Five types of multipopulation surveys are defined and described: perio
dic surveys; multidomain designs; multinational survey designs, cumula
ted and combined samples; and controlled observations or quasi-experim
ental designs. I emphasize the deliberate designs of these surveys, no
t the mere post hoc or ad hoc utilization of survey results. Most impo
rtantly, I emphasize a sharp distinction between seven listed aspects.
Similarity and standardization of the survey aspects (definitions, me
thods, measurements) are essential to avoid biases in comparisons, tho
ugh admittedly difficult. In contrast, flexibility in sampling designs
and sizes, to reduce variances, is permissible, also desirable to red
uce variances and costs. This divergence in strategies still needs to
be understood and applied.