Thursday, November 24, 2005
DIBELS:The Perfect Literacy Test
Language Magazine in press
If Katrina came close to being the perfect storm-in the awful sense of the storm that had all the attributes to do the most harm to the lives of those whose destructive power and irresistible forces it touched, then there is a perfect literacy test sweeping through American schools and doing the maximum amount of damage to the lives of those it touches.
American education has been overdosing on all kinds of tests in recent years as politicians and groups with their own agendas have put pressure on schools to show measurable results for the funding they receive. And many tests have become “high stakes tests” in that they are used as criteria for admissions, promotion, graduation and even wages.
But the perfect test is not like any of the traditional tests in popular use. It is not “norm referenced” like the Iowa or Stanford tests which are widely used to measure achievement. It is not like the barrage of high-stakes state “criterion referenced tests” promulgated to test reading and writing and judge whether pupils can pass from grade to grade or receive a high school diploma. It is not the National Assessment of Educational Progress which has been used to paint dire pictures of whole states failing to produce proficient readers and writers.
No, the perfect Literacy test is the Dynamic Indicators of Basic Literacy Skills developed by a federally funded group at the University of Oregon. It is being widely mandated as part of the No Child Left Behind plan each state must submit to the federal bureaucracy that controls NCLB funding. It’s acronym DIBELS has, according to Education Week, “become a catchphrase in the schoolhouse and the statehouse.” (Manzo, Education Week, 9/28/2005)
What makes DIBELS the perfect literacy test is that it takes total control of the academic futures and school lives of the children it reaches from the first day they enter kindergarten when they are barely five years old. It keeps control of their literacy development and indeed their whole school experience for four years from kindergarten through third grade.. And the more poorly the children respond to DIBELS the more they experience it. Norm referenced tests usually are not given until third grade and then only once a year. Diagnostic tests are usually used selectively with pupils to provide teachers with information on what strengths and weaknesses learners may have. DIBELS, once it gains a foothold, is administered a minimum of three times a year at the beginning, middle and end of each grade from kindergarten to third.
Within a few days of entering school five year-olds have their first opportunity to fail to achieve DIBELS arbitrary “bench marks.” Each month DIBELS is also used to “monitor progress” and those who are marked for “intensive instruction” are monitored weekly. Such tests have sometimes been called test-teach-test models. In that model a pre-test is given, then the content is taught and then a post-test measures gain. DIBELS uses a test-test-test model because increasingly frequent testing is the fate of those who fail to achieve the bench marks of DIBELS. There are reports of children practicing for the DIBELS while they wait in line to use the toilets.
Unlike Katrina whose path and approach could be monitored permitting those who had the means to get out of her way to avoid her dangers, DIBELS has arrived in most of the schools it takes control of as an irresistible force with neither pupils or teachers having any opportunity to get out of the way. That’s because the federal ideologues who have the power to review state NCLB proposals have strongly “encouraged” the use of DIBELS and most states have obliged, some even mandating its use in all of the state’s K-3 classrooms. Every K-3 teacher in New Mexico gets a palm pilot programmed with DIBELS. The scores the testers enter go directly to Santa Fe and the computers at the University of Oregon.
Where DIBELS has been mandated state-wide the only escape from DIBELS is home-schooling which of course is not an option for most working parents. Some parents have down-loaded the whole test themselves and drilled their children at home so they will “bench-mark” in school and avoid the intensive interventions of DIBELS.
According to the DIBELS manual (available on-line at http://dibels.uoregon.edu/ ) “for the 2004-2005 school year, 8293 schools used the Dibels scoring service , across 2582 districts in 49 states and Canada, totaling over 1.7 million students (K-3). It’s likely that its reach has expanded considerably beyond that for the current school year.
So what is DIBELS that it should have such awesome powers? It is a package of sub-tests designed to be administered in 1 minute each. It’s basic premise is that it can reduce reading development to a series of tasks, each measurable in one minute. Each test has arbitrary “benchmarks” which get more difficult to achieve in successive grades. The test authors claim that the sub-tests are “stepping stones” to reading proficiency and each prepares the child for the next test. That means that children who fail one test are failing in reading development according to the authors. And in fact children are being retained in kindergarten and first grade solely because they fail one sub-test in DIBELS. In fact, only a small number of states require children to attend kindergarten. So children entering school without kindergarten are already a year behind from the DIBELS perspective.
Testers who are most often not the children’s teachers are given minimal training and admonished not to deviate in anyway from the procedures or the wording of the tester’s manual. “You’ll be proud of me” said one five year old to her teacher when she came back from being Dibeled.. “I didn’t talk to those strangers”. Her scores were perfect zeroes.
