Explain Test Percentile Scores to me please.

[kajm]

We completed Woodcock-Johnson testing recently and I received percentile scores. This is the first time we have had the percentiles and I don't quite understand them. For instance, if they score a few grades above level in a subject their percentile score may be a 73.0, but if they are right on grade level in a subject, their percentile score may be a 51.0.

 

Obviously, this must not be the same percentile score that gives an A, B, C, D or F grade. But what does it mean if their percentile score is 50 or 99? Percent of what?

 

Please help my ignorance on this subject!

[Momling]

When your kids were little, did you take them to the pediatrician who told you they were in the 75th%ile for height (or something like that...)?

You can think about it like this: if you were to line up 100 kids from the shortest to the tallest (with #1 being the shortest and #99 the tallest), your child would have been number 75 in line. That is, they are taller than 75% of the population.

Same with test scores. A score in the 73rd %ile is better than 73% of the test takers at that grade level. Most likely, the 73rd percentile would be a higher grade level equivalent than a score in the 51st %ile, which would be pretty close to average.

[bolt.]

Percentile is "of the people who take this test" such-percent got a score infirior to yours. 52% means that you scored better than half of all test takers, but 48% of test takers scored better than you. The 'average' test taker has 50% above them and 50% below them -- it is entirely possible that all the test takers are "on grade level" or that only 10% of test takers are getting unacceptable grade-level scores. 99th percentile would mean only 1% of all test takers scored better than you.

 

Percent is "of the questions asked or points awarded" you earned such-percent of the total available points for correct answers. 52% means that you only know slightly more than half of the material that was being tested, that 48% of what they asked, you got wrong. 99% means that only 1% of the test material was unknown to you.

 

Very different measures.

[Sebastian (a lady)]

We completed Woodcock-Johnson testing recently and I received percentile scores. This is the first time we have had the percentiles and I don't quite understand them. For instance, if they score a few grades above level in a subject their percentile score may be a 73.0, but if they are right on grade level in a subject, their percentile score may be a 51.0.

 

Obviously, this must not be the same percentile score that gives an A, B, C, D or F grade. But what does it mean if their percentile score is 50 or 99? Percent of what?

 

Please help my ignorance on this subject!

 

The other thing that comes into play with standardized test results that give both a percentile and a grade level is that the grade level assignment means that the child did as well answering the questions given as a XX grader would have done on those same questions. It doesn't mean that they are doing XX grade work.

 

Let me give an example. My kid takes a test at the 4th grade level. His results might say that his answers were 6th grade with a 90th percentile. That means the average 6th grader would have answered the 4th grade questions with the same right/wrong results.

 

The percentile would indicate that he scored better than 90% of all the 4th grade students who took the test. It has nothing to do with the percentage of items he answered correctly.

 

Now Woodcock-Johnson may be a little different, because I don't think there are different grade levels, so much as a long continuous battery. Even so, you would expect that in areas where they were ahead of grade level, they would score above average (50%) and areas where they score around the middle percentile they would be at grade level.

 

Your test report should have some kind of section on understanding the score. Maybe on the back side of the report or in a separate handout. If not, I would go back to your test provider for more info.

 

One thing that I have to remember with standardized tests is that half of the students taking the test will be below average. It's not like a spelling or math test where the student studies the material and you are expecting they will demonstrate mastery with a score in the 80-100% range. 50% could be quite normal for a kid who is moving on pace with his grade peers. It wouldn't indicate something to be concerned about.

[kajm]

Thank you everyone. I think I'm getting it, but like Sebastian said, I don't believe the Woodcock-Johnson is based on what grade the child is in when they give the test. They simply keep asking questions until the child struggles too much. So in that case would the saying "he's doing as well as an 8th grader on a 4th grade test" apply? I'm just trying to understand and am obviously in a brain fog. Help!

 

But I think I do get the fact that the 50th percentile score means that they are right in the middle of the pack and where it would be expected they would be for their age.

 

I loved Momlings analogy with the pediatrician and percentile scores for height. That made so much sense to me and I could relate to that. I also think I understand that a higher than 50 percentile score doesn't necessarily mean they are performing at a higher grade level, they simply could answer the questions asked as well as someone would who would be at X grade level, right?

[Pamela H in Texas]

The WJ is different from the CAT, for example. With CAT, if it says 4.3 grade level, it means your kid did as well as a child that was in the 3rd month of 4th grade would do on that test whether it was a 2nd grade or 7th grade test. If it says 4.3 on the WJ, it means your child was judged to be at the level of a person in the 3rd months of 4th grade. They are two totally different kinds of tests. WJ pinpoints at what grade level the child is where the CAT compares students who take that certain level test.

