How Reading Assistant Down-Levels
- Forrest Rose
- Jeff Gregson
When students read the Benchmark assessment, the intent is for the student to achieve a sufficient amount of “productive struggle.”
Every student begins the Benchmark assessment by reading a grade-level passage. If a student struggles to progress on the first page of the passage, Reading Assistant prompts the student to try a new story and presents a down-leveled passage, which is one grade level below the current passage.
An algorithm determines the new level. The conditions for immediate down-leveling are:
Failure to make sufficient progress for 1 minute
Reading at an accuracy rate of 60% or below
As the student reads the following passage, Reading Assistant will continue to down-level using the criteria above.
The kindergarten passage (which contains CVC words and Dolch high-frequency words) is the lowest-level passage a student can be given in Reading Assistant. If a student has trouble completing the passage, even at the kindergarten level, encourage the student to try his/her best to sound out words, read the known words, skip unknown words, and/or say the letters he/she knows. As long as some effort has been made for one minute, Reading Assistant will give the student credit for the reading session.
Students who cannot complete the kindergarten passage are given the assessment status of Early Reader on the Tracking Report. Students with the Early Reader status will automatically be placed in the Early Reader Skills Scaffold the next time they log into Reading Assistant.
During the ORF Assessment, Reading Assistant tries to discern if the child has enough momentum to get through the passage. She does not want a student to reach a significant frustration level, but she also wants to provide teachers with grade-level data. To strike this balance, Reading Assistant will typically down-level the student in the following cases:
The student is struggling to get through the first 6-7 words (after 60 seconds)
The student has struggled to read more than 40% of the words after the second page of the story
If the student is reading at a significant pace but missing several words, Reading Assistant will err on the side of continuing to see them through the passage. Allowing them to continue attempting to read the passage also enables Reading Assistant to gather the most data possible to provide the most accurate results of a student's abilities. If, after reviewing the results of a benchmark assessment activity, you believe that the student would benefit from being placed in the Early Reader Skills Scaffold even if Reading Assistant did not place that student in the ERSS automatically, you can do so via the Tracking Report.
How are my student’s scores affected if they get down-leveled into a passage that is lower than their grade level passage?
Reading Assistant down-levels students from an on-grade passage when the student fails to read a meaningful number of words in that passage. Down-leveling occurs so that the Assessment can gather enough information to provide helpful insight to the teacher.
Different test forms (meaning different stories in the case of Oral Reading Fluency passages) have various difficulty levels, which impact a student’s ability to read accurately and quickly. For example, a passage at a one-grade level lower is likely more straightforward for the student to read and comprehend than an at-grade passage. Variations in passage difficulty will result in variations in raw student scores. Without adjustments, Reading Assistant will produce scores that are comparatively over-state or under-state student ability.
Therefore, Reading Assistant’s psychometric models make adjustments to the raw WCPM scores to account for the difficulty level of any given passage in relation to other passages. This way, the adjusted WCPM scores account for the fact that a “down-leveled” student is reading an easier passage than their peers who tackled the on-grade passage. In other words, a 2nd grader who reads the 2nd grade passage at 30 raw WCPM will end up with a higher adjusted WCPM score than a 2nd grader who reads the 1st grade passage at 30 raw WCPM. The magnitude of this adjustment varies depending on the grade and screening window, and each grade/window’s adjustments are derived from a combination of Reading Assistant’s empirical data on the passages, measures of text complexity of each passage, and historical data on each passage collected by TPRI.
To provide a ballpark sense of the magnitude of this adjustment, a 2nd grader reading the Winter (MOY) 2nd grade passage at a raw WCPM of 30 can expect their raw WCPM to be adjusted about 12 points higher. In contrast, a 2nd grader reading the Winter (MOY) 1st grader passage at a raw WCPM of 30 can expect their raw WCPM to be adjusted about 2 points lower. Therefore, at the same raw WCPM of reading, the 2nd grader reading the grade level passage will effectively have an adjusted WCPM 14 points higher than a 2nd grader reading the 1st-grade passage.
The plots below illustrate this adjustment. They represent histograms of the WCPM scores for 2nd graders taking either the grade level story (blue) or the 1st grade story (orange). The left plot shows the distribution of the raw (unadjusted) WCPM scores for the two populations. In contrast, the right plot shows the distribution of the adjusted WCPM scores, where the adjustment accounted for the difficulty level of the story. The plots show that the center of mass of the distribution for on-grade-level readers and down-leveled readers is much more discriminable after the adjustment.
Looking at the Adjusted WCPM graph on the right side, it is important to note that there is still an overlap in the adjusted WCPM scores of students who complete the down-leveled story and those who complete the grade-level story. Down-leveling is something that occurs when the software makes a live judgment while the student is taking the assessment. It will likely be able to generate a more psychometrically valid score if the student reads an easier passage than the one they are currently on.
The adjustment in WCPM does not guarantee that the student will receive a lower score than that of all students who take the grade-level story. The only guarantee is that the adjustment applied to grade level vs. down leveled story scores is appropriate to account for the difference in text reading difficulty between the two passages. Our research indicates that down-leveled students are sometimes better readers than those not down-leveled because:
The student did not try hard or engage with the initial passage
The student gave up after encountering a few hard words
The student was just below the cut-line for success on the on-grade passage and then demonstrated strong fluency with slightly less text complexity.
So, while down-leveled students nearly always (about 75% of the time) score below the 30th PR for their grade level, there are occasional exceptions.
Support
Email: Submit a Case
Phone: 800-225-5750, Option 3 (Technical Support)