Laserfiche WebLink
C Data -- The raw data and calculated results for all QC and field <br /> samples and standards should be maintained in the manner described in the <br /> preceding paragraph. Documentation should allow correlation of sample <br /> results with associated QC data. Documentation should also include the <br /> source and lot numbers of standards for traceability. QC samples include, <br /> but are not limited to, control samples, method blanks, matrix spikes, and <br /> matrix spike duplicates. <br /> Correspondence -- Project correspondence can provide evidence supporting <br /> technical interpretations. Correspondence pertinent to the project should <br /> be kept and placed in the project files. <br /> Deviations -- All deviations from procedural and planning documents should <br /> be recorded in laboratory notebooks. Deviations from QAPjPs should be <br /> reviewed and approved by the authorized personnel who performed the <br /> original technical review or by their designees. <br /> Final Report -- A copy of any report issued and any supporting documenta- <br /> tion should be retained. <br /> 5.0 DEFINITIONS <br /> The following terms are defined for use in this document: <br /> ACCURACY The closeness of agreement between an observed value and <br /> an accepted reference value. When applied to a set of <br /> observed values, accuracy will be a combination of a <br /> random component and of a common systematic error (or <br /> bias) component. <br /> BATCH: A group of samples which behave similarly with respect to <br /> the sampling or the testing procedures being employed and <br /> which are processed as a unit (see Section 3.4. 1 for field <br /> samples and Section 4.4.3 for laboratory samples) . For QC <br /> purposes, if the number of samples in a group is greater <br /> than 20, then each group of 20 samples or less will all be <br /> handled as a separate batch. <br /> BIAS: The deviation due to matrix effects of the measured value <br /> (x, - x„) from a known spiked amount. Bias can be assessed <br /> by comparing a measured value to an accepted reference <br /> value in a sample of known concentration or by determining <br /> the recovery of a known amount of contaminant spiked into <br /> a sample (matrix spike) . Thus, the bias (B) due to matrix <br /> effects based on a matrix spike is calculated as: <br /> B = (x, - xu ) K <br /> where: <br /> 4. <br /> ONE 23 Revision 1 <br /> July 1992 <br />