The New Art and Science of Teaching Reading. Robert J. Marzano
word recognition (which implies accurate word recognition) is necessary for skilled reading because the brain cannot pay attention to two tasks at once. It must either switch back and forth between tasks, or one of the tasks must become automatic. Reading researchers have been aware of this phenomenon since the late 19th century (for example, see Cattell, 1886; Huey, 1908). In 1974, David LaBerge and S. Jay Samuels articulated their theory of automatic information processing in reading. They posit that, in mature readers, surface-level reading processes (such as identifying letters, decoding them, and connecting words to meanings) operate with little to no conscious effort, allowing these readers to focus cognitive resources on making meaning from text (comprehension). Charles A. Perfetti and Thomas Hogaboam (1975) and Michael I. Posner and C. R. R. Snyder (1975) go a step further, claiming that word recognition has to become automatic because comprehension can’t; understanding a text always requires conscious cognitive control. Later researchers agreed with and extended these theories (see Logan, 1997; Stanovich, 1990), but the fundamental idea remained: “Rapid word recognition frees up mental resources for thinking about the writer’s intent and the meaning of the text rather than what word the print represents” (Roberts, Christo, & Shefelbine, 2011, p. 229). Research also indicates that many struggling readers’ difficulties stem from a lack of automaticity with word recognition (Fletcher, Lyon, Fuchs, & Barnes, 2007; Vellutino, Fletcher, Snowling, & Scanlon, 2004); these students can’t focus on understanding a text because they have dedicated their cognitive resources to figuring out what the words say.
For skilled readers, some researchers call automatic word recognition “the obligatory nature of word recognition” (Roberts et al., 2011, p. 230). In other words, a skilled reader can’t resist reading a word; to see a word is to know what it says. Reading “happens automatically without the influence of intention or choice” (Ehri, 2005, p. 135). To illustrate this idea, consider figure 1.3. Try to name the animal in the picture while ignoring the printed word.
Figure 1.3: Picture-word interference task.
For skilled readers, the task in figure 1.3 is difficult because they recognize the printed word automatically; the word creates a cognitive cue that the reader must resist in order to accurately name the animal in the picture. Similarly, Stroop tasks (named after the researcher who invented them in 1935) list color words (red, yellow, blue, green, purple) printed in nonmatching ink colors (red appears in purple ink, yellow appears in blue ink, blue appears in red ink, and so on). The reader must name the color of the ink while ignoring the printed word; most find this more difficult than picture-word tasks because colored ink is a weaker stimulus than a picture. Tasks such as these provide proof for the idea that skilled readers no longer sound out words but recognize them automatically.
In 2005, reading researcher Linnea C. Ehri synthesized eight theories about the development of automatic word recognition (Chall, 1983; Ehri, 1998, 1999, 2002; Frith, 1985; Gough & Hillinger, 1980; Marsh, Friedman, Welch, & Desberg, 1981; Mason, 1980; Seymour & Duncan, 2001; Stuart & Coltheart, 1988). She concludes that all students learning to read in English progress through four phases of word recognition development: (1) the prealphabetic phase, (2) the partial alphabetic phase, (3) the full alphabetic phase, and (4) the consolidated alphabetic phase. We describe and summarize each in the following sections.
Prealphabetic Phase
During the prealphabetic phase, students read words using visual or context cues. A visual cue is a distinctive feature of a word’s appearance; for example, a student named William might recognize his name because it has “two lines in the middle” (the Ls). A context cue is something that typically occurs in a word’s immediate environment—for example, the golden arches that often accompany the word McDonald’s.
Philip B. Gough, Connie Juel, and Priscilla L. Griffith (1992) examined the nature of context cues by teaching preschoolers sets of four words in which one of the words appears with a thumbprint (see figure 1.4 for an example).
Figure 1.4: Set of words with a thumbprint word.
Students could quickly learn to “read” the thumbprint word (in figure 1.5, lamp) with the thumbprint next to it, but when the researchers removed the thumbprint, they no longer knew the word. When the thumbprint appeared by itself, with no print accompanying it, students successfully produced the thumbprint word (lamp). When researchers placed the thumbprint next to a different word (for instance, stick), almost all of the children produced the original thumbprint word (lamp) rather than the new word (stick). In essence, these prealphabetic readers were associating the target word’s pronunciation and meaning with the thumbprint image, rather than with the letters of the word.
Patricia E. Masonheimer, Priscilla A. Drum, and Linnea C. Ehri (1984) found that preschoolers who could read common signs and labels such as McDonald’s or Pepsi could still read them if letters were altered; for example, if the researchers changed Pepsi (shown with its logo) to Xepsi, children still read it as Pepsi. Even when prompted to look for mistakes, students failed to detect the change. These findings indicate that, during the prealphabetic phase, learners do not associate words with their spellings, but rather with context or visual cues.
Additionally, Brian Byrne (1992) finds that prealphabetic readers attend to print-meaning correspondences, but not to print-sound correspondences. In an experiment, he taught prealphabetic readers that triangle-square meant little boy and circle-square meant big boy (see figure 1.5). Then he showed them triangle-cross and asked, “Does this say little fish or big fish?” Most of these prealphabetic readers were able to match the word little with the triangle and answer correctly (it says little fish).
Figure 1.5: Print-meaning correspondence task.
In contrast, when Byrne (1992) taught prealphabetic readers that triangle-square meant fat and circle-square meant bat, the learners were unable to figure out whether triangle-cross meant fun or bun (see figure 1.6). That is, they could not isolate the /f/ sound and match it to the triangle.
Figure 1.6: Print-sound correspondence task.
Bynre’s (1992) findings indicate that prealphabetic readers understand that printed symbols can correspond to spoken words (which have meaning), but they do not understand that printed symbols correspond to spoken sounds (which do not always have meaning by themselves).
Partial Alphabetic Phase
The partial alphabetic phase, according to Ehri (2005), emerges when students begin to form connections between the sounds of letters and the sounds in words. They stop using visual or context cues and start using phonetic (letter-sound) cues to identify words. To make this switch, students must grasp the alphabetic principle—the realization that there is a relationship between printed letters and spoken sounds (Purcell-Gates et al., 2016). Students typically begin to grasp the alphabetic principle as they acquire alphabet knowledge and phonological awareness (Paratore, Cassano, & Schickedanz, 2011).
Alphabet knowledge involves knowing the names of the letters and the sound or sounds associated with each letter. Research indicates that students acquire alphabet knowledge in a specific order (Paratore et al., 2011), typically starting with letters in the student’s name and letters in frequently encountered words. They usually learn uppercase letters before lowercase letters, and generally distinguish letters with few overlapping features (such as O/E or h/s) before letters with many overlapping features (such as E/F, O/Q, K/X, or m/n). Last of all come letters that differ only in orientation (such as b/d or p/q).