Construct validity is concerned with both the definition and measurement of variables in a study. Construct validity regarding definition focuses on how well researchers define their concepts. The second aspect of construct validity is concerned with the ways in which variables are operationalized. That is, the methods researchers used to measure variables are crucial to the operationalization of those same variables.Researchers must carefully identify and define the independent variable.If not, alternative explanations may be offered to explain the changes in the dependent variable. For treatment purposes, the treatment model must be clearly defined. This means researchers should have a treatment manual including adherence checks (do the therapists actually stick with the model) and competence checks (how competent are the therapist doing the therapy).

For this week's discussion board,we will follow these steps.

For the purpose of evaluating construct validity, the first step is to identify all the variables being studied. How do researchers define their variables? How do they measure these variables? For example, if researchers are studying the effectiveness of cognitive therapy on depression, then they need to clearly specify cognitive therapy (use a manual; identify the competence and adherence of therapists to the manual). They also need to define depression. Do the researchers use their own measures for depression? If so, did they convince you they have measured depression well?

The second step of evaluating construct validity is to describe how well the researchers measured their variables of interest. You could use the ten threats to construct validity:

1. Inadequatepreoperationalexplication - don't clearly define all relevant variables
2. Mono-operation bias - Single measure is used for the independent and dependent - measure
3. Evaluation apprehension - Faking good
4. Experimenter expectancies - Investigators communicate their expectancies to participants which biases outcomes
5. Confounding constructs and levels of constructs - A specific level of a construct is measured which effects the outcome of the study
6. Restrictedgeneralizabilityacross constructs - By narrowly defining constructs, one risks missing important relationships among the constructs
7. Mono-method bias - Using the same method of measuring a construct -Self report measure
8. Hypothesis guessing in the experimental condition - Participants try to figure out what the investigation is about and either conform or rebel.
9. Interaction of different treatments - There is a potential confound when participants are exposed to multiple treatments -The outcome of the study could be a result from the last treatment in the sequence or a combination of the treatments
10. Interactions between testing and treatment - The pretest may educate and sensitize the participant to the posttest, therefore, improved performance may be the result of training on the pretest.

The final aspect of evaluating construct validity is to detail your opinion as to the researchers' methods. That is, do you think they did a good job based on step 2?