Exploring the frequency contours in close reading texts

Many initiatives have emphasized the importance of challenging students in text using readability formulas. Almost all formulas employ mean word frequency yet long-standing issues abound. Researchers question using a singular mean with a skewed variable like frequency. They also question the degree to which frequency pinpoints complex words, noting that infrequent words can be rare, but not rich (e.g., smock). Unfortunately, no studies have examined how often this phenomenon occurs. In fact, word frequency is a “black box” because we do not know the actual words represented at different frequency points. This study used a new frequency metric to analyze 249 texts designed for close reading. Texts averaged 640 words and were written for Grades 2–8. For each text, curves were compared by grade and type. Words at each point on the curve were identified and analyzed. Findings illustrated the curves of word frequency throughout a text and unpacked the actual words at different points along the curves. Regardless of grade or type, curves were nearly identical in shape. Curves were characterized by four frequency strata. At the lowest frequency level were the rich, rare words that advance learning. However, analyses showed that between 34 and 72% of the time these least frequent words were rare but not rich, a pattern that occurred more frequently in narrative texts. Findings also showed that narratives were characterized by one very rare word that could influence the mean word frequency. The study included a series of graphics illustrating the functions of words at different tiers and listing the words along the curves.