Laserfiche WebLink
Mid-Continent Report; Nov. 15, 1993 <br /> Steve Renner/Harry Posey <br /> Page 5 <br /> evaluating the data. Rather, the test was to see whether the positive outliers bear any <br /> relationship to time of sampling. <br /> 5b. Conductivity and TDS Outliers. Table 1 shows the data and the variation relative to the <br /> standard deviation of the population at station #20. This evaluation of variance relative to the <br /> mean and standard deviation is different from that shown on the graphs. The graphs highlight <br /> the outliers at any specific value of flow whereas Table 1 highlights the outliers relative to the <br /> entire population, regardless of the rate of flow. <br /> TDS outliers normalized to flow range from about 430 to 1312; conductivity outliers normalized <br /> to flow range from 600 to 1900. However, TDS outliers relative to the mean of the entire data <br /> set range from 568 to 1312, and conductivity outliers relative to the mean of the entire data set <br /> range from 900 to 1900. <br /> Outliers Range <br /> low high <br /> TDS relative to flow 430 1312 <br /> TDS relative to population 568 1312 <br /> Cond relative to flow 600 1900 <br /> Cond relatiove to population 900 1900 <br /> This could have been an important distinction because the company reported their average yearly <br /> values relative to the entire population even though the entire population does not account for <br /> the variation in any of the parameters relative to the flow. However, even though there are <br /> large differences in the low range of the values, these were caused by only one sample. Even <br /> by comparing results with the mean of all values, as was done by the company, anomalous <br /> results could have been discerned in nearly all cases without the different approach used in this <br /> report. <br />