More Evidence that Good Instruction Can make a Big Difference

Last spring I reported on the Gering School District (Nebraska) which had taken their Reading First grant back in 2004 and, with the help of NIFDI, implemented the full-immersion Direct Instruction for Reading and Language Arts back in 2004 district wide.

Let's see what the results look like four years later now that that first cohort has reached 3rd grade. We can also take a look at the performance of the 4th - 6th grades who participated in the program for some of their school years. Also, we can take a look at the 7th and 8th grades that received a remedial version of the reading program only.

As you can see from my fancy Excel 2007 graph below, Gering students are now performing better than the Nebraska state average.

The blue bars on the left represents the mean performance of all students in Nebraska on the Nebraska state NCLB test (a test with an artificially high pass rate to be sure, but we are concerned here with relative performance). Nebraska is about 70% white and 25% Hispanic so the average performance of all students comprises mostly white students.

The red bars represent the performance of all Gering students which is higher than the Nebraska average in all grades. This is significant because the demographics of Gering are below the Nebraska average. (Adults are less educated and there is more poverty.)

Now let's take a look at the subgroup data to see who is making the progress.

The green and purple bars represent the performance of Hispanic and white students respectively. Both groups are performing better than the state average for all students in all grades but grade 5 where Hispanics perform slightly below. Gering Hispanics are actually performing better than Gering white students in grade 7 which is something you rarely see on a district wide basis and/or without selective admissions.

The light blue bars (on the right) represent the performance of students receiving free and reduced price lunches -- a proxy for poverty. Poor students in Gering perform better than the Nebraska average for all students (not just poor ones) with the except of grades 5 and 7. This is something you don't see everyday. I've run quite a few analyses like this for different states and I have yet to find an entire school district whose low-SES students performed above the state average for all students.

Apparently the effects of poverty and being a non-Asian minority can be mitigated through good instruction alone.

I challenge anyone to produce similar effects for their favored non-instructional intervention (on a District-wide basis without selective admissions) that produce effects similar to those found in Gering in grades 3-8.

I'll drill down deeper into the data in a future post.

When is a decoding error not an error?

I recently came across the Critical Reading Inventory, a balanced literacy diagnostic, which purports to assess reading ability.


In case you were still under the misapprehension that balanced literacy was a state-of-the-art blend of the best of code-based reading instruction and the best of whole language theory, the Critical Reading Inventory should help disabuse you of that notion.


The problem with balanced literacy is that even though it typically includes phonics-like activities, it is not based on alphabetic code-based reading instruction. The dirty little secret behind balanced literacy is that for all the lip-service that's paid to phonics, in balanced literacy the decoding of text using the alphabetic code can be replaced with the guessing of word identity based on non-phonological information. In balanced literacy, decoding errors can be ignored if the reader guesses a word that means about the same thing as the word that the reader is unable to properly decode. That doesn't sound very balanced to me. Doesn't sound like decoding either.


The Critical Reading Inventory helpfully makes public their scoring assistant website in which teachers learn the black art of miscue analysis (i.e., ignoring decoding errors). Two case studies are provided laying out all the gory details. (Beware this is not for the faint of heart.) Let's take a look at a scoring sheet from the second case study in which "John," a fifth grader, struggles to decode a third grade text.

(Click on the graphic to see a full sized version.)

As you can see, John has made numerous (over 20) decoding errors half of which have been ignored because of the miscue analysis. All the errors with a plus sign (+) in the margin have been ignored because John's errors didn't affect the meaning of the text.

This is how you make it seem like a student has learned how to read when he really hasn't yet. But doesn't it all look so scientific and professional. We have "case studies" which sounds like science. And science is good. right? And we have fancy jargon like "miscue analysis" which makes it seem so professional like these educators really know what they're doing. You can trust us with your child's education.

And, that's just the decoding part of the test, wait until we get to the comprehension part.

All the Cool Kids Are Doing It Now

Uncle Jay is blogging.

Reminds me that I have to do a post on Jay's latest book on KIPP which I enjoyed.

Any other new bloggers to recommend?

What I learned this year

Another school year down. One child completed kindergarten. One child completed third grade. A ptotracted negotiation with my school district over my third grader's GIEP. A hundred blog posts in the hopper. Innumerable education articles read. So what, if anything, did I learn this school year educationwise?

The concept of reading "comprehension" should be abandoned.

There do not seem to be any good instruments for testing reading comprehension that are valid, reliable, and objective and that actually measure how well the student is able to comprehend text.

There is insufficient content instruction in the late elementary grades. Many students are fluent decoders at this level. Too much time is wasted on learning reading strategies and reading simple fiction books instead of learning content knowledge which we know many low-SES students lack and is somewhat easy to teach.

There's a lot of fat in math curriculum from fourth grade until algebra begins (typically in eighth grade). Half the year is typically devoted to review.

Despite all this math review, many students fail to learn elementary math as demonstrated by standardized tests which by and large test nothing but elementary math.

Since when did being able to write a short essay about how you solved a math problem become the end game of elementary math. This seems to be testing the student's reading ability twice rather than the student's math ability. See Reading Comprehension, Not Knowing How to Teach above.

History, geography, science, and civics instruction continue to get short shrift.

Many schools practice education socialism in kindergarten by not formally teaching reading at least to the at-risk students in favor of not teaching anyone in the hope that the at-risk will mature and to slow down the capable.

Students are grouped instructionally for teacher convenience rather than student need. The at-risk and special education students are spread out so no teacher is over-burdened. Same deal with the more capable students.

Instructionally differentiation is largely a waste of time. The more capable students are given make-work and other sundry time-wasters too compensate for an instructional pace that is too slow for them. The less capable students get a watered down version of the curriculum. Is it any wonder why they never catch up?

School Districts don't always follow the regulations when the regulations conflict with how they want to run their school.

Education policy makers aren't too concerned with the nitty-gritty details (see above) that really determine the effectiveness of schools. There's a good argument that they are incapable of knowing the details needed to effectively regulate, so its more important for the incentives to be aligned at the school level so educators will run their schools most effectively themselves. Currently, those incentives are far from aligned.