Predicting trust in online advertising with an SEM-artificial neural network approach

•A nonlinear ADTRUST and Trust Building Model was integrated to predict trust.•Age, gender, education and hours spent were incorporated as the control variables.•A 10-fold cross-validated SEM-ANN analysis with the FFBP algorithm was applied.•Reliability, website quality, willingness, reputation & hours spent are significant.•76.74% of the variance of trust in online advertising were explained by the model. Trust has an imperative role in online advertising because the effectiveness of the adverts will be greatly affected when consumers distrust online adverts. Currently, the level of consumers' trust in online advertising remains low. The current study will assess the drivers of trust by integrating the Trust Building Model and the ADTRUST scale. Unlike present literature that utilized linear models, a Structural Equation Modelling-Artificial Neural Network (SEM-ANN) approach was used. This is because consumers’ trust-building is a complex process and linear models will over-simplify the complexity in the decision-making processes. Thus, the outcomes from linear models are inadequate and inaccurate to explicate the mechanism of trust creation in online advertising. Data were gathered from 500 online consumers using a mall intercept technique. The outcomes from the sensitivity analysis show that reliability is the most imperative antecedent of trust followed by website quality, willingness to rely on, reputation, and hours spent. The model predicts 76.14% trust in online advertising. The theoretical implication is the integration of the ADTRUST scale with the Trust Building Model. The methodological implication is the use of the SEM-ANN approach that captured both linear-nonlinear and compensatory-non-compensatory associations. The findings provide some useful practical implications for online advertisers, service providers, and retailers. The study has contributed useful theoretical and practical implications to the online marketing literature.