DIBELS provides no time for thoughtful responses. It allows for only one speed- fast. Like a whirlwind DIBELS seizes young children and drives them into each task. Each test is administered with a stop watch in hand. Children are permitted three seconds for each response and the test is stopped at one minute or when the child is wrong on five items. All scores are quantitative and the tester makes no judgements of the quality of the response, so in no sub-test is there any information about how well the child is understanding- and indeed in only one test is there any meaningful text to be read.
Here are the names of the tests and what they actually test in the order that children would encounter them.
Letter Naming Fluency: The child is given a page with lines of mixed capital and lower case letters in a font that is not the most common one in early reading material. The score is the number of letters correctly named in one minute. If the child says a sound instead of a letter she is told “names not sounds”, but only once. Some five year olds respond with the name of a child whose name starts with the letter. No points for that.
Initial Sound Fluency: The child is shown a page with four pictures. The tester says a word for each picture and then asks which picture starts with “buh”. The child must remember the names of the pictures and then abstract out the first sound. The picture may look like a bear but the tester called it a cub. That big yellow grasshopper was called an insect. Is that picture a frosted donut or a bagel with cheese on it? The score is the number of right initial sounds the child can say in 1 minute.
Phonemic Segmentation Fluency: The tester has a sheet with one syllable words. If the tester says “cat” the child must respond kuh- ah- tuh in a few seconds. One point for each correct sound produced in one minute. Mismatches between the dialect of the tester and the child certainly effect the score.
Nonsense Word Fluency: The child has a sheet with what are supposed to be two or three letter “make-believe” words. The tester tells the child to either say the whole word or each sound. In either case the score is the number of sounds right in one minute. In this test children already reading are handicapped because many of the nonsense words are either possible spellings of real English words or actual words in English or Spanish. There are stories of teachers making nonsense bulletin boards so the children can practice reading nonsense.
Oral Reading Fluency. Starting in first grade the children are given a five paragraph essay on a topic written in first person. The score is the number of words read correctly in one minute. The children learn to skip any words they don’t know and say the words they know as fast as they can. The tester says any word the child stops at after a few seconds. Some children use that as a signal that they should wait for the tester to say the word before proceeding. And a minute goes by very rqapidly.
Oral Retelling Fluency. Teachers complained that counting correct words didn’t show what the children understood. So the DIBELS folks added an oral retelling. The score is the number of words the kids produce in one minute that are more or less on topic. No attention is paid to the quality of the retelling. Honest.
Word Use Fluency ; Starting in kindergarten, the tester says a word and tells the child to “Use the word”. The score is the number of words the child uses in using the words in one minute. It’s hard to see what this would have to do with reading since no reading or print is involved.
Notice that each test name includes the word fluency. How can one be a fluent word namer or sound sayer? Apparently fluency to the Dibelers means speed and accuracy.
There are many things wrong with DIBELS.
It turns reading into a set of abstract decontextualized tasks that can be measured in one minute. It makes little children race with a stop watch.
It values speed over thoughtful responses.
It takes over the curriculum leaving no time for science, social studies, writing, not to mention art music and play.
It ignores and even penalizes children for the knowledge and reading ability they may have already achieved..
Reading is ultimately the ability to make sense of print and no part of DIBELS tests that in any way. In DIBELS the whole is clearly the sum of the parts and comprehension will somehow emerge from the fragments being tested.
On top of that the sub-tests are poorly executed- the authors do badly what they say they are doing. Furthermore the testers must judge accuracy, mark a score sheet and watch a stop watch all at the same time. And, to be fair, testers must listen carefully to children who at this age often lack front teeth, have soft voices, and speak a range of dialects as well languages other than English. Consistency in scoring is highly unlikely among so many testers and each tester is likely to be inconsistent.
And lets add that DIBELS encourages cheating. There is a thin line between practicing the “skills” that are tested and being drilled on the actual test items, all of which are on-line to be downloaded.. With so much at stake why wouldn’t there be cheating?
In summary DIBELS, The Perfect Literacy Test, is a mixed bag of silly little tests. If it weren’t causing so much grief to children and teachers it would be laughable. It’s hard to believe that it could have passed the review of professional committees state laws require for adoption of texts and tests . And in fact it has not passed such reviews. There is strong evidence of coercion from those with the power to approve funding of state NCLB proposals and blatant conflicts of interest for those who profit from the test and also have the power to force its use.. A congressional investigation is now underway into these conflicts of interest .
In training sessions for DIBELS teachers are not permitted to raise questions and are made to feel that there is a scientific base to the test they lack the competence to understand. It is , after all, The Perfect Literacy Test.
Ken Goodman, Professor Emeritus
Language, Reading and Culture,
University of Arizona Kgoodman@u.arizona.edu
Ken Goodman is a researcher and teacher educator in language and literacy. He is past-president of the International Reading Association and the National Conference of Research in Language and Literacy. His reading miscue research and model of the reading process have won a number of national awards. His books include On Reading, Phonics Phacts ( both Heineman) In Defense of Good Teachers (Stenhouse) and Saving Our Schools (RDR Books)