Integration and elaboration of the De Goede (2007) and Burger (2012) learning potential structural models

Venter, Braam (2019-04)

Thesis (MCom)--Stellenbosch University, 2019.

Thesis

ENGLISH SUMMARY : South Africa’s turbulent past has left Human Resource managers in South Africa with a unique challenge. Apartheid legislation unfairly discriminated against certain groups of people, which led to these groups’ skills and competencies being underdeveloped. The consequence of this is that the skills of a large number of employees in the South African labour market are underdeveloped, which has subsequently led to adverse impact in valid, fair strict-top-down selection. This has fundamentally been caused by the fact that the competence and human capital in South Africa has not been uniformly developed across groups. The current situation should be addressed by organisations, not only because it is required by legislation, but because it is central to the economic survival of South African organisations. In the final analysis it should be addressed by organisations because it is the morally correct thing to do. Unrest is growing in South Africa especially under those South African groups that have been previously disadvantaged. The masses are tired of not having the opportunity to productively take part in economic activities and experience economic freedom. A testimony to this is the meteoric rise of the Economic Freedom Fighters (EFF) party that, in its first national election as an official party, obtained 9% of the total votes. To address this unrest individuals from previously disadvantaged groups, with the necessary learning potential, need to be identified and developed. Therefore, a method is needed in South Africa that will identify individuals who display a high potential to learn and that will gain maximum benefit from affirmative development opportunities. In order to successfully address the negative effects of South Africa’s past through affirmative development the complex nomological network of latent variables underlying learning performance needs to be understood. It will be possible to rationally contribute to successful accelerated affirmative development when a comprehensive understanding of the factors that underlie learning performance, and how these factors combine to determine learning performance, exists. The primary objective of this study is to integrate the De Goede (2007) and Burger (2012) learning potential structural models and to expand and modify the integrated De Goede- Burger model. More specifically the objective of the current research was to: - Identify additional latent variables not currently included in the integrated DeGoede- Burger learning potential structural model that might directly or indirectlyinfluence classroom learning performance and learning performance duringevaluation; - Develop hypotheses on the manner in which these additional latent variablesshould be embedded in the integrated De Goede- Burger learning potentialstructural model; - Empirically test the expanded De Goede- Burger learning potential structural modelby evaluating the model’s absolute fit and the testing the statistical significance ofhypothesised paths in the model. Once additional latent variables were identified and hypotheses were developed on the manner in which these additional latent variables are embedded in the integrated De Goede- Burger learning potential structural model, the expanded model was empirically tested. The attempt to obtain measurement model fit was constrained by the fact that the number of observations (114) that were obtained were smaller than the number of freed parameters in the congeneric measurement model in which the intercepts were not modelled. The measurement model was subsequently fitted as a tau-equivalent model. The fitted measurement model did not provide a sufficiently credible description of the process that generated the observed inter-item parcel covariance matrix to have faith in the measurement model parameter estimates. The researchers consequently deemed it pointless to proceed with the fit of the structural model via structural equation modelling. In-order to remedy the situation the decision was made to take a more robust approach by evaluating the path specific substantive hypotheses via multiple regression analysis. This meant dissecting the structural model into 7 separate regression models, fitting each of these via multiple linear regression analysis and testing the path-specific substantive hypotheses by testing the significance of the partial regression slope coefficient estimates. The regression analysis results indicated that most of the independent variables explained unique variance, which was found to be statistically significant (p < .05), in the specific dependent variables that the independent variables are proposed to influence. However, no support was obtained for the path-specific substantive research hypotheses that learning performance, exerts a unique positive influence on academic self-efficacy. Also, no support was found for the path-specific substantive research hypotheses that the information processing capacity*time cognitively engaged interaction effect exerts a unique positive influence on automisation. Limitations to the research methodology are noted. Practical recommendations are made. Recommendations for future research are made.

