Internships, like other forms of cooperative education, involve students undertaking work as an integrated component of their tertiary education programme. It is only relatively recently that research has been undertaken to consider what it is that students actually learn when undertaking such work. This is because workplace learning is complex, informal, and subject to the contextual influences of the particular workplace. Such complexities are heightened in some disciplines, like business, where the work occurs in diverse workplace settings, with the work requirements being unique to each student. Not surprisingly, there is even less research that may assist practitioners to find ways of assessing such learning. Most forms of summative assessment are based on adherence to the principles of criterion-referencing, which require using the same criteria and set of standards for all students. But when the learning takes place away from the formal, structured environment of an educational setting, underpinned by a fixed and 'known' curriculum, adoption of such principles is problematic, and can create the conditions for assessment to be inherently unfair (and therefore invalid). This is because they can fail to take account of the individual and variable nature of the work, the contextual influences involved, and the conscious and unconscious biases of the assessors. So how does one assess student performance and learning in cooperative education? This thesis sets out to address this question in relation to a business internship that is part of an undergraduate degree programme in a large New Zealand polytechnic. A multi-theoretical approach was taken to the study, which provided valuable frames of reference for viewing assessment of learning. By adopting an interpretive methodology, primarily driven through participatory action research, the contextual complexities involved were able to be incorporated into the research design. Through engagement with the practitioner, a self-assessed, evidence-based portfolio model of assessment was created. A key feature of the model is that the 'truth' of students' performance emerges through consensus, based on an informed understanding of the subjective elements and contextual influences present. An important contribution is the on-going dialogue that occurs, throughout the work placement, between the stakeholders (employers, students and academics). The study has found that the assessment model developed was able to address the complexities involved. The stakeholders supported and valued the portfolio assessment model, and it was apparent that the formative aspects of the portfolio contributed positively to its summative outcome, without seemingly compromising the nature of either. The portfolio also had a high 'backwash' effect on learning, contributing to its consequential validity. Such learning included students' increased awareness of the important competencies required in the workplace and how such competencies contribute to effective performance. In addition, the self-assessed nature of the model contributed to students' development as lifelong assessors of their own learning; preparing them to become self-regulating professionals. Finally, it was apparent that informal, emergent learning, derived from the sociocultural influences present, was an important feature of students' workplace experiences.