"This paper discussed the possibility to improve public communication campaign theory, by making use of data obtained through mass media health communication campaign evaluations. The idea of an ‘engineering’ approach to campaign design, where theory and scientific findings are systematically used and adopted for practical problems, plays an important role in the discussion ... In the third part, a sample of 33 evaluation reports for mass media health communication campaigns was analyzed. 32 of these reports have not been published in an scientific journal. The evaluations were conducted in 22 different countries. The analysis of the reports focused on the campaign goals, evaluation outcome measures, research design and methods, and on questions of validity. The findings suggest that theory is not widely and consequently used to inform health mass communication campaigns or their evaluations – with notable exceptions. While there is a large number of outcomes measured, they seem to be taken out of theoretical context. Neither the campaign goals nor the evaluation measures reflect the large number of possible communication strategies that the various communication or behavior-change models and theories imply. Unintended campaign effects were mostly ignored. In very few cases the campaign designers or evaluators make use of an effects model or program logic model. This is one of the areas where I see the possibility of an important improvement. The methodology of campaign evaluation is relatively homogenous across the 33 cases in regards to data collection method. Standardized questionnaires are the dominating data collection instrument. Non-reactive observation or tracking methods are very rare. A surprising two thirds of the evaluations did not use multivariate analysis, and the reliance on self-reports raises questions of reliability." (Summary, p.120-121)