Summary from three previous publications by Peter Kelterborn.
This article characterizes some basic aspects of carrying out an archaeological experiment.

Note from the editors: We believe that our readers, particularly the younger ones in the arena of experimental research, could profit from reflecting on some constant principles of scientific experimentation. Therefore, and with the permission of the author, we condensed his publications on the subject as follows.

When discussing in theory the strengths and weaknesses of experimental archaeology, or when observing in practice current projects, invariably some aspects of the following four key issues arise. What follow is the personal opinion and the experience of the author, presented in the form of very concentrated and simplified headlines.

Key issue 1: state of the art

When does an experiment agree scientific standards as they are accepted today? When it is:
1.1 Clearly goal and solution oriented. Experimenting is not learning by doing.
1.2 Correctly modelled.
1.3 Measurable.
1.4 Repeatable.
1.5 Professionally planned and supervised during all seven basic activities.
1.6 Executed with the correct manual skill, not too high and not too low.

Key issue 2: basic activities

What is common amongst all trustworthy experimental projects? When all seven basic stages are well executed:
2.1 Procure, analyze, exploit the existing data base and make logical conclusions with a regard to the future project. The data base includes literature, archaeological originals and opinions of experts.
2.2 Conceive and plan, not only the experiment, but the whole project in its archaeological context.
2.3 Prepare and equip the infrastructure and the location of the experiment (lab or field).
2.4 Supply all original or substitute raw materials.
2.5 Make or buy tools, instruments, fixtures and gadgets.
2.6 Run the experiment, evaluate and draw conclusions simultaneously. If needed, make corrections.
2.7 Document, store and report.

Key issue 3: convincing quality

What is needed to become an expert in experimental archaeology? Besides experience:
3.1 A profound understanding of the technology and archaeology of period and culture being researched.
3.2 Familiarity with the experimental approach to problem solving.
3.3 Having the practical skills necessary for the activities in the project.
3.4 Talent to organize and improvise.
3.5 Ability to observe and report with precision.
3.6 Curiosity, honesty and self critique.

Key issue 4: winning strategies

Just as a game may be won in different ways, so a problem may be solved by following different paths or strategies. What can lead quicker to success in which situations?
4.1 "one after the other"
Indicates that all 7 basic activities must be thoroughly executed in the exact order as given in key issue 2. Recommendable for routine experiments or when surprises and changes are surely not to be expected.
4.2 "the art of reduction"
Is chosen when it is possible to isolate key problems and to define relevant information to be measured, even when it requires pre-testing. A very useful speed-up strategy when large volumes of possibly useless numerical data could be gathered in the course the experiments or already when procuring the data base. Also a helpful strategy before deciding to use computer supported working methods, which tend to accumulate superfluous data, costing time and work.
4.3 "cross link and backfeed"
Instructs to conduct all 7 basic activities as much as possible simultaneously with constant input of every bit of progress into ail the other activities under way. This very flexible approach is best when there is much interdependence between the seven basic activities to be expected or when the learning steps in one activity immediately influence the work in others.
4.4 "divide and control"
Advises to split a large problem into smaller independent problems and to solve these separately with separate experiments. This is particularly suitable for questions which involve large systems or when different stages are likely to occur in a single problem.
4.5 "overview first, details later"
Gives a clear priority to approximate and quick solution of the overall problem, before looking at exacter details or optimizations. This is for large scale projects and also as a defence against being side-tracked by secondary discoveries or secondary difficulties.
4.6 "eliminate first, solve later"
Uses the experiments to find what does not work or leads to a dead ends first. This can save much time, while giving the opportunity to gain practical experiences in the area of the problem.
4.7 "try now, think later"
Means that the actual experiment is started right away. While all other basic activities are postponed until absolutely required. This can be used as an opening move or as a break-through strategy when the problem is very complex or very new, or when the barriers are so high that the project would get stuck even before an experiment is started.

Earlier versions of this article were published in - see Bibliography.