|
SLS Programs: | |||
 |
||||
| Stevenson Reading | |
|||
|
||||
| Cursive Writing | ||||
|
||||
| Essential Grammar | ||||
|
||||
| Other Programs | ||||
|
||||
|
Latest News | |
 |
||
| Workshops & Conferences | |
|
|
||
| Stevenson Forums | |
|
|
||
| FAQs | |
|
|
||
| Our Newsletter: "Food for Thought" |
||
|
||
| Consulting | ||
|
||
| Teaching Resources | ||
|
||
| Links | ||
|
||
Recent Article from Spring 2007:
Some Interesting Test Results from Ohio
Anyone who has seen the Test Score section of our Background Information Folder knows we have some impressive scores demonstrating the effectiveness of the Stevenson Program. However, some test results are more interesting than others. Recently we received some DIBELS (Dynamic Indicator of Basic Early Literacy Skills) scores that, although not definitive, captured our interest.
Laurie Pottorf is an excellent, dedicated teacher now working in South Carolina. She recently shared with us the results of an experiment she did as part of her Masters research at the University of Dayton. Laurie was teaching in Dayton last year in a resource room with urban students, who were all functioning at a similar academic level and who were all diagnosed as having some degree of learning difficulties. Laurie divided them into two comparable groups and administered the DIBELS Second Grade Benchmark Assessment I, Nonsense Word Fluency as a pre-test. This test requires students to decode a series of nonsense words presented in typical short vowel (closed syllable) structures (CVC or VC). She performed a simple experiment by using Stevenson clues with one group and conventional phonics instruction with another, then administering the same test to both groups two more times.
Prior to the experiment, all students had some exposure to both the general education, phonics-oriented basal reading system and to the Stevenson Program. All students had encountered several different types of words (i.e., closed syllable or short vowel words, long vowel silent e words, etc.), and they all exhibited varying degrees of decoding problems. On the day following the pretest, Laurie taught half the students, whom we will call Group A, the Lonely Vowels clues from the Stevenson Language Skills Program. These clues are used to teach short vowel words. The other half, Group B, were taught a more conventional approach to the short vowel patterns. Both groups were then immediately tested again using the same section of DIBELS. After a two week interval and a brief refresher course in short vowels (using the respective approaches), the groups were tested again.
The results were intriguing – for several reasons. First, we believe the results tend to support the use of mnemonics in phonics and the effectiveness of the Stevenson Program. Second, the experiment raises some questions about the assessment process in general. Third, the first two points make it clear that it would be helpful to collect results from other similar, but more extensive, experiments.
Regarding the first point, the outcome that stood out to us was this: the decoding accuracy of Group A, which received the mnemonic instruction with Stevenson, improved dramatically, while the decoding accuracy of Group B, which received the conventional phonics instruction, did not improve at all. Then, after two more weeks, Group A retained their gains in accuracy and even improved slightly, while Group B displayed a minor amount of deterioration.
The results reinforce what we have long believed to be true: teaching phonics with mnemonics is more effective than teaching phonics in conventional ways. Certainly there is a large body of evidence from published, controlled studies (a bibliography of which is supplied in our Background Information Folder) that prove how effective mnemonic instruction is. However, only some of those studies focused on phonics, which is one reason why Laurie Pottorf chose to explore this area as part of her Masters work. As we previously stated, these results are not definitive. The total number of students involved is small. The actual instruction time (both mnemonic and conventional) before retesting was brief, and another follow up beyond two weeks would be good. While more extensive experiments would be welcome, we believe some numbers are telling. The Stevenson group experienced a growth in accuracy from 75.7 percent to 95.9 percent (20.2 percentage points or 26.7 percent of the base) in a short period of time. The conventional phonics group did not improve their accuracy at all. Experienced teachers know that students need to read with more than 90 percent accuracy to be competent readers.
It is important to note a certain point, which leads to our questions about the assessment process: this particular DIBELS test was not designed to measure accuracy rates. In this test, students look at a series of nonsense words (e.g., vov or lig) and produce sounds for the letters. They are scored on the number of correct sounds produced in one minute. For the score it does not matter whether pupils sound out one letter at a time (e.g., /v/ /o/ /v/ for vov) or a group blended together (e.g., /vov/ for vov). It also does not matter what the accuracy rate is. In order for us to present the accuracy rates described earlier, we had to extrapolate them from the raw data. The actual DIBELS scores showed small effects. In this test, it is theoretically possible for a student to attempt 80 letters, miss half of them and score higher than a student who attempted only 35 letters but got them all correct. This, and other DIBELS tests, measure decoding fluency. Of course, a student who decodes very slowly will have difficulty reading with good comprehension; however, a student who decodes with a low rate of accuracy will also not be able to read in a meaningful way.
DIBELS is a very popular and very useful assessment. It highlights particular aspects of the reading process effectively, particularly in the areas of phonemic awareness and decoding. It is easy to administer, and it is available free on line. The Nonsense Word Fluency test is only one of five measures that DIBELS offers, but it is used frequently in many schools at key junctures in a students’ development. Therefore, we should be careful about how the scores are used.
Today educators know they must teach phonics and also keep data on their students’ progress. With DIBELS, an experienced teacher can gather a great deal of useful information about a student’s strengths and weaknesses. However, the data must be placed in context. Laurie noticed some students in both groups intentionally read more slowly after they received the instruction that followed the pretest. They appeared to be trying harder to blend sounds and figure out words. Therefore, in several cases, a student’s official DIBELS score went down while his or her accuracy improved. We have heard heated debates about the relative importance of speed and accuracy, but any reading teacher knows you need both. In our data driven world, it is essential that people who are experienced with reading instruction interpret the data if it is actually going to help the student.
At Stevenson Learning Skills we are often confronted with test related issues. The fact that we teach long vowel words before short vowel words creates short term challenges. (To learn more about this subject request our free information piece entitled Thoughts on Long Vowels Short Vowels and the Stevenson Program.) Of course, in the long run, no one can be functionally literate until they can handle both long vowels and short vowels, as well as diphthongs and many other structures. Stevenson does cover all the major patterns (in a unique order), and that is why we can produce the impressive test results given in our Background Information Folder.
Still we would like to collect more data. We want to demonstrate how specific elements of the program can produce specific positive effects, and we want to identify appropriate assessments to measure reading progress. The results that Laurie Pottorf shared may not be conclusive, but they are very valuable, and we would like to duplicate and enlarge the experiment. It would be good to have more data from DIBELS because it is such a common assessment tool. We are also interested in data from the DAR, the TPRI and other instruments. For the data to be useful, control groups are required.
Unfortunately, we cannot perform the experiments ourselves. We publish materials and offer training, but we do not have students of our own. Also, if we perform the experiments ourselves, the results are suspect. However, we are eager to help out any schools using Stevenson who can provide the data and the necessary conditions. If you have any ideas, please contact us at 800-343-1211, M-F, 9-5, Eastern Time. Also call us if you want to examine the data discussed in this article. We will be happy to send you a description of the experiment and a table of results.
|
|
 |
|
©
2002 Stevenson Learning Skills, Inc. All rights reserved.
Stevenson Reading • Cursive Writing • Essential Grammar • Other Programs News • Workshops/Conferences • Forums • FAQs • Our Newsletter • Consulting Services • Ordering/Catalogue About Us • Contact Us • Site Map |
