In cognitive load theory, the superiority of the Example-Problem sequence over the Problem-Example sequence has become a classic paradigm. The comparative effectiveness of these sequences, however, is subject to the influence of the factors of element interactivity and prior knowledge, and studies have examined these influences focused mostly on procedural rather than conceptual knowledge. This paper takes a deeper look at the effect of types of knowledge concentrating on conceptual knowledge. An experiment is reported comparing the Problem-Example and Example-Problem sequences on two levels of element interactivity, low versus high, which were associated with two types of conceptual knowledge (general principle knowledge and knowledge of principles underlying procedures, accordingly). Since there was no difference found between these sequences for either level of element interactivity, the paper discusses conditions of effectiveness of example-based instructions for different knowledge types in the broader context of Explicit Instruction First and Problem-Solving First approaches.