USGS Inorganic Blind Sample Project (IBSP)

Monitoring and Evaluating Laboratory Analytical Quality

(Based on Fact Sheet FS-136-97)

The U.S. Geological Survey (USGS) collects and disseminates information about the Nation's water resources. Surface- and ground-water samples are collected and sent to USGS laboratories for chemical analyses. The laboratories identify and quantify the constituents in the water samples. Random and systematic errors occur during sample handling, chemical analysis, and data processing. Although all errors cannot be eliminated from measurements, the magnitude of this uncertainty can be estimated and tracked over time. Since 1981, the USGS has operated an independent, external, quality-assurance project called the Inorganic Blind Sample Project (IBSP). The purpose of the IBSP is to monitor and evaluate the quality of laboratory analytical results through the use of double-blind quality-control (QC) samples. The information provided by the IBSP assists the laboratories in detecting and correcting problems in the analytical procedures. The information also can aid laboratory users in estimating the extent that laboratory errors contribute to the overall errors in their environmental data.

A "double-blind sample" is a QC sample submitted for analysis for which the identity of the sample as well as its concentration levels are unknown to the analyst. Double-blind QC samples containing selected inorganic and nutrient constituents at various levels of concentrations are prepared and disguised as routine environmental samples. The IBSP submits these samples to the National Water Quality Laboratory (NWQL) at the Denver Federal Center in Lakewood, CO. The NWQL provides analytical services for all USGS national programs and many local or regional projects conducted by the USGS. Standard reference samples (Farrar and Long, 1997) are used to make the QC samples for the IBSP. They are used diluted with deionized water, undiluted, and mixed in varying proportions with other standard reference samples. This sample-mixing procedure produces a large number of unique samples available for quality-assurance purposes. The standard reference samples are usually natural matrix samples collected from different sources such as snowmelt, streams, and ground water.

The IBSP samples are made to appear as much like environmental samples as possible and are subjected to identical laboratory handling, processing, and analytical procedures. After the laboratories analyze the samples, IBSP personnel compile and review the analytical results. The resulting data are stored in the USGS National Water Information System data base.

The laboratories are evaluated by how closely their analytical results approximate the most probable value of the blind QC samples. Most probable values are the median concentrations reported for each constituent in the round-robin evaluation of the standard reference sample. The assessment of whether an analytical result is acceptable is based on the number of standard deviations that the measured concentration differs from the most probable value. Analytical results that are within two standard deviations of the most probable value are considered acceptable. Analytical errors fall into two major categories: bias and variability. Bias is systematic error that causes consistently positive or negative deviation in the results from the most probable value. Variability is random error that affects the ability to reproduce results. Repeated measurements of the IBSP samples over time provide estimates of both systematic bias and random variability in the laboratory analytical procedures. The IBSP uses a variety of graphical and statistical tools to evaluate laboratory analyses of the blind QC samples. These tools include control charts, relative standard deviation charts, box plots of error distributions, Wilcoxon signed-rank test for bias, binomial-probability distribution test for variability, and statistical summaries.

Control charts are produced for each analytical method as a review of the laboratory performance. They are a graphical display of the analytical deviation from the most probable value with respect to time. Control charts show if analytical results are within the expected control limits. The IBSP sets control limits at ±2 standard deviations of the most probable value. These charts can illustrate a systematic or sudden shift in bias or variability. The control chart (fig. 1) for whole water-recoverable iron analyses indicates a change from a positive bias in February 1996 to a negative bias in September 1996.

Another graphical tool used by the IBSP is the relative standard deviation chart, which displays analytical variability with respect to concentration. These charts allow a data reviewer to estimate analytical variability at a given concentration. The relative standard deviation chart (fig. 2) for dissolved zinc analyses is typical in showing that variability decreases with increasing concentrations.

The IBSP maintains an interactive online computer program and data base (QADATA Application) for the retrieval and assessment of blind QC sample analytical results. Currently, the QADATA Application contains more than 235,000 QC analyses for inorganic and nutrient constituents, and physical-property measurements dating from October 1984 to present. New analytical data released from the laboratories are added twice-weekly to the QADATA Application. Data retrievals from the QADATA Application can be customized to document the laboratories' analytical bias and variability relative to the time period, analytical procedures, and concentration ranges of individual water-quality projects or programs. The QADATA Application is available through the USGS computing environment. For instructions on accessing and using the QADATA Application contact Ted Struzeski at 303.236.1872 (struzesk@usgs.gov).

Farrar, J.W., and Long, H.K., 1997, Report on the U.S. Geological Survey's evaluation program for standard reference samples distributed in September 1996-T-143 (trace constituents), T-145 (trace constituents), M-140 (major constituents), N-51 (nutrient constituents), N-52 (nutrient constituents), P-27 (low ionic strength constituents), and Hg-23 (mercury): U.S. Geological Survey Open-File Report 97-20, 145 p. Table of Project Statistics by Water Year.

For further information, contact:

Inorganic Blind Sample Project Chief (struzesk@usgs.gov) U.S. Geological Survey
Denver Federal Center
Mail Stop 401, Box 25046
Denver, CO 80225
303.236.1872