A researcher uses $t$-tests to compare a new study technique against traditional learning across 15 different topics in a biology course. Without adjusting for multiple comparisons, what is the expected number of statistically significant results ($p < 0.05$) due to chance alone, even if the study technique has no actual effect?