We investigated what impact elaborated feedback has on sixth graders' deep-level comprehension of texts within a computer-based assessment. Experiment 1 (N = 566) focused on the contents of computer-provided elaborated feedback (i.e. inference-prompts, error explanations, or monitoring-prompt) using a control-group design. Results showed that none of the feedback treatments had an effect on performance. This appeared to result from participants' low commitment to processing the feedback. Experiment 2 (N = 251) focused on the feedback presentation type by varying computer-mediated and person-mediated inference-prompts within a control-group design. Results showed that only the person-mediated inference-prompts had significant effects on performance with respect to the correction of initially false answers to comprehension questions and the performance on subsequent test questions. Findings of both experiments indicate the impact of inference-prompts on text comprehension within performance assessments, highlighting the need to explicitly account for motivational issues in feedback interventions on higher-order reading processes.