ABSTRACT
In 1956 Herbert Blumer wrote a devastating critique of ‘variable analysis’. He poured scorn on that research strategy which, regardless of the issue at hand, reduces social inquiry to the task of identifying the crucial variables which constitute that issue and goes on to investigate the patterns of statistical associations between the said variables. Many, many others have followed in his wake and by now we have reached the stage, in some quarters, where more sociologists can be relied upon to argue his case than might be expected to know the difference between an independent and a dependent variable.
