Biomedical Engineering Reference
In-Depth Information
The physical description of the act should have suffi-
cient detail to allow new observers to recognize the
response. For example, “bite” may be defined as the
animal's grasping their partner on any body part with
the mouth, and “grimace” as a facial expression where the
lips are drawn back, exposing the teeth. These definitions
allow multiple observers to record an animal's responses
using consistent scoring schemes. If multiple observers are
recording data, it is recommended that each observer
independently record an animal's behavior across the same
time period so that an inter-rater reliability can be
measured. Cohen's Kappa is an index typically used to
correlate independent observations, as it takes into account
the amount of agreement that could be expected to occur
through chance.
Ethograms should also include whether a response is to
be recorded as a frequency per some time period or as
duration over time. Grabs, bites, pushes, and other quick
interactions have little duration and are thus best recorded
as a number of occurrences per time block. Interactions
such as groom, sit next to (proximity), or chase have longer
durations. The frequency of these interactions may be
important, but their lengths and the total time spent in the
interaction may be of equal or greater importance. Deciding
what responses need to be recorded and the type of data to
be collected, frequency or duration, will influence what
observation method you can use. If you intend to look at
responses that are recorded as frequency data, you may be
limited to the all-occurrences (group) scan or focal animal
techniques. The probability of observing a bite at the
moment of an instantaneous scan is small because the
duration of the response is small. If your study involves
responses typically recorded as durations, then instanta-
neous scans could be useful. This may become important if
you can spend only limited time recording data. Focal
animal sampling can give you the best picture of what is
occurring but can be very labor intensive. Instantaneous
scans will miss the quick responses but a lot of data can be
collected quickly on longer duration responses.
Additionally, you may be interested in how your new
group or pair compares with other stable groups or pairs.
Data on stable social groups can be collected at the same
time as it is collected for the study group or it can be
collected independently. Data collected independently can
serve as a population norm to compare with the data
collected as part of the social group formation with two
caveats. First, for the data to be directly comparable, it
should be collected using the same sampling techniques.
Second, husbandry, housing, and other facility factors that
might influence social behavior need to be consistent across
the two datasets. Since almost all facilities change
husbandry procedures over time, it would be best to
conduct routine observations of stable groups to ensure that
the “normal” sample incorporates these changes. This
procedure will also give the established observers practice
and new observers training.
Aggressive Behavior in a Colony Management
Context
There are a number of circumstances in which it is
important to observe and record nonhuman primates'
interactions, but when discussing aggressive behavior,
three situations stand out: pair formation, group formation,
and intra-group aggressive behavior. Specific methodolo-
gies for pairing animals or forming groups are covered
more fully in Chapter 6, but some things to consider before
designing behavioral data collection protocols are pre-
sented below.
Pair Formation
When unfamiliar animals are paired, there will be a period
of possibly severe fighting while dominance relationships
are established. Questions to consider include how the level
of aggressive behavior in the pair compares with levels in
other, more established pairs and how the level of aggres-
sive behavior has changed since the pair's first introduction.
For the former, a previously collected dataset from estab-
lished pairs is essential.
Whether you put the animals together directly from the
beginning or use a series of increasing levels of contact,
collecting consistent and reliable data will help you decide
whether the pair is compatible. Since only two animals are
involved, using a focal animal technique and switching
between individuals will build a picture of the relationship
from both sides but will require more time to build a large
database. The confined space and the fact that there is only
one potential social partner make it possible to use all-
occurrence sampling to collect more data faster. A grab,
bite, or time spent grooming can only be with one other
animal.
Control Group Importance for Comparing
Results
When your ethogram is set up, you have your dependent
variables defined, but hypothesis testing for changes in
rates of behavior requires that you have something to
compare against. During a pair or group formation, you can
compare rates of interactions over time. If you record data
at set times over the hours, days, or weeks after a pair
formation, you can ask whether the rates of bites or grabs or
the duration of grooming bouts, for example, has stayed the
same, increased, or decreased. Plotting the behavioral data
over time can give you a picture of how interactions within
the group are changing.
Search WWH ::




Custom Search