10.0      DATA REDUCTION, VALIDATION AND REPORTING

 

10.1      DATA REDUCTION

 

      Data reduction includes all activities that convert instrument/computer responses into reportable results.  These activities may involve mathematical calculations, compound identification and summary statistics.  The final results may be obtained in two ways:

      1.   Direct readings from the instrument; or

      2.      Calculations based on instrument output, readings or responses.

 

      The initial data reduction is the responsibility of the analyst or field technician who operates the analytical instrument.  In addition to the general duties specified below, additional responsibilities for manual and computer related data reduction have been specified.

      1.      Calculate spike recoveries and precision for duplicates; 

      2.   Identify quality control data (blank, spikes, duplicates, etc.) for review by quality assurance officer;

      3.   Assure accurate transcription of sample identification numbers on all records;

 

10.1.1               Manual Data Reduction

 

      1.   If applicable, assure that all readings or output are precisely measured and noted on strip charts;

      2.   Select appropriate formulae for calculating final results;

      3.   Enter the formulae and at least one complete sample calculation on the strip chart or in the notebook;

      4.   Assure that all data are accurately transcribed into notebooks, forms or spreadsheets;

      5.   Enter all manual calculations into notebook or data records;

      6.   Check raw data entries with final computer output to assure accurate initial data entry;

      7.   Record appropriate and accurate information concerning sample identification, operating conditions, etc.

 

      If raw data is entered into a computer program or spreadsheet for data reduction, the organization must be aware of and have on file a record of the mathematical formulae that are being used by the computer.  If such information is not available, the organization shall verify the formula by manual calculations and maintain a record of the verification process.

 

      All raw data output (strip charts, tabular printouts, etc.) must be retained as a part of the records.  These records at a minimum must be identified with the following information:  Date of run; sample ID numbers; analyst or operator; type of analysis (nitrate, metals, etc.).  In addition, the following information must be maintained:  instrument operating conditions (if applicable); detector and column types; instrument configuration; etc.  The latter information may be kept in cross referenced records or may be entered on the various output records.

 

10.1.2               Computer/Integrator Reduction

 

      1.   Assure that all data to be used in final calculations are entered accurately:  sample weights or volumes; final extract volumes; dry weight factors; dilution factors; surrogate standard concentrations, etc.;

      2.   Properly interpret the computer output in terms of properly identified components, positive or negative identifications, and appropriate confirmatory measures;

      3.   Record appropriate and accurate information concerning sample identification, operating conditions, etc.;

      4.      Calculate surrogate recoveries and internal standard responses (if applicable);

      5.   GC and GC/MS analyses should be checked to verify that target components are within acceptable retention time windows and that additional confirmation (if needed) is initiated.

 

      Many analytical instruments are interfaced with computers or integrators that automatically evaluate, identify and calculate final values.  The results are printed in combinations of graphic (ex. chromatograms) and tabular forms.  As with manual data reduction, the organization must be aware and should on file a record of the mathematical formulae or algorithms that are being used by the computer.  If the information is not available, the organization shall maintain records which demonstrate that the software is providing the expected results (e.g. check sample or check standard data is acceptable).

 

      Typically computer data files are identified by a queue number or a data file number.  In such cases, the organization must maintain a cross reference index or log to identify the computer data files with sample ID numbers.  Additional information that should be entered into the data file records are:  date of run, analysis type, and analyst initials.  Cross referenced auxiliary records are required to identify instrument operating conditions (if applicable); detector and column types; instrument configuration; etc.

 

10.1.3               Formulae and Calculations

 

      The final results of each test shall be calculated by the formula specified in the analytical method that is being used.

      The final result should be rounded off to an appropriate number of significant figures (typically 2 significant figures).  If the digit 6,7,8 or 9 is dropped, increase the preceding digit by one unit; if the digit 0,1,2,3, or 4 is dropped, do not alter the preceding digit.  If the digit to be dropped is 5, round off the preceding digit to the nearest even number:  2.25 becomes 2.2 and 2.35 becomes 2.4.

 

      As a general rule the results should be converted to the reporting units presented on Table 10.1.  Other reporting conventions (i.e. wet weight instead of dry weight) should be clearly identified on the final reports with appropriate justification.

 

      Note:  If components of interest are detected in any quality control blank(e.g. method blanks, digestion blanks, etc.), the blank concentration must be reported.  The blank concentration shall not be subtracted from any associated sample data.

 

10.2      DATA VALIDATION

 

10.2.1               Data Integrity

 

      Data integrity involves checking all field and laboratory data entries and calculations for errors and mistakes.  It also involves reviewing all documentation to assure that sample ID numbers are correct, and that the tests have been performed within appropriate holding times.  Data integrity does not include assessment of quality control measures.

