What are the potential limitations of an evaluation? How can researchers watch for and address them?

What will be an ideal response?

Evaluation research frequently encounters some potentially limiting factors. For example, the internal validity of evaluation may be limited by the unfeasibility of random assignment as a result of resistance from subjects, sponsors, administrators, or staff. The resulting selection bias and the later differential attrition rates could enhance or mask the true impact of the program.
The external validity of evaluation may be threatened by other commonly encountered features of health programs. For example, the program may be effective in a particular geographic setting where external factors are crucial to its implementation. When the effectiveness of a program is dependent on the personal qualities and interests of the staff who administer it, program results might not be generalizable to widespread application of the program by staff who may be less capable, committed, or interested in program success. The knowledge that one is a participant in an experimental health program can generate enthusiasm and cooperation that might not be there when the program is no longer novel or being evaluated. These potential threats to external validity should be watched carefully, or the extension of a program or policy to other participants or beneficiaries could be hampered.
Time, financial limitations, and political climate are important constraints affecting the scope and depth of program evaluation. The amount of time the evaluator can devote to the project may determine the choice of evaluation methods and affect the ultimate breadth and quality of an evaluation. The amount of time and effort the evaluator will devote to the project is also dependent on available budget. The political climate could influence an evaluation in several ways. It might place constraints on evaluation design and data collection. The use of research results may be influenced by political considerations. When evaluation results contradict deeply held beliefs or vested interests, they may not be taken very seriously. After all, an evaluation is only one ingredient in a political process of balancing interests and making decisions.
In addition, answers should also try to address these potential limitations related to evaluation by pointing out what researchers can do to either avoid, reduce, or caution users of these potential limitations.