d 3 0.14<br />

4 47.04 46.79 d 4 0.25<br />

d 0.25<br />

p (d d) 2<br />

sd<br />

0.099<br />

N 1<br />

0.25<br />

t p4 1 4.30<br />

0.099<br />

From Table 2.27, t0.975 3.18 (at 95% probability) and t0.995<br />

5.84 (at 99% probability). The<br />

difference between the two methods is probably significant.<br />

If the t-value falls short of the formal significance level, this is not <strong>to</strong> be interpreted as proving<br />

the absence of a systematic error. Perhaps the data were insufficient in precision or in number <strong>to</strong><br />

establish the presence of a constant error. Especially when the calculated value for t is only slightly<br />

short of the tabulated value, some additional data may suffice <strong>to</strong> build up the evidence for a constant<br />

error (or the lack thereof).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!