[Alessandra]

Thank you everyone. I think I'm getting it, but like Sebastian said, I don't believe the Woodcock-Johnson is based on what grade the child is in when they give the test. They simply keep asking questions until the child struggles too much. So in that case would the saying "he's doing as well as an 8th grader on a 4th grade test" apply? I'm just trying to understand and am obviously in a brain fog. Help!

 

 

 

As pp have said, the test is not for a specific grade level -- I believe it is for ages 2-90+. The sampling is excellent, so accuracy of results are reliable.

 

As for the bolded -- examiners determine a basal level and a ceiling level. For a given subtest, the examiner begins at where they think child should start and asks 6 questions. If 6 questions in a row are answered correctly, the examiner goes on until 6 questions in a row are missed, with the last correct answer being the ceiling. On the other hand, if the child cannot answer the first 6 questions correctly, the examiner goes back (page by page, if I understand correctly) until 6 questions are answered correctly and then continues with the test.

 

The idea is to have the examiner make an accurate assessment with as few questions as possible.

 

I like a book in Wiley's Essentials series, Essentials of WJII Tests of Achievement.

 

Btw, I think it's great your ds did that test. The subtest scores and the examiner's comments can be a gold mine of info that will help you understand your dc.

[kajm]

You guys are great - thank you! It is so good to know the difference in some of the tests and that they do test differently and give results in a different fashion. This is only our 2nd year, so I'm learning more all the time.

 

Thank you also for that book suggestion. I had no idea such a book existed, so I will definitely look into it.

 

Do anyone know, is there a certain test that is "better" than other tests or that future colleges will look at with more favor, etc.?

[Sebastian (a lady)]

Colleges will use the ACT or the SAT reasoning (which is no relation to the Stanford Achievement Test) and any SAT subject tests or Advanced Placement tests. They won't be interested in standardized tests taken in elementary or middle school.

 

So pick what best suits your family and state requirements.

[Aslana]

Hmm, just to throw a wrench into this explanation, as I understood it from a teacher friend, saying someone scored on a 4.3 level, means that person scored on a 4th grade, 3rd month level NOT that they share a level with anyone else.

 

Meaning, if you have your 4th grader take a test marked for 4th graders and he scores a 4.3 on the math (let's say), then your 4th grader is scoring on level or behind (depending on when in the year he takes this test. If he takes it at the beginning of his 4th grade year, he's 3 months ahead of a typical 4th grader. If he takes it at the end of the year, he's actually behind the typical 4th grader) other kids but he isn't compared to them because comparing them is the percentile score.

 

So, 4th grader takes a complete 4th grade test. He scores a 4.3 on the overall test with a 51% percentile score (ignore any bad math, just throwing examples out). This means that YOUR 4th grader:

 

is at a 4.3 (4th grade year, 3rd month) of his learning and is behind 49% of the 4th grade population ( out of 100, this means he's #51 on the list. 50 students did better than him and 49 students did worse). So that 4.3 only tells you where he sits normally in a school 4 th grade class, not that he's 4.3 ahead of anyone.

 

 

But here's the kicker--you can have your 4th grader take a 5th grade leveled test. If they scored that 4.3 on that 5th grade level, it doesn't mean they are behind! It only means he's not ready for 5th grade yet. You only have to worry about a child being "behind" anyone when you have that child take an on grade-level test and it's the end of the school year and they score a grade level that indicates beginning of school year (that 4.3 instead of 4.8).

[Alessandra]

Do anyone know, is there a certain test that is "better" than other tests or that future colleges will look at with more favor, etc.?

 

 

I can't answer your question about what tests colleges prefer, although I think most colleges would use widely taken tests, like the SAT, for general admission. Since WJ is an individually administered test, there would be a significantly smaller population that takes it -- so there would not be much basis for comparison. I am not sure, however, if WJ comes into play for any kind of special needs admissions.

 

Even though WJIII has 'Achievement' in the title, it is primarily a diagnostic test. The 'Achievement' label sets it apart from tests of cognitive abilities (like IQ tests). WJ can help tell you/teacher/therapist what to focus on. For example, a kid with reading difficulties who scores very highly on word identification or word attack may be better served by a vocabulary or comprehension program than by an Orton program.

 

WJIII is highly respected. The sampling is excellent, so that the accuracy of the scoring is very high. It is also widely used, so if you move to another area, you should easily be able to find people to administer it. And since professionals are familiar with the scoring, they will understand what your dc's individual needs are more quickly than they might with an unfamiliar test.

 

I put off buying the Essentials book for several years. It was expensive & not in my library system. But it was very much worth the money to me.