identifies five important

issues that managers should consider when evaluating screening models:

$1.$ Realism. An effective model must reflect organizational objectives, including a firm$$s strategic

goals and mission. Criteria must also be reasonable in light of such constraints on resources as

money and personnel. Finally, the model must consider both commercial risks and technical

risks, including performance, cost, and time. That is$,$ will the project work as intended? Can

we keep to the original budget or is there a high potential for escalating costs? Is there a strong

risk of significant schedule slippage?

$2.$ Capability. A model should be flexible enough to respond to changes in the conditions under

which projects are carried out. For example, the model should allow the company to compare

different types of projects $($long$-$term versus short$-$term projects, projects with different technol$-$

ogies or capabilities, projects with different commercial objectives$).$ It should be robust enough

to accommodate new criteria and constraints, suggesting that the screening model must allow

the company to use it as widely as possible to cover the greatest possible range of project types.

$3.$ Flexibility. The model should be easily modified if trial applications require changes. It must,

for example, allow for adjustments due to changes in exchange rates, tax laws, building codes,

and so forth.

$4.$ Ease of use. A model must be simple enough to be used by people in all areas of the

organization, both those in specific project roles and those in related functional positions.

Further, the screening model that is applied, the choices made for project selection, and the

reasons for those choices should be clear and easily understood by organizational members.

The model should also be timely: it should generate the screening information rapidly, and

people should be able to assimilate that information without any special training or skills.

$5.$ Cost. The screening model should be cost$-$effective. A selection approach that is expensive to

use in terms of either time or money is likely to have the worst possible effect, causing orga$-$

nizational members to avoid using it because of the excessive cost of employing it$.$ The cost

of obtaining selection information and generating optimal results should be low enough to

encourage use of the models rather than diminish their applicability.

Let$$s add a sixth criterion for a successful selection model:

$6.$ Comparability. The model must be broad enough to be applied to multiple projects. If a

model is too narrowly focused it may be useless in comparing potential projects, or foster

biases toward some over others. A useful model must support general comparisons of project

alternatives.