Construct  Validity:

In  cqnstruct validity, the investigator  is  concerned with the questions  the  concept under investigation being adequately measured? Is  there  a valid basis  for  inferring the  scores? This begins when  the investigator formulates hypotheses, based  on  theory, about characteristics  of those who  have high scores on  a  test (compared to those with  low  scores). The measure  can be dependent variable  in  some hypotheses  and  an independent variable  in  others. Evidence  of construct validity is not established within a single study. 

Variables like anxiety, creativity, job  satisfaction, attitude towards  numerous phenomena, intelligence, empathy, dependency, power, self-concept, etc.  are  all constructs. These variables have been constructed from theory and empirical data  (i.e. quantitative  data  from research). For measures of variables that are constructs, evidence must be presented that  scores represent the degree to which an individual  does  indeed possess  or exhibit the  construct. One common approach  to  construct validation  is  "known-groups technique" where the group which is expected to  differ on  the particular attribute because of  some known characteristics  are administered the test. For example,  in  validating  the  tool to measure  fear in pregnancy of primigravida, the researcher would also administer the tool to multigravidae.  Because one would logically infer that women who  have not given birth (primi) would  express more  fear than multigravidae.