 

      All data integrity checks should be performed by an individual (preferably a supervisor) who was not involved in the original data reduction process.

      The following records and documentation must be checked:

      1.   A minimum 10 - 20% of raw data entries for transcription accuracy;

      2.   A minimum 10 - 20% of all calculations randomly checked for mathematical errors including dilution factors, final volumes, dry weight factors and sample volumes or weights.  Note:  field calculations for well and purge volume are excluded;

      3.   Verify acceptability of initial and/or continuing calibration data;

 

Table 10.1

DATA REPORTING UNITS

 

TEST NAME OR COMPONENTS	REPORTING UNITS
	WATER	SED	FISH	WASTE
Metals except:	ug/L	mg/kg	mg/kg	mg/kg
Reports for potable(drinking water),
Calcium, magnesium, sodium, potassium	mg/L	mg/kg	mg/kg	mg/kg
Purgeable organic components (VOCs and VOAs)	ug/L	ug/kg	mg/kg	mg/kg
Extractable organic components Including pesticides and herbicides except:	ug/L	ug/kg	mg/kg	mg/kg
Dioxin/Furan Scan and Dibenzo dioxins and dibenzofurans	ng/L	ng/kg	ng/kg	ng/kg
Odor (60 Degree C)	TON
Odor (Room Temp)	TON
pH (Laboratory)	pHUN
Color (True-PTCO)	CU
Color (Apparent-PTCO)	CU
Conductivity	uMHOS/cm
Corrosivity
Flash Point				DEG F
Hardness (as CaCO3)	mg/L
Settleable Solids	ml/L/h
Total Solids	mg/L	mg/kg		mg/kg
Volatile Total Solids	mg/L	mg/kg		mg/kg
Total Suspended Solids	mg/L	mg/kg	mg/kg
Volatile Total Suspended Solids	mg/L	mg/kg		mg/kg
Total Dissolved Solids (180 Degree C)	mg/L
Volatile Total Dissolved Solids	mg/L	mg/kg	mg/kg
Toxicity (EP and TCLP)	mg/L
Turbidity	NTU
Radium-226, Total	pc/L	pc/g
Radium-228, Total	pc/L	pc/g
Radium-226, Diss	pc/L	pc/g
Radium-228, Diss	pc/L	pc/g
Gross Alpha, Total	pc/L	pc/g
Gross Beta, Total	pc/L	pc/g
Acidity	mg/L	mg/kg	mg/kg	mg/kg
Alkalinity, Bicarbonate (as CaCO3)	mg/L
Alkalinity, Carbonate (as CaCO3)	mg/L
Alkalinity, Total (as CaC03)	mg/L	mg/kg	mg/kg	mg/kg
Bicarbonate (as HCO3 ION)	mg/L	mg/kg
Carbonate (as CO3 ION)	mg/L	mg/kg
Ammonia (an N)	mg/L	mg/kg	mg/kg	mg/kg
Ammonia, Dissolved (as N)	mg/L

Table 10.1

DATA REPORTING UNITS, cont.

 

TEST NAME OR COMPONENTS	REPORTING UNITS
	WATER	SED	FISH	WASTE
Ammonia, Unionized (as NH3)	mg/L
Bromide	mg/L
Chloride	mg/L	mg/kg		mg/kg
Chlorine Residual	mg/L	mg/kg	mg/kg	mg/kg
Cyanide	mg/L	mg/kg	mg/kg	mg/kg
Cyanide Amenable to Chlorination	mg/L	mg/kg	mg/kg	mg/kg
Cyanide, Free	mg/L	ug/kg	mg/kg	mg/kg
Dissolved Oxygen (Winkler)	mg/L
Dissolved Oxygen (Electrode)	mg/L
Fluoride	mg/L	mg/kg	mg/kg	mg/kg
Nitrate-Nitrogen	mg/L	mg/kg	mg/kg	mg/kg
Nitrite-Nitrogen	mg/L	mg/kg
Nitrate+Nitrite Nitrogen	mg/L	mg/kg	mg/kg	mg/kg
Nitrate+Nitrite Nitrogen, Dissolved	mg/L
Ortho-Phosphate Phosphorus	mg/L	mg/kg	mg/kg	mg/kg
Silicon (Si)	mg/L	mg/kg	mg/kg	mg/kg
Silica (SiO2)	mg/L	mg/kg	mg/kg	mg/kg
Sulfate	mg/L	mg/kg		mg/kg
Reactive Sulfides (as H2S)	mg/L	mg/kg		mg/kg
Sulfides	mg/L	mg/kg		mg/kg
Sulfite	mg/L
Temperature	Deg C
Total Dissolved Phosphorus	mg/L
Total Kjeldahl Nitrogen	mg/L	mg/kg	mg/kg	mg/kg
Total Kjeldahl Nitrogen, Dissolved	mg/L
Total Phosphorus	mg/L	mg/kg	mg/kg	mg/kg
Bio-Chemical Oxygen Demand, 5 Day	mg/L	mg/kg	mg/kg
Bio-Chemical Oxygen Demand, 5 Day, Dissolved	mg/L	mg/kg	mg/kg
Bio-Chemical Oxygen Demand, 20 Day	mg/L	mg/kg	mg/kg
Bio-Chemical Oxygen Demand, 60 Day	mg/L	mg/kg	mg/kg
BOD, Carbonaceous, 5 Day	mg/L
Chemical Oxygen Demand	mg/L	mg/kg	mg/kg
Chemical Oxygen Demand, Dissolved	mg/L
Linear Alkyl Sulfonate	mg/L	mg/kg	mg/kg	mg/kg
Oil and Grease	mg/L	mg/kg	mg/kg	mg/kg
Phenols (4AAP)	ug/L	mg/kg	mg/kg	mg/kg
Total Organic Carbon	mg/L	mg/kg	mg/kg	mg/kg
Total Organic Carbon, Dissolved	mg/L
Purgeable Organic Carbon	mg/L
Total Organic Halogen	ug/L	ug/kg	mg/kg