AFRIKAANSE OPSOMMING : Suid-Afrika se onstuimige verlede het menslike hulpbronbestuurders in Suid-Afrika gelos met 'n unieke uitdaging. Wetgewing gedurende Apartheid het op ‘n onregverdige wyse gediskrimineer teenoor sekere bevolkingsgroepe wat daartoe gelei het dat dié groepe se vaardighede en bevoegdhede onderontwikkel is. Die gevolg hiervan is dat die vaardighede van ‘n groot hoeveelheid werknemers in die Suid-Afrikaanse arbeidsmark onderontwikkel is wat gelei het tot nadelige impak in geldige en billike bo-na-onder seleksie. Die fundamentele oorsaak hiervan is die feit dat die bevoegdhede en intellektuele kapitaal in Suid-Afrika nie eenvormig oor groepe ontwikkel is nie. Die huidige situasie behoort aangespreek te word deur organisasies, nie net omdat dit vereis word deur wetgewing nie, maar omdat dit van kardinale belang is vir die ekonomiese oorlewing van Suid-Afrikaanse organisasies. Nie net is dit die ekonomiese regte ding om te doen nie, maar dit is ook die morele regte ding om te doen. Daar is ‘n onrus wat besig is om te groei in Suid-Afrika, veral onder voorheen benadeelde groepe. Dié groepe se ongelukkigheid is besig om te groei omdat hulle nie die geleentheid gegin word om deel te neem aan ekonomiese aktiwiteite en om ekonomiese vryheid te ervaar nie. Die feit dat die Ekonomiese Vryheidsvegters (EFF) in hulle eerste nasionale verkiesing 9% van totale stemme ingepalm het, is getuienis van die feit dat voorheen benadeelde groepe honger is vir ekonomiese geleentheid en vryheid. Om hierdie onrus aan te spreek moet individue van voorheen benadeelde groepe, wat beskik oor die nodige potensiaal, geïdentifiseer word en ontwikkel word. Om die identifiseering van potensiaal te bewerkstellig word ‘n metode in Suid-Afrika benodig wat individue wat ‘n hoë potensiaal het om te leer, en wat maksimum voordeel uit regstellende ontwikkelinggeleenthede sal kry, te kan identifiseer. Om die negatiewe gevolge van Suid-Afrika se verlede op ‘n suksesvolle wyse reg te stel deur regstellende ontwikkeling moet die komplekse nomologiese netwerk van latente veranderliks onderliggend aan leerprestasie verstaan word. Dit sal moontlik wees om op ‘n rasionele vlak by te dra tot suksesvolle versnelde regstellende ontwikkeling wanneer ‘n omvattende verstaan ontwikkel is oor die faktore wat onderliggend is aan leerprestasie, asook hoe die faktore kombineer om leerprestasie te bepaal. Die primêre doelwit van die studie is om die De Goede (2007) en Burger (2012) leerpotensiaal strukturele modelle te integreer en om die geïntegreerde De Goede- Burger model uit te brei en aan te pas. Meer spesifiek was die doelwit van dié huidige navorsing om: - Addisionele latente veranderlikes te identifiseer wat nie tans in die geïntegreerdeDe Goede-Burger leerpotensiaal strukturele model ingesluit is nie, maar wat moontlik direk of indirek ‘n invloed het op leerprestasie in die klaskamer en leerprestasie gedurende evaluasie; - Hipoteses te ontwikkel oor die wyse waarop dié addisionele latente veranderlikesingesluit moet word in die geïntegreerde De Goede-Burger leerpotensiaalstrukturele model. - Empiries die uitgebreide De Goede-Burger leerpotensiaal strukturele model tetoets deur die model se absolute passing te evalueer en deur die statistiesebeduidenheid van die voorgestelde paaie in die model te toets. Na addisionele latente veranderlikes identifiseer is en hipotesese ontwikkel is oor die wyse waarop dié addisionele latente veranderlikes ingesluit is in die geïntegreerde De Goede- Burger leerpotensiaal strukturele model, was die uitgebreide model empiries getoets. Die poging om aanvaarbare metingsmodelpasgehalte te vind is aan bande gelê deur die feit dat die getal waarnemeings (114) gelyk was aan die getal vrygestelde parameters in die kongeneriese metingsmodel waarin die afsnitte nie gemodelleer is nie. Die model is vervolgens as ‘n tau-ekwivalente model gepas. Die gepasde model het nie ‘n genoegsaam oortuigende beskrywing gebied van die proses wat die waargenome inter-itempakkie-kovariansiematrys gegenereer het om vertroue in die metinsmodelparameter-skattings te hê nie. Die navorsers het gevolglik besluit dat daar geen punt daarin is om voort te gaan met die passing van die strukturele model nie.. Aangesien daar nie passing vir die metingsmodel verkry is nie, het die navorsers het besluit om ‘n meer robuuste benadering te neem deur die baanspesifieke substantiewe hipoteses te evalueer via meervoudige regressie-analise, Dit het beteken dat die strukturele model vereenvoudig moes word na 7 afsonderlike regressie-modelle. Elkeen van die modelle is gepas word met behulp van meervoudige lineêre regressie-analise en die baan-spesifieke substantiewe hipotesese is getoets deur die statistiese beduidenheid van die gedeeltelike regressie-helling-koëffisiënt-ramings te toets. Die resultate van die regressie-analise het aangedui dat meeste van die onafhanklike veranderlikes unieke variansie in die spesifieke afhanklike veranderlikes wat die onafhanklike veranderlikes voorgestel is om te beïnvloed verklaar, wat as statisties beduidend gevind is (p < .05. Daar was egter geen ondersteuning gevind vir die baanspesifieke substantiewe navorsinghipotesese dat leerprestasie, unieke positiewe invloed uitoefen op akademiese selfdoeltreffendheid nie. Daar is ook geen ondersteuning gevind vir die baanspesifieke substantiewe navorsinghipotesese dat die informasie prosessering kapasiteit*tyd kognitief ingespan interaksie effek ‘n unieke positiewe invloed uitoefen op outomatisasie nie. Tekortkominge in die navorsingsmetodiek word uitgewys.. Praktiese aanbevelings word gemaak. Aanbevelings vir toekomstige navorsing word gemaak.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/106079
This item appears in the following collections: