Little is known about the effectiveness of district plans in protecting built heritage, which is a matter of national importance under New Zealand's Resource Management Act 1991 (RMAct). This is despite the fact that the RMAct directs planning agencies to evaluate the effectiveness of plan provisions. This lack of evaluation is not unique to New Zealand or merely symptomatic of heritage planning. Instead, it is a shortcoming in planning theory and practice internationally; a well recognised impediment being that planning lacks a suitable evaluation approach. This thesis aims to address this deficiency by proposing a methodology for evaluating plan effectiveness and applying it to the built heritage provisions of two district plans. The methodology adopted has been shaped by the theory-based and realist evaluation approaches, as developed in the field of programme evaluation. Both approaches share a common ontology regarding claims of causality, which stresses 'knowledge in context'. Thus, a central endeavour of the research is not only to identify the environmental outcomes arising from plan implementation, but also to understand how and why the implementation context promoted or inhibited the achievement of plan goals. In so doing, the causal and implementation theories underpinning the plans' heritage provisions are exposed, modelled and tested. The findings reveal that plan implementation failed to prevent the loss of built heritage values in many instances. While the plans' causal theory was largely sound, key aspects of the implementation theory were not realised during the development control process. Plan quality was a significant factor, as was the commitment and capacity of developers to comply with the plans. The institutional fixation on consent processing speed rather than environmental outcomes was a further impediment. Overall, the theory-based approach provided a useful framework for determining plan effectiveness and holds promise for evaluating plan issues other than built heritage.