The development of a scale for talent retention

Abstract

Introduction Talent retention remains a key goal of successful organisations even though there is a lack of academic clarity on the most appropriate path to follow in order to retain talent or minimise the loss of valuable employees. Research purpose From a methodological point of view, the purpose of this study was to produce a valid and reliable measure that could contribute to academic research in talent retention in South Africa. Motivation for the study Within the South African context, talent retention faces the additional challenges of scarce skills, resource constraints and labour legislation. Education, together with training and innovation, is viewed as central to solving the national challenges of poverty and inequality (National Planning Commission, 2012:261). Concern about the retention of high quality and scarce skill employees in general education and higher education in South Africa has been raised as being of critical importance in addressing the social and economic challenges that the country faces (National Planning Commission, 2012; Spaull, 2013). Research design, approach and method The research could be classified as a methodological research study. The scale development process of DeVellis (1991) and Hinkin (1995) provided the framework for the research design. A mixed method research strategy was followed involving the collection of qualitative data and quantitative data. Semi-structured interviews with key respondents from six organisations were conducted to help determine the parameters of turnover and retention data that were included in the measurement scale. The qualitative data generated from these interviews was analysed using thematic analysis. The developmental study was conducted in two distinct samples of employees, n=153 in Higher Education Institutions and n=1148 in general education. The quantitative data produced during the developmental study was analysed using descriptive statistics, exploratory factor analysis, and when indicated, confirmatory factor analysis. By means of multiple group structural equation modelling it was possible to establish a comprehensive model for the Talent Retention Scale. Validity, reliability and invariance testing for the measurement scale were described. Main findings The qualitative data from the key respondent interviews enabled the development of scale items with contextual relevance to South African organisations. The developmental study in Higher Education Institutions and general education highlights concerns about job satisfaction in education and the risk that employees would leave as soon as possible or when alternative opportunities become available. The risk for the turnover of young, newly qualified educators in general education has been identified in this research. Being unhappy about financial compensation was identified as the most likely reason that employees in this study would consider leaving their institutions. A supportive relationship with a direct line manager is a potential employee retention factor identified in this study and employees in the higher education and general education samples were generally satisfied with the employee-direct line manager relationship. The results showed that the Talent Retention Scale, with higher order dimensions Compensation and Recognition; Manager Relationship and Institutional Practices showed significant negative relationships with Intention to Quit. Therefore, for the entire sample, it can be assumed that the proposed Talent Retention Scale seems to be useful to predict Intention to Quit, and the model fits the data adequately. Validity and reliability in the HEI study were sufficient although the sample size was too small for a thorough psychometric analysis. Internal consistency reliability was established in both samples and Cronbach’s alpha reliabilities ranged between 0.764 and 0.928 for the scale factors. In the general education sample support for content validity and construct validity were obtained with the overall measurement scale and second-order CFA producing good results in terms of overall model fit. However, in terms of measurement invariance the Talent Retention Scale had adequate invariance over gender groups but invariance was not clearly established over employment equity groups.