A retroactive decision was made to compare the correlations of sequential pairs of words and how often each of such pairs appears in direct relation within the model. This was able to be done effectively because both subjects reported having not revised any entries, so the original order of the entries remained in tact. This correlation, reported as the frequency of directly-connected pairs from the original list, was high for the expert and quite low for the novice. High correlation may suggest that the subject generated a subsequent term in the list because of its semantic relevance to the previous concept; this would also signal a more structured mental representation of basic firefighting.

Evidence from Part II of the study supports the theory that experts have better-defined scripts and schemata than those of their novice counterparts. The expert firefighter had better memory accuracy in general (greater number of details recalled with fewer errors), but it’s key to note that every single accurate detail recalled by the expert was that of fire-scene (script-consistent) information. For the novice, this was not the case: only about 73% was such. This suggests that the expert firefighter was relying less on working memory and more his scripts and schemata stored in long-term memory. The excerpt from the book was chosen specifically because of its high proportion of information that was likely to fit a firefighter’s general script; it should be no coincidence that the firefighter recalled more than twice as much information as the novice firefighter. (In the future, this type of data should be analyzed with multiple subjects and appropriate statistical analyses to quantitatively assess for significance or confidence.)