The frameworks for prose literacy, document literacy, and numeracy met the criteria established for inclusion and were referred to international item development. Chapter 4, 5 and 6 document the frameworks, the item development process and how the assessment protocols were validated at the international level. The initial Problem Solving framework and approach to assessment proposed by Harry O'Neil and this team at CRESST, UCLA proved to be unreliable in initial multi-country piloting and a second group of experts, headed by Jean-Paul Reeff, was recruited to refine the framework and to develop on an alternative approach to assessment. This redevelopment, effort, documented in Chapter 6 of this report, proved successful.

Chapter 7 in Part III also documents the development effort for teamwork headed by O'Neil and his colleagues at CRESST. It produced a useful framework but failed to provide an approach to assessment of sufficient quality to merit inclusion in an international comparative study. A second team, headed by David Baker at the America Institutes for Research, developed a slightly modified framework and a different approach to assessment. This effort also failed to meet the standards set for inclusion in ALL. Despite failing to yield the desired result - a valid, reliable comparable and interpretable teamwork scale - the development effort did provide some useful results.

Analysis of small scale pilot results suggests that teamwork is a complex multilevel phenomenon with effects at the individual, the occupational, the workgroup, the firm, the industry and the national level as well as a temporal aspect that depends on where the firm is in its product lifecycle and corporate strategy. To be informative, sampling strategies must reflect all of these levels explicitly, something that is beyond the scope of the current ALL household-based design.

The framework and associated measures for practical intelligence have been included as Part III of this report (Chapter 8). Although these were deemed to be adequate the statistical techniques employed to summarize proficiency were judged to be difficult to interpret within the context of an international comparative study. This domain was dropped from the international study pending additional work.

Development of a framework for measuring information and communication literacy, headed by Jean-Paul Reeff, then of the Luxembourg Ministry of Education, failed to provide a clear approach to measurement. As a result Graham Lowe of the University of Alberta was recruited to develop a behavioural module, focusing on familiarity and use of ICT's, to replace a direct measure of ICT proficiency. This module was eventually included in the ALL assessment. In a parallel activity Statistics Canada chose to invest in an ETS project designed to develop a framework and viable approach to measurement of ICT skills. Although this development failed to yield results in time for inclusion in the ALL study it did establish the basis for doing so in future assessment cycles. Interested readers are referred to chapter 9 for additional detail on this work.