Conclusion validity focus on how sure we can be that the
treatment we used in an experiment really is related to the
actual outcome we observed. Typically this concerns if there
is a statistically significant effect on the outcome.
If there is a statistically significant relationship the Internal
validity focus on how sure we can be that the treatment
actually caused the outcome. There can be other factors that
have caused the outcome, factors that we do not have control
over or have not measured.
Construct validity focus on the relation between the theory
behind the experiment and the observation(s). Even if we
have established that there is a casual relationship between
the treatment of our experiment and the observed outcome,
the treatment might not correspond to the cause we think we
have controlled and altered. Similarly, the observed outcome
might not correspond to the effect we think we are measuring.
Finally, the External validity is concerned with whether we
can generalize the results outside the scope of our study. Even
if we have established a statistically significant casual relation
between a treatment and an outcome and they correspond to
the casue and effect we set out to investigate the results are
of little use if the cause and effect we have established does
not hold in other situations.
Conclusion validity focus on how sure we can be that the treatment we used in an experiment really is related to the actual outcome we observed. Typically this concerns if there is a statistically significant effect on the outcome.
If there is a statistically significant relationship the Internal validity focus on how sure we can be that the treatment actually caused the outcome. There can be other factors that have caused the outcome, factors that we do not have control over or have not measured.
Construct validity focus on the relation between the theory behind the experiment and the observation(s). Even if we have established that there is a casual relationship between the treatment of our experiment and the observed outcome, the treatment might not correspond to the cause we think we have controlled and altered. Similarly, the observed outcome might not correspond to the effect we think we are measuring.
Finally, the External validity is concerned with whether we can generalize the results outside the scope of our study. Even if we have established a statistically significant casual relation between a treatment and an outcome and they correspond to the casue and effect we set out to investigate the results are of little use if the cause and effect we have established does not hold in other situations.
Validity Threats in Empirical Software Engineering Research - An Initial Survey