Submitted in fulfillment of the requirements of the Master of Information & Communication Technology, Durban University of Technology, Durban, South Africa, 2016. / Community Development service practitioners in South Africa have been increasingly burdened with a broader range of responsibilities. Using ICT proficiently could enhance their productivity. The use of ICT is prevalent in almost every organization, therefore, it can also serve as a major factor in providing flexibility of service to communities, but only if the practitioners possess adequate skills in retrieving, presenting and disseminating valuable and adequate information within the workplace to the parties involved. Training interventions are however not always successful and their impact need to be measured. This study focussed on developing a Multi-Stage assessment model to measure the impact of an ICT training intervention, to provide evidence of the effectivity of the training. A syllabus was developed, based on the European e-Competence Framework for ICT Users - Part 1 (CEN, 2013) and contextualized to suit the work environment of Community Development Workers (CDWs) from KwaZulu-Natal, training material was created and a Learning Management System was used to deliver the training. A self-assessment questionnaire was used to determine the pre-training skills level of the CDWs, 189 CDWs from KwaZulu-Natal were trained and a second questionnaire was used to measure the impact of the training. Results were analysed using Item Response Theory, which provided a way to measure not only overall competency but responses to specific items. Classical Test Theory measuring frequencies and averages were also use and the two sets of responses were compared. The training was found to be effective. Recommendations towards the development of a Multi-Stage assessment model are made. / M

The Rapid Estimate of Adult Literacy in Medicine (REALM) is a standardised health literacy test developed in the United States that has been previously administered and researched in South Africa (Lecoko, 2000). It is an objective vocabulary test which uses item recognition of 66 health-related words where recognition is measured by pronunciation. It is designed as a screening instrument to identifY the health literacy levels of patients in clinics. Lecoko (2000) found the REALM to be largely inapplicable in a South African setting, in that only eight out of the 66 words could be deemed acceptable, in that they could be either both adequately pronounced and adequately comprehended or both inadequately pronounced and inadequately comprehended. This may have occurred for a number of reasons, including the administration of the test in a population for whom English is a second or other language, and inaccurate measurement of pronunciation and comprehension of words, with a key problem being that the choice of words in the original REALM may not accurately represent the range of conditions and issues in a South African healthcare setting. This thesis was therefore premised on the principled relexicalisation of the REALM, that replacement of the words used in the test, using a sample gathered from health information and promotion texts in local clinics, would improve its applicability. In this regard, an exhaustive sample was gathered and analysed and 66 new words were chosen. The test was also modified to include a more principled approach to pronunciation and comprehension issues, and to account for language proficiency differences in administering an English language test in an English second language population. This modified test, referred to as the REALM-M was administered to a group of respondents who were statistically similar to the group to whom the REALM had originally been administered, and the results were compared. It was found that relexicalisation increased the number of acceptable words on the test from eight to 38. However, researching the key discourses surrounding health literacy and comparing these with current discourses about literacy beyond the field of health care revealed that despite improved content validity over the REALM, the REALM-M lacked construct validity. This provided the opportunity to discuss the discourses of health literacy and to suggest the application of alternative paradigms in this field.