Friday, May 9, 2008

Digging Into "What Kids Are Reading"

Renaissance Learning, the company that runs the Accelerated Reader program, has just released a report, "What Kids Are Reading." They report the top twenty most read books, by grade level and gender; the top twenty titles by grade level and region of the country; and the top twenty most read titles by the top 10% most proficient readers.

One of the challenges of discerning what are good books for children relates to how unreliable much of the data is. Just because a book is selling really well doesn't mean either that it is being read, or if being read, is having much of an impact on the reader. Advertising and marketing budgets as well as external events like movie tie-ins can distort these numbers enormously.

The data Renaissance Learning is reporting has some advantages over other sources of information.
• The data population sizes are large which is always advantageous. Their smallest grade level cohort is 162,000.
• After a child has read a book they take a brief test that measures their proficiency and comprehension which can then be used to guide their next choice of book. Because of this, you know that the books have at least been read.
• Children usually have a lot of leeway in choosing which books they read, reflecting, presumably what they enjoy reading.
• Accelerated Reader covers a very large portion of available children's titles, versus other reading programs which are sharply limited in terms of the range of what can be read.
It is interesting to scan through the report and see what might be gleaned from the data. I have taken their reported information and drilled down a little further to see if there are any interesting tid-bits. As always, when using someone else's data, there are the caveats of not knowing the real quality of the information. In particular, I have no insight as to the validity of the key assumption that readability level reflects reading capability. I also am assuming that the cohort of top 10% of readers by proficiency is reasonably homogenous over the years, i.e. that while there may be some drift of new individuals in and others out as they progress over the years, there is some core majority that remain the same. This is not an unreasonable assumption based on real world experience but I have no means to validate that it is good for this data set.

With those caveats then, here are some of the observations:

Most titles are read at a single grade level but there are some that have a depth and richness that exerts an appeal to many readers across several grade levels.

Out of 132 titles read amongst all the grade levels, there are fourteen which are popular across four or more grades (i.e. they are among the twenty most read titles in four or more grades). These are:
• Harry Potter and the Sorceror's Stone by J.K. Rowling which crosses 6 grades including Fourth Grade to Ninth Grade and above
• Holes by Louis Sachar which crosses 6 grades including Fourth Grade to Ninth Grade and above
• Bridge to Terabithia by Katherine Paterson which crosses 5 grades including Fourth Grade to Eighth Grade
• The Bad Beginning by Lemony Snicket which crosses 5 grades including Fourth Grade to Eighth Grade
• Harry Potter and the Chamber of Secrets by J.K. Rowling which crosses 5 grades including Fifth Grade to Ninth Grade and above
• Harry Potter and the Goblet of Fire by J.K. Rowling which crosses 5 grades including Fifth Grade to Ninth Grade and above
• Harry Potter and the Prisoner of Azkaban by J.K. Rowling which crosses 5 grades including Fifth Grade to Ninth Grade and above
• Charlotte's Web by E.B. White which crosses 4 grades including Third Grade to Sixth Grade
• Because of Winn-Dixie by Kate DiCamillo which crosses 4 grades including Third Grade to Sixth Grade
• Hatchet by Gary Paulsen which crosses 4 grades including Fifth Grade to Eighth Grade
• The Reptile Room by Lemony Snicket which crosses 4 grades including Fifth Grade to Eighth Grade
• Eragon by Christopher Paolini which crosses 4 grades including Sixth Grade to Ninth Grade and above
• Harry Potter and the Order of the Phoenix by J.K. Rowling which crosses 4 grades including Sixth Grade to Ninth Grade and above
• Harry Potter and the Deathly Hallows by J.K. Rowling which crosses 4 grades including Sixth Grade to Ninth Grade and above

In the younger grades there is a tendency to read more books that have been around for awhile and at the older grades to be more focused on current titles

For first grade, the mean for the top twenty books was a publication of 24 years ago. There is a progressive slide towards more current books, till, by the older grades, the top preferred books were written a mean of 10-12 years ago, i.e. reasonably current and topical.



















Age of books by Grade
  Median
Publication Date
Written,
Years Ago
First Grade198324
Second Grade198522
Third Grade198918
Fourth Grade199314
Fifth Grade199314
Sixth Grade199314
Seventh Grade199710
Eight Grade199710
Ninth Grade >199512


Given that there are some very old outliers among the High School titles (Shakespeare and Nathaniel Hawthorne for example), the High School median is effectivley overstated.


The amount of reading done by the participants plunges from a high in the early grades of 46 books per year to a low in High School of 6 books per year.

