曝光臺 注意防騙 網曝天貓店富美金盛家居專營店坑蒙拐騙欺詐消費者

variables. It must be noted that in some case, the aim of the observations may be to
build a topographic map of existing C2 behaviours, such as teamwork, that can be
used to assist in guiding future system design. If this is the case, then the issue of
dependent and independent is not applicable. If the goal is to examine the effect of a
system such as BCSS on C2 team processes, the dependent variables will include
communication, cognitive and physical workload, and frustration. These are
behavioural variables. The independent variable in this case will tend to be the
various components of BCSS. Other possible independent variables include
communication architecture and team structure. It is important to notice that the
independent and dependent variables are interchangeable. In a field setting, clear-cut
variable boundaries tend not to exist, with distinctions hovering between the
arbitrary and the outright fuzzy.
3.4 When to Collect Data
Another area of observational research that needs to be clearly defined is when the
observations should occur. This can only be determined through preliminary
observations. However, when it comes to a military exercise, it is usually necessary
to attend for the full duration (using sampling techniques to collect the data). Only
observing for peak activity periods will act as a cue to the subjects, and will bias the
data (particularly if a behaviour taxonomy is the aim). In real world activities,
observation will largely be opportunistic. Ideally, it would consist of at least three
full days of observation spread across one week. While this time frame is somewhat
arbitrary, it has been shown to be useful as a “rule of thumb” for conducting
observational research projects (Lehner, 1979). To determine the duration
systematically, it is necessary to perform a saturation sample pilot study. This
involves observing everything for a week, identifying the points that give
representative coverage, then going back and observing these points to test whether
it is adequate. It must be noted that in a Defence setting, collecting the data necessary
to perform a saturation sample study is not usually possible. As a consequence,
wherever possible the above rule of thumb should be adopted.
DSTO-TR-1034
8
3.5 Developing a Behaviour Taxonomy
In general, the first step is to formulate a list of well-named, clearly defined
behaviours relevant to the research question. Important aspects of observational
research are the behaviour categories and data parameters. This requires knowledge
of the relevant literature, as well as preliminary observations in the research
environment. Knowing what has been done before may avoid unnecessary repetition
and may suggest possible methods of data collection. As will be discussed, a survey
of the literature on team performance suggests clearly defined team and task
categories. If, as is the current case, relevant categories have been suggested in the
literature, preliminary observations can be used to assess whether they are relevant
to the new context. If data is to be collected in a systematic fashion, the categories
must also be clearly defined. A precise operational definition must be written out to
ensure that observers do not drift from the original definition, and to enable other
researchers to use the same system.
Generally, a select number of behaviours must be chosen to reduce the chances of
being swamped while observing. During data analysis, the behaviours are often
collapsed into function categories, or factors. It should be noted that there is a degree
of arbitrariness to this process. As will be discussed, inter-observer ratings improve
the reliability of information.
Once the relevant behaviours have been established, it is necessary to determine
what data parameters are important. For example, is it important to know how often
the behaviour occurs, how much time is spent in a behaviour, how long a bout of
behaviour lasts once it is triggered, and/or what sequences of behaviour are
important. It is also necessary to determine whether the identification of individuals
is important, or whether it is the overall level of a particular behaviour within a
group that is the critical parameter. If the focus is on social dynamics, individuals
need to be identified. However, if the goal is to delineate tasks within a cell,
individual data may not be necessary.
Once these issues are established, it is valuable to determine whether the research
question is answerable with the data parameters chosen. This involves analysing
some data as early as possible. It may be necessary to refine or even dramatically
alter techniques. For example, collection may be focusing on rate of behaviour, when
　
中國航空網 www.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：航空資料29(63)