      4.   Sample preparation logs and instrument or analytical logs to assure that samples were prepared and analyzed within prescribed holding times;

      5.   Cross check all records for completeness and for transcription errors associated with the sample number;

      6.   Verify all chain of custody records (if applicable) for completeness and acceptability;

 

10.2.2               Data Validation

 

      Data validation is accomplished through a series of checks and reviews that are intended to assure that the reported results are of a verifiable and acceptable quality.

 

      A majority of these tasks should be performed by the Quality Assurance Officer, but some may be delegated to another individual (ex. supervisor) who was not actively involved with generating the data.

      1.   Verify that all quality control blanks meet criteria;

      2.   Review all other quality control data (spikes, duplicates, quality control check standards, quality control check samples, etc.) for acceptability;

      3.   Review all surrogate and standard additions spike recoveries and internal standard responses for acceptability;

      4.   Identify any sample set or data that are unacceptable and initiate appropriate corrective action measures;

      5.   Assign data qualifiers (if needed) to reported values;

      6.   Verify mass spectral interpretation (if applicable) and/or component identification;

      7.   Assign data qualifiers to all applicable data (see Table 10.2).  Note:  the reported value always precedes the data qualifier code.

 

10.3      DATA REPORTING AND OVERALL PROJECT VALIDATION

 

10.3.1               Laboratory Data Reports

 

      The final reports from the laboratory may be generated in several different ways:

      1.   Hand written report forms;

      2.      Manually typed reports and narrative;

      3.      Computer generated reports;

      4.   Any combination of the above methods.

 

      All parties who are involved with the data review and validation process are responsible for providing data entry operators or clerical personnel with accurate records for transcription.  If data are automatically reported through a LIMS system, the final reviewer must assure that the appropriate commands have been input to release the data for final reports.

 

10.3.2               Engineering Project Reports

 

      The final reports from an engineering firm or the organization responsible for a project involves assimilating and presenting data from both the laboratory and field.  These reports may also include narratives on site history, an analysis of current findings; and conclusions and/or recommendations on further project work.

 

10.3.3               Project Validation

      Project validation is the process by which all project data is reviewed prior to reporting the data to the client.  This task is normally assigned to the project manager but may be performed by an individual who is responsible for overall management operations.

 

Table 10.2

DATA QUALIFIER CODES

 

A                      Value reported is the mean (average) of two or more determinations.  This code shall be used if the results of two or more discrete and separate samples are averaged.  These samples shall have been processed and analyzed (e.g. laboratory replicate samples, field duplicates, etc.) independently.  Do not use this code of the data are the result of replicate analyses on the same sample aliquot, extract or digestate.  Under most conditions, replicate values shall be reported as individual analyses.

 

B                      Results based upon colony counts outside the acceptable range.  This code applies to microbiological tests and specifically to membrane filter colony counts.  The code is to be used if the colony count is generated from a plate in which the total number of coliform colonies EXCEEDS the method indicated ideal ranges which are:

                                    Total Coliforms:  20 - 80 colonies

                                    Fecal Coliforms:  20 - 60 colonies

 

F                      When reporting species: F indicates female sex.

 

H                      Value based on field kit determination; results may not be accurate.  This code shall be used if a field screening test (i.e. field gas chromatograph data, immunoassay, vendor-supplied field kit, etc.) was used to generate the value and the field kit or method has not been recognized by the Department as equivalent to laboratory methods.

