Difficult to test the predictions
Sutherland aimed to produce scientific framework within which future offending behaviour could be predicted and this means that the predicts must be testable
Problem is the many of the concepts aren't testable as they can't be operationalised
IE hard to see how the number of pro-crime attitudes can be measured
Based on assumption that offending occurs if pro-crime > anti-crime attitudes without being able to measure it we can't know at what point the urgent to offend is realised + when the career is triggered
doesn't have scientific credibility