Studying worked examples has been found to be effective for learning problem solving, especially among students. However, students need to actively process example content to benefit from it and content must be structured in a manner that facilities knowledge construction. This study investigated the use of worked examples for teaching and learning programming. Programming involves problem analysis and solution generation. But students tend to jump to solution generation without adequately analysing the problem. Consequently, the current study designed and implemented a new worked example design that emphasised problem analysis and utilised highlighting through web technology to encourage active processing of example content. This study also evaluated the new design in a quasi-experiment in a university course in Malaysia, compared to subgoal labelled worked examples, and conducted over three sessions. Posttest performance was analysed using independent samples t-test and frequency distributions. The results suggested that worked examples based on the new design were more effective than subgoal labelled worked examples, with statistically significant difference in performance, and medium effect size for the first session. For the second and third sessions, performance was marginally better, with learning in both groups possibly limited by the complexity of the worked examples and assessments.