There are any number of possible explanations. With our own kids, we have seen a pattern of them being excited to participate in such reading programs in the early grades but as they get older, more independent and more particular about making their own choices, they are no longer interested in participating at all or participating diligently (actually being tested for a book they might have read). They are not reading any less, they are just not bothering to record what they are reading. Sports, social commitments, the opposite sex, extracurricular activities (such as scouting, band, etc.) and of course increased structured school work are all other candidate explanations, particularly when you look at where the steepest declines occur (Fourth through Sixth Grade) which is when those variables begin to really kick in. All these activities take away larger and larger chunks of time that might formerly have been dedicated to reading. I also suspect that there are many "literature" books being read for classes that are just not getting captured in the later grades.

















  Number of Books
Read by Grade Level
First Grade38.6
Second Grade46.2
Third Grade40.2
Fourth Grade29.2
Fifth Grade19.1
Sixth Grade12.9
Seventh Grade10.7
Eight Grade7.1
Ninth Grade >6


The data is insufficient to identify a cause or causes for the decline but there is one other way of viewing the available information that might be meaningful.

The most accomplished readers (top 10% in reading proficiency) follow the same pattern of reading decline but with two critical differences.

















  Number of Books Read
by Grade Level Among
Top 10% of Readers
First Grade74.2
Second Grade84.8
Third Grade60.9
Fourth Grade41.5
Fifth Grade31.7
Sixth Grade26.8
Seventh Grade23.4
Eight Grade20.2
Ninth Grade >25.5


The first difference is that they are reading a much higher volume of books per year at the beginning and at the end. In First Grade, they are reading nearly twice as many books as the the whole group and this ratio remains roughly the same through Seventh Grade (roughly one and a half to twice as many books read in each grade). Then, beginning in eighth grade, they jump to reading nearly three times as many books as the average and by High School are reading more than four times as many as the average.

This up-tick in the relative number of books read is related to the second difference. Among the top 10% and the Average group, there is a sharp decline in the number of books read per year in every grade till Fifth Grade. From Fifth Grade onwards, the trends diverge. The average readers keep showing a material drop (15-30%) in number of books read from year to year. For the top 10% readers, there is still a decline but much more slowly, in the range of 10-15%. Then, when they hit High School, their number of books read per year bumps back up again. So they start by reading many more books, they decline more slowly and then they recoup more later.

Since the two groups are subject to the same external distractions (sports, opposite sex, etc.) there has to be some other explanation for the different patterns of decline. There is no way I can see to unlock the causality - are they reading more books because they are better able to read or are they better able to read because they have more access to books? It almost doesn't matter because I think the conclusions are the same - 1) get more good books in front of them as early as possible, and 2) encourage through all means, especially by word and by example, at the earliest age the message about the importance, wonder and pleasure of reading. If they have the habit and love of reading instilled early, the declines associated with distractions will self-right themselves later on.

The Readbility Level gap of books popular among the most proficient readers (top 10%) and all readers is narrow in early and later years but very large at Third and Fourth Grade.

Renaissance Learning uses their own ATOS system of measuring readability level and is not too dissimilar to Lexiles. As Renaisance Learning describes it, "ATOS uses four factors to measure a book's readability score: average sentence length, average word length in number of letters, word difficulty level, and total number of words in the book." For practical calibration purposes, The New York Times and The Washington Post average a readbility level of 7.8, USA Today is at 6.6 and People magazine is at 5.4.




















ATOS Readbility Level
 Top 10%AllDifference
First Grade21.70.3
Second Grade32.40.6
Third Grade5.441.4
Fourth Grade6.24.81.4
Fifth Grade6.55.60.9
Sixth Grade6.65.90.7
Seventh Grade6.66.20.4
Eight Grade6.46.20.2
Ninth Grade >6.56.10.4


In First Grade the gap is 0.3 units and then balloons to 1.4 by Third Grade. At Fifth Grade, the top readers are at a plateau - from Fifth Grade onwards they only fluctuate up or down by 0.1 units around a readability score of 6.5. The average readers though keep on building their readbility level through to High School and by High School the gap between the top readers and everyone else has narrowed back down to only 0.4 units (6.5 versus 6.1). Tragically though, at this point, even though they are able to read at about the same level, one group is reading a lot and the other group hardly anything.

From these numbers then you can see that the best readers are reading a larger number of books from the outset, they are reading more challenging books earlier and end school reading a reasonable number of books (in the scheme of the numbers being reported) per year. In contrast, the average population are reading fewer books at the outset, suffer a steeper decline in the number of books they read with each advancing grade and end school reading few books at all. However, in their school years they have narrowed the skills gap of reading so that they are reading close to the same level of the best readers as measured by readability level.

The above four observations would point towards home/cultural values being a determining factor in reading patterns. The schools appear to be doing a good job of closing the readability gap in the years they have the students under their tutelage. The best readers appear to already be launched on a path that values reading by the time they arrive in First Grade as reflected in the number of books they are reading compared to their cohort. The habit, cultural appreciation and personal enjoyment of reading would explain why their decline in number of books read at each grade level is slower than for their cohort; it would also explain the rebound in numbers in High School. Correspondingly, home/cultural values would also seem to explain why children that start out with with marginally lower reading capabilities arrive reading fewer books and then follow a continuing path of fewer and fewer books being read each year even though their accomplished readability level keeps rising. They can increasingly read a broader range of books but they are choosing, for unknown reasons, not to.

