The Scanning Task demonstrates perceptual effects on a reader's ability to recognize words. Perceptual effects observed in the task include word-length, word-superiority, and word-frequency effects. These effects are described below, followed by a description of how the scanning task items have been constructed to generate these effects, then followed by a review of a typical set of results in which the three effects can be observed.
There is strong evidence to suggest that readers process words one letter at a time. As a word's length increases the time we spend fixated on it also increases. Readers spend approximately 30 extra milliseconds fixated on a word for each additional letter (Just & Carpenter, 1980, 1981;). But, this pattern changes when words or letter patterns become so well learned that they are recognized more quickly as visual wholes rather than individual letters.
In the Scanning Task the speed advantage scanning for words has over scanning for non-words, and that scanning of non-words has over scanning for letter strings, is the result of a reader's ability to identify visual patterns in text. Words and word patterns are so well learned that they are relatively easy to spot as a visual "chunk". Non-words on the other hand are not so recognizable, but they do contain letter patterns we often identify as groups (such as "tion"), and we see them visually much like we see words. Letter strings on the other hand can be down right difficult to find because the letter patterns are generally new, not corresponding to any well-learned word or letter pattern. The relative advantage scanning of words has over non-words, and non-words has over strings has been termed the "word-superiority effect". The primary assumption of the word-superiority effect is that words and parts of words are stored in long term memory as whole patterns, rather than individual letters. These stored letter patterns allow for quicker perception of familiar words than would be the case if individual letters were perceived one letter at a time.
The ability to perceive words is not always improved by the familiarity of letter patterns, as might be predicted by the word-superiority effect. Very high frequency words, such as the word "the", are often more difficult to find than less frequently occurring letter patterns. Word frequency refers to the number of times a word appears in a very large sampling of English text that is typical of what an average reader might encounter in everyday reading, over their lifetime. While frequency effects are usually associated with frequency of meaning (the location of the word "the" is easy to predict from the surrounding meaning), the frequency with which the letter pattern itself appears visually in English text also seems to affect how easy or difficult they are to consciously recognize. Presumably within the context of the Scanning Task, these higher frequency words are perceived with little conscious effort, if any, thus making them more difficult to find than less frequently appearing letter patterns.
As a brief experiment that demonstrates word frequency effects, read the following passage. With a pencil and paper at your side, as you read record the number of times the letter "t" appears with a tick mark. Read to understand the passage. You'll be tested.
|
It was the first of the month. When the bell rang, most of the room emptied, clearing before the speaker could finish the the entire speech. Fortunately the point of the speech had been made, so the effort was not in vain. |
Actually, you won't be tested. You were instructed this way to reduce the chance that you would methodically scan the paragraph for t's, in which case you were not really reading. How many times does the letter "t" appear? Most people will incorrectly count 18 or 19 occurrences. There are 21. If you didn't count 21, scan (don't read) over the passage to find letter t's. Does scanning without paying attention to meaning increase the number of t's you find? In most cases when readers miss the letter "t", they miss it in the word "the", even after being told of this fact. The word "the" appears so frequently within English text and is so highly recognizable and predictable, that we often don't fixate on it when reading, but rather assume it's there based on the meaning in the surrounding words. When we do fixate on it, it is likely that so little effort is required to process it, that we don't consciously perceive the word.
Now read (don't scan) the paragraph again but this time as you read count the occurrences of the letter "f".
A similar effect occurs when searching for the letter "f" as occurs when searching for the letter "t". Readers will often miss the letter "f" in the word "of". This is likely due, in part, to the frequency with which the word "of" appears in English text, but in this case the pronunciation of "f" is not the typical pronunciation. It sounds like a "v", deceiving readers who use a verbal strategy to search for the typical "f" sound. There are nine occurrences of the letter "f" in the passage. How many did you find?
Finally, did you catch the duplication of the word "the" at the end of line three and the beginning of line four? If you caught it, was it when you were reading or scanning the paragraph? Readers tend to skip the beginnings and ends of lines as they read, particularly where the beginning or ending word is highly predictable.
These simple experiments demonstrate the effect of word frequency, and the effect of context on our ability to perceive words. With grammatical and semantic context (syntax and meaning) skilled readers perceive about 80% of content words and about 40% of function words (Just & Carpenter, 1987). The rest of the words are predicted, or filled in based on the meaning in surrounding words. As a result of context effects the highly predictable words are often skipped over while reading.
The Scanning Task is not a reading task, like the one above is (arguably). It is a pattern or word recognition task, differing considerably from normal reading because effects of meaning and syntax have been removed, both key features of skilled reading that mediate our ability to perceive words. The Scanning Task demonstrates the perceptual effects of word recognition, independent of the processes associated with meaning and syntax.
Paragraphs, or text blocks as described in the Scanning Task, have been extracted from a number of post-secondary level readings. The text blocks and target letter patterns are manipulated in different ways to produce 12 paragraphs made up of either words, non-words, or letter strings, controlling word length for target patterns, and for actual words, controlling word frequency.
Word length, word-superiority, and word-frequency effects can be observed by reviewing results in Table 1. The top portion of the table lays out the raw results for each item in the two exercises, while the bottom portion lays out the average reaction times for each exercise and each type of letter pattern (i.e.words, non-words, and strings).
Effects for individual users, though likely to follow a typical pattern like that in Table 1, will not always follow that pattern. Typically results will show a relatively strong word-superiority effect and an interaction between word length and word frequency across the two exercises. Because of the small number of items in the exercises, a "lucky scan" or two can throw off results significantly. Some times a user's eyes just happen to fall on the right spot right away. Review others results if you had any lucky scans, and notice the typical pattern that emerges in their reaction times.
Table 1: Results for User 2005
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
In the bottom of Table 1 the effect of word-superiority shows up in a comparison of the average reaction times for words, non-words, and letter strings in Exercise 1, and in a comparison of non-words and strings in Exercise 2. This typical word-superiority effect breaks down however with an interaction between word length and word frequency. In the Scanning Task the combined effect of length and frequency is demonstrated by comparing the average scanning times for words in Exercise 1 with those in Exercise 2.
In the Scanning Task the predictability produced by the surrounding meaning in which words appear has been removed by having readers scan, rather than read the text. The surrounding meaning and syntax are assumed to be ignored (at a conscious level at least). Word frequency and word length on their own, independent of meaning and syntax, affect how we perceive words.
Take a few moments to study the raw results in Table 1 and notice the times reported in the "Scantime" column for words, non-words, and strings. Also take a few moments to review the responses to the questions in the questionnaire that followed the Scanning Task. Take note of how users think while doing the task. It is interesting to compare the thoughts of users on this task with those of users who completed the questionnaire following the Stencil Stacking Task. Read through the pages accompanying the Stencil Stacking task to learn more about strategy use and metacognition.
The goal of this exercise has been to understand perceptual aspects of word recognition. Generally, the longer words are, the longer it takes to perceive them. But this is not always true. As words and letter patterns become more recognizable to readers as they become more skilled, the conscious effort required to perceive them decreases. Presumably as the effort to perceive words decreases more mental space becomes available for comprehension, and other metacognitive activities associated with learning from text.
-------------------------------------------