08.06.2015 Views

PCM-2 Manual.pdf - Voss Associates

PCM-2 Manual.pdf - Voss Associates

PCM-2 Manual.pdf - Voss Associates

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Chi-squared Distribution<br />

0.995<br />

1.5<br />

···························r················7:··:::-···-l.;.;,···;,;,;··:.:.:··.:.:··.:.:···:··:.:.:··.:··:··:.:.···~··~··fr·=·:-::··~··~··~ ..=.::.=.:::: ..~..:..:..~.="' ••:.J====:.: ..:..:.:.:.: ..~..~..~.'j.j<br />

~::::<br />

0.990<br />

0.900<br />

............,<br />

. ~ ,<br />

1 ~ ~ ~ .<br />

0.500<br />

~--l ;<br />

0.750<br />

0.250<br />

'" u<br />

0.5<br />

5 10 15<br />

Degrees of Freedom<br />

20 25 30<br />

Figure 5, Chi-squared Distribution<br />

Table 3 is entered at the row that corresponds to the number of degrees of freedom (n - 1).<br />

Associated with the closest x 2 /v found in that row is a probability listed at the top of the column.<br />

That value is the probability that x2/V would be less than or equal to the number found in the table.<br />

Very large or very small probabilities indicate poor fit. When considered in terms of observed results<br />

divided by expected results, intuitively, an ideal experiment would produce a ratio of 1. A 50 %<br />

probability is equivalent to saying that in a random sampling x 2 /v is just as likely to be above as<br />

below the number indicated, and therefore indicative of a good fit of real data to ideal.<br />

NOTE: The fact that the degrees of freedom is one less than the number of data points can cause<br />

some consternation to students of statistics. To clarify the relationship between "n" and "v", take the<br />

case of n=2, the smallest value of n from which an average can be computed. Knowing the average<br />

and the value of only one of the two data points, it is possible to derive the second data point. Also,<br />

the two data points are equidistant from their average. In like fashion, if the average of n data points<br />

is known, then the value of anyone of those data points can be derived from the values of the<br />

remaining (n - 1) data points. Any value could be assigned to any of the (n - 1) points, but the n tl1e<br />

<strong>PCM</strong>2.MAr\

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

Saved successfully!

Ooh no, something went wrong!