There are material differences in the choice of titles girls and boys choose to read but this difference also follows a pattern.


















Boy Girl Reading Preferences Overlap
  Percent Overlap M/F
First Grade85%
Second Grade70%
Third Grade65%
Fourth Grade55%
Fifth Grade60%
Sixth Grade75%
Seventh Grade70%
Eight Grade50%
Ninth Grade >70%


Boys and girls read a similar range of books in their early school years. In First Grade 85% of the top twenty titles are the same for girls and boys. This correlation then falls steadily to a low point in Fourth Grade where just above half of the most popular books are in common. The commonality of reading then plateaus from that point forwards with roughly 70% of the books being common between the genders with one anomaly in Eight Grade when only half the books are common to both.

Looking at the actual titles reveals even more interesting items. Most of the books that are read by one gender and not the other are part of series. In Fourth Grade, fully (and repugnantly) 40% of what the boys are reading are books from the Captain Underpants series. In no grade does Captain Underpants show up among the top twenty most popular titles among girls. Correspondingly, though less preponderantly, girls are in Fourth Grade reading the Junie B. series and later the Clique series, no titles of which ever show among the boys' top twenty. There are some hybrid series as well. Boys appear to take to the Harry Potter books sooner (starting in Fourth Grade) and with greater avidity (25-30% of the top twenty titles over several grades) than girls. Girls do read the Harry Potter books but in later grades and with less intensity (never more than 10% of their top twenty titles). Similarly, girls take to the Lemony Snicket books earlier than boys, Fourth Grade instead of Fifth Grade, and also with greater intensity (15-25% of the titles each year till Eight Grade).

At Eighth Grade, the second divergence in reading patterns appears. The books that boys are reading which girls are not are Adventure (for example Call of the Wild and Hatchet) and Fantasy (still a lot of Harry Potter plus Eragon and Eldest). The books girls are reading that don't show up in the boys column include titles from the Clique series, Westerfeld books such as Uglies and Meyer books such as New Moon and Brashares' Sisterhood of the Travelling Pants.

There is a material difference in the reading difficulty by gender at the later grade levels.

This finding caught me by surprise as it is counter to what I have seen and read most consistently over the years which was that girls tend to perform better in reading and verbal tests than boys. Now reading difficulty levels are not necessarily the same thing as effective reading as measured by standaridized tests. As mentioned above, Renaissance Learning uses their own ATOS system of measuring readability level. Reiterating from above, as Renaisance Learning describes it, "ATOS uses four factors to measure a book's readability score: average sentence length, average word length in number of letters, word difficulty level, and total number of words in the book." For practical calibration purposes, The New York Times and The Washington Post average a readbility level of 7.8, USA Today is at 6.6 and People magazine is at 5.4.

The average reading scores of the top twenty books most preferred by boys and girls are basically the same in grades one through three. A gap then opens up at Fourth and Fifth Grade which settles at about a 0.7 unit spread between the difficulty of the books the boys are choosing to read and those that are being read by girls from Sixth Grade onwards. Now it may be that the degree of reading difficulty does not correlate well with standardized tests which focus more on reading comprehension, etc. It is possible to hypothesize that girls are reading somewhat easier books and understanding more (i.e. they have lower ATOS readability scores on the books they are reading but higher standardized test results because they are more proficient in their comprehension.)




















Readbility Level by Gender
 MaleFemaleDifference
First Grade1.71.70.0
Second Grade2.32.4-0.1
Third Grade4.14.00.1
Fourth Grade4.94.50.4
Fifth Grade5.85.30.5
Sixth Grade6.15.40.7
Seventh Grade6.35.60.7
Eight Grade6.45.80.6
Ninth Grade >6.45.70.7


It is possible to make that hypothesis, but it doesn't really ring right. You would expect that children are choosing to read books that are within their comprehension range and that that would reflect itself on standardized reading tests. Perhaps it is a function of some quirk of the Accelerated Reader measurement system for measuring reading achievement but I would sort of doubt that too - the program has been used long enough for something like that to be caught and rectified. So - don't know, but surprising.

The books that are read by the most proficient readers (10%) are basically the same as those read by all their peers, the most proficient readers are just reading them sooner.

The overlap between the books read by the best readers and everyone else is shown below. At first glance it would appear that there is a material difference.


















Reading Commonality by Ability
 Overlap
First Grade55%
Second Grade35%
Third Grade40%
Fourth Grade45%
Fifth Grade55%
Sixth Grade65%
Seventh Grade65%
Eight Grade80%
Ninth Grade > 70%


However, when