Reading Last…

MDRC does a real service with this policy brief on the late Reading First program (pdf).

2 Replies to “Reading Last…”

  1. Two things stand out about the MDRC report on Reading First: We still don’t know how well, if at all, Reading First really worked. And, even schools that implemented the “scientifically-based” reading practices may not have shown many gains.

    To me, this points to a serious potential for error in our understanding of the science of reading.

    I have watched young children learn (and not learn) to read in hundreds of classrooms all across the country. When teachers spend a lot of time (as suggested by Reading First) on decontextualized phonics and phonemic awareness activities, students tend to make slow and painful progress, if they progress at all.

    However, when teachers concentrate on differentiation by teaching kids how to choose their own “just right” books, individual conferencing, and modeling of explicit comprehension strategies, kids really seem to take off. Furthermore, though our government tells us that fluency is vitally important (and it is) no one seems to have any lessons on how to teach it. (Other than the DIBELS “read it faster” approach, I’ve never seen a teacher give any explicit instruction in fluency to a class of children.) To correct this, our company developed such a set of instructional methods for fluency and expression and reading rate — all of which seem to work well for students at all grade and ability levels.

    My problem with Reading First has always been two-fold: It never seemed to be worth the money we were spending and it didn’t identify the most effective approaches to teaching reading. I wish the MDRC brief, and other writings on Reading First, would address these issues directly — especially the issue of scientific research on reading. I believe we have made a serious mistake over the last decade in identifying areas of reading practice that are not the most effective. If Reading First has proven nothing else, I believe it has proven this.

  2. Yet another perspective…

    Perhaps the reason Reading First was not all that successful was because the government was not specific enough as to which actual teaching practices might yield the best results. The government advocated instruction in five areas but did not specify what that instruction should look like specifically.

    In our company, we have developed specific practices for early readers that have made a huge difference in how quickly they learn to read. These practices include:

    1. Teaching letters by sound rather than by name to improve the rate at which kids learn the Alphabetic Principle.

    2. Teaching letters in small groups and making words out of them so that letters are not learned in isolation.

    3. Teaching common spelling patterns.

    4. Using writing, from day one, as the best possible phonemic awareness activity (along with a spelling correction procedure that keeps children from learning incorrect or invented spellings).

    5. An explicit strategy for sounding out words (we use a little poem).


    This repertoire of specific practices makes a dramatic difference to teachers who use them. Along with a larger set of comprehension strategies, (again, modeled and explicitly taught), we see great gains in young readers.

    The problem with most government policy (federal or state) is that it leaves implementation to people who, by definition, have not been successful implementers. This leads to bad choices about specific practices — or to no practices at all.

    Too many teachers think they can have success by merely following a textbook or program. The government supports this notion by providing funding for the purchase of these materials. But programmed instruction simply has not been shown to be successful. And yet, there are simple practices readily available to teachers (free, in the case of practices that come from our company) that would make a significantly greater difference in student achievement.

Leave a Reply

Your email address will not be published.