Table B-II

DOE-SPEC-1142-2001

B.2.

For each Treatment Group calculate the median of the Ln counts (see Table B-II

column 6).

Table B-II

Treatment Group

Ln (well counts)

Median

Day5 controls

7.1066

7.7795

7.4810

6.8533

7.2819

Day5 controls

7.3126

7.3576

7.2513

7.0309

7.2819

Day5 controls

6.8763

7.7253

7.4634

6.5903

7.2819

Day5 Be1

7.4827

7.5443

7.4396

7.5417

7.5122

Day5 Be10

8.1221

8.8847

7.2951

8.0382

8.0801

Day5 Be100

8.1973

8.2039

7.8047

7.3988

8.0010

Day7 controls

8.1931

9.7648

8.2913

8.0533

8.0123

Day7 controls

6.5058

7.1365

7.3112

8.4029

8.0123

Day7 controls

7.9714

8.3366

7.2196

7.0493

8.0123

Day7 Be1

7.4206

7.6898

6.4441

7.1420

7.2813

Day7 Be10

5.7991

6.3936

5.5373

5.5759

5.6875

Day7 Be100

8.1917

8.3975

9.5786

9.6086

8.9880

PHA

11.534

10.697

10.991

10.841

10.916

CONA

11.658

11.553

12.438

11.979

11.819

B.3 For each beryllium concentration, calculate the Ln(SI) by subtracting the median of the

control wells from the median of the Be stimulated wells, e.g., for Day5 Be100 the

Ln(SI) = 8.0010 - 7.2819 = 0.7191 and the SI is exp(0.7191) = 2.05.

B.4 Calculate the standard deviation of the Ln counts (corresponds to CV on original scale). In

EHP96 this is referred to as "phitilde". The median absolute deviation (MAD) estimate

given on page 960 (called here Sm--- the MAD estimate of the standard deviation), is

Sm = 1.48 * median[ absolute(residual) ] * √[n/(n-p)] where:

absolute(residual) = the absolute value of the (Ln well count - the median Ln well count);

n the number of data values is 24, if 12 control wells and 12 exposed wells; and

p the number of parameters is 4, if 1 median for control wells and 3 medians for exposed

wells.

A separate estimate of Sm is calculated for Day 5 and Day 7, since it has been observed that

there is generally more variability on Day 7. Using Day 5 as an example, the residuals are

listed in Table B-III.