untitled

1 The provision of nest boxes for wood ducks will increase the size of the

population.

2 The provision of nest boxes for waterfowl will benefit their overall ecology.

The first is a hypothesis testable against a predicted outcome. The second is of the

type that covers everything and probably cannot be disproved. What is an “overall

ecology?” How is it measured? What species should it be measured on? Wildlife man-

agement objectives that are framed in an unverifiable form (example 2) are not very

useful whereas those in the form of testable hypotheses (example 1) allow us to learn

more about the system.

Fourth, even when there are such criteria to judge success, the wildlife manager

is seldom in complete control of the situation and hence cannot be held fully and

personally responsible for the outcome. A failure is usually referable to the acts of

many people, often interacting with changes in the environment (sometimes referred

to as acts of God).

Fifth, the wildlife and its habitat usually forms a robust ecological system. Within

rather wide limits that system will absorb the most inappropriate or irrelevant of

management and still look good.

Because the criteria for success are often fuzzy in wildlife management, the out-

comes of different management systems are sometimes difficult to rank. For example,

when managing deer populations do we shoot only bucks, shoot only does, shoot

70% bucks and 30% does, shoot 30% bucks and 70% does, shoot neither? All these

schemes have been tried and all have been reported as highly successful. Highly

successful with what end in view? How highly successful? Perhaps we ask those

questions less often than we should.

Agriculture made a major advance because R.A. Fisher invented the “analysis of

variance” and because a few agriculturalists recognized that here was a technique

that could differentiate the effect of different management treatments. More import-

antly they believed that differentiation was necessary. Wildlife management can learn

from the history of agriculture by incorporating more statistical design in manage-

ment programs.

In contrast to the value judgmentdiscussed above, the wisdom of a technical judg-

mentcan be evaluated according to strict criteria. If a manager decides that supple-

mentary feeding will increase the density of quail then that can be tested and the

decision rated right or wrong. If a manager decides that elephants must be culled

because, if not, they will eliminate Acacia tortilistrees from the area, that decision

is right or wrong and it can be demonstrated as right or wrong by an appropriate

experiment. Note that the decision on whether the local survival of the acacia

justifies the proposed reduction of elephants is a value judgment and hence not testable.

So there are value judgments and technical judgments and these must not be con-

fused one with the other. Technical judgments can be tested and should be tested.

By this means we learn from our failures as well as from our successes. A recurring

theme of this book is that wildlife management advances only when the efficacy of

a management treatment is tested. For that to happen the technical decision as to

the appropriate treatment must be stated in a form that predicts a verifiable outcome.

Research questions are usually phrased in positive form such as, for example, does

the mean body weight of black bears (Ursus americanus) change as we move from

EXPERIMENTAL MANAGEMENT 269

16.3 Technical judgments can be tested

16.3.1Hypotheses

WECC16 18/08/2005 14:47 Page 269