|
 |
|
 Types
of Data (Text Version)
| Laboratory data |
Conditions can be controlled so that variables can be measured
or evaluated one at a time. Responses are usually less variable
and smaller differences are easier to detect. On the flip side,
too many controls can lead to inaccuracy because some of the
variables that exist in the field are eliminated. |
| Data from field studies |
Field experiments or field surveys, as their name suggests,
convey real interactions. Because of their nature, they cannot
be controlled. Compared to laboratory studies or theoretical
models, field surveys usually represent exposures and effects
(including secondary effects) more accurately, so they are most
useful forField experiments or field surveys, as their name
suggests, convey real interactions. Because of their nature,
they cannot be controlled. Compared to laboratory studies or
theoretical models, field surveys usually represent exposures
and effects (including secondary effects) more accurately, so
they are most useful for:
- Linking stressors and effects (as long as stressor and
effect levels are measured concurrently).
- Assessments of multiple stressors or where site-specific
factors significantly influence exposure.
- Larger geographical scales and higher levels of biological
organization.Bear in mind that field survey data are not
always necessary or feasible to collect for screening level
or prospective assessments. Also, because treatments may
not be randomly applied or replicated, classical statistical
methods need to be applied with caution. linking stressors
and effects (as long as stressor and effect levels are measured
concurrently).
- Assessments of multiple stressors or where site-specific
factors significantly influence exposure.
- Larger geographical scales and higher levels of biological
organization.
Bear in mind that field survey data are not always necessary
or feasible to collect for screening level or prospective
assessments. Also, because treatments may not be randomly
applied or replicated, classical statistical methods need
to be applied with caution.
|
| Modeling data |
Modeling outputs are estimates. They simplify reality, so
it's important to evaluate any data you use to build the model
so this simplification comes out as accurate as possible. Models
are particularly useful when it's impossible to make direct
measurements. For example, models can:
- Determine the concentration of air contaminants downwind
of an industrial facility if (1) the facility is not yet
operational or (2) the facility is operational but no sampling
equipment is located at downwind (or downstream) locations.
- Predict the effects of a chemical that has yet to be manufactured.
|
| Analogous data |
When assessors can't generate data specifically for a particular
risk assessment, they must sometimes rely on analogous data
from previous studies-studies performed in similar environments,
on similar organisms, or with a similar chemical. Analogous
data are particularly useful for analyzing a stressor's effect
prior to its release into the environment. One example of this
is the toxicity of a newly manufactured (i.e., unstudied) chemical.
Use of analogous data without knowledge of the underlying processes
may substantially increase the uncertainty in a risk assessment. |
|
|
