Abstract
This article examines some recent British research in the field of attitude measurement, and contends that a few of these studies display serious faults in instrument design and usage. Effective research in this field demands the construction of attitude scales which clearly reflect some underlying theoretical construct; it requires distinct variables to be kept separate and not added together to produce meaningless total scores; and it requires instruments chosen as outcome measures in research on treatment effects to display some defensible connection with the treatment being studied. This article was written during 1973 while the writer was on study leave at the Centre for Science Education, Chelsea College, University of London.

This publication has 4 references indexed in Scilit: