Basic HTML Version
25
from similar analyses. Only comparable data sets can be readily combined. Table 5 presents nine
indicators of comparability, and questions are considered related to each.
Table 5. Indicators of comparability.
Indication of Comparability
Related Questions
Samples within data sets
should be selected in a similar
manner
Sample design: Were the samples selected in a similar manner? Are they equally
representative of the population of interest? If samples in one data set were selected
using a judgmental sample design, and another data set is based on a statistical design,
then combining these data may not be appropriate for some uses.
Data should be
temporally and
spatially consistent
Sample collection dates: Were samples collected in the same sampling event? Are there
temporal factors such as seasonality or holding times that could directly affect
interpretation of the data?
Sample location: Were the samples taken from the same area? Are they representative
of the same population spatially? If they are from different areas, how are they expected
to be similar? How are they expected to differ?
Matrix: Were the samples from the same matrix? This relates to how the samples were
collected, location of the samples, and when the samples were collected. If matrices are
different, are they expected to be related in some way?
Data sets should contain the
same set of variables of
interest
Variables of Interest: Which variables are of interest and are necessary for grouping or
analyzing the data? Were these variables reported for all data sets?
Units in which these variables
were measured should be
convertible to a common
metric
Units: Units should be reported for all data sets. Are the units all convertible to a
common metric? For example, some results may be reported in wet weight and some in
dry weight, which are not directly comparable without additional information
Field collection methods
should be
similar
Field methods: What instrument was used and which procedure was followed? Were
single or composited samples collected?
Sample handling: Some samples require special handling such as preservatives or
special containers. Differences in sample handling may cause variations in the results,
which may affect comparability. Were the samples filtered or unfiltered? Are there
chain-of-custody forms available for all samples?
Similar sample
Preparation methods should be
used
Laboratory: Was the same laboratory used for all analyses? The use of routine methods
and procedures simplify the issues of comparability because the same standards should
be met. In addition, this will increase confidence in the comparability of methods used.
Sample preparation: Was the same sample preparation used for all samples? If not, are
the sample preparation methods comparable?
Similar procedures
and quality assurance should
be used to collect and analyze
samples for all data sets
Analytical method: Was the same analytical method used for all samples? If not, are
any of the analytical methods comparable? The use of routine methods simplifies the
determination of comparability because all laboratories used the same standardized
procedures and reporting parameters. However, when reviewing the analytical
methods, consideration must also be given to options that may be available within the
method. Although the analytical method may be the same, options such as matrix or
concentration level will affect results reported.
Analytical method options: If the analytical methods are comparable, were the same
options within each method chosen? The options available within each method must
also be checked because the same analytical method using different options may
produce very different results.
Measuring devices
used for both data
sets should have
approximately similar
detection levels
Detection or quantitation level: Are non-detects generally reported at the same level?
Are the detection or quantitation levels acceptable for use in decision making?
Combining data sets having different detection or quantitation levels leads to difficulties
in analytical interpretations.
Quality control of data entry, storage, transfer, and retrieval: Were results reported into
the database in a consistent manner? Have all data sets been checked for completeness?