 

J                       Estimated value; value not accurate.  This code shall be used in the following instances:

                                          1.   surrogate recovery limits have been exceeded;

                                          2.   no known quality control criteria exists for the component;

                                          3.   the reported value failed to meet the established quality control criteria for either precision or accuracy;

                                          4.   the sample matrix interfered with the ability to make any accurate determination; or

                                          5.   the data is questionable because of improper laboratory or field protocols (e.g. composite sample was collected instead of a grab sample).

                                    Note:    1.         A "J" value shall be accompanied by justification for its use.

                                                      2.   A "J" value shall not be used if another code applies (ex. K, L, M, T, V, Y, PQL)

 

K                      Off-scale low.  Actual value is known to be less than the value given.  This code shall be used if:

                                          1.   The value is less than the lowest calibration standard AND the calibration curve is known to be non-linear; or

                                          2.   The value is known to be less than the reported value based on sample size, dilution or some other variable.

                        This code SHALL NOT be used to report values that are less than the laboratory practical quantitation limit or laboratory method detection limit.

 

L                      Off-scale high.  Actual value is known to be greater than value given.  To be used when the concentration of the analyte is above the acceptable level for quantitation (exceeds the linear range or highest calibration standard) AND the calibration curve is known to exhibit a negative deflection.

Table 10.2

DATA QUALIFIER CODES, cont.

 

 

M                     When reporting chemical analyses:  presence of material is verified but not quantified.  The reported value shall be the laboratory practical quantitation limit.  This code shall be used if the level is too low to permit accurate quantification, but the estimated concentration is GREATER THAN the laboratory method detection limit. If the value is less than the method detection limit, use "T", below.  When reporting Oxygen Reduction Potential or Temperature:  indicates a negative value  When reporting Species:  indicates male sex.

 

N                      Presumptive evidence of presence of material.  This qualifier shall be if:

                                          1.   the component has been tentatively identified based on mass spectral library search;

                                          2.   if the presence of the analyte is indicated but there is evidence of possible interferences; or

                                          3.   there is an indication that the analyte is present, but quality control requirements for confirmation were not met (i.e. presence of analyte was not confirmed by alternate procedures).

 

O                      Sampled, but analysis lost or not performed.  Note:  if reporting data to STORET, a numerical value must be entered.  Such values are not meaningful and shall not be used.

 

Q                      Sample held beyond the accepted holding time.  This code shall be used if the value is derived from a sample that was prepared and/or analyzed AFTER the approved holding time restrictions for sample preparation and analysis.

 

T                      Value reported is less than the laboratory method detection limit.  The value is reported for informational purposes only and SHALL NOT be used in statistical analysis.

 

U                      Indicates that the compound was analyzed for but not detected.  This shall be used to indicate that the specified component WAS NOT detected.  The value associated with the qualifier shall be the laboratory method detection limit.  Unless requested by the client, less than the method detection limit values shall not be reported (see "T" above).

 

V                      Indicates that the analyte was detected in both the sample and the associated method blank.  Note:  the value in the blank shall not be subtracted from associated samples.

 

Y                      The laboratory analysis was from an unpreserved or improperly preserved sample.  The data may not be accurate.

 

Z                      Too many colonies were present (TNTC), the numeric value represents the filtration volume.

Table 10.2

DATA QUALIFIER CODES, cont.

 

 

PQL                  The reported value is between the laboratory method detection limit and the laboratory practical quantitation limit.

 

REJ                  Data is rejected and should not be used.  Some or all of the quality control data for the analyte were outside criteria, and the presence or absence of the analyte cannot be determined from the data.

 

NAI                   Not analyzed due to interference.

 

 

                        If more than one code applies, and the data is to be entered into STORET, only one code shall be reported.  The code shall be selected based on the following hierarchy:

                                    REJ

                                    NAI, O

                                    Y

                                    V

                                    H

                                    J

                                    B, K, L, M, PQL, T, Z

                                    A

 

      The following codes deal with certain aspects of field activities.  The codes shall be used IF the laboratory has knowledge of the specific sampling event.  The codes shall be added by the organization collecting the sample, if they apply:

 

D                      Measurement was made in the field (i.e. in situ).  This applies to any value (ex pH, specific conductance, etc.) that was obtained under field conditions using approved analytical methods.  Note:  when data is to be entered into STORET, and the parameter code specifies a field measurement (e.g. "Field pH"), this code is not required.

 

E                      Indicates that extra samples were taken at composite stations.

 

R                      Significant rain in the past 48 hours.  This code shall be used when the rainfall might contribute to a lower than normal value.

 

HIS                   Data deviates from historically established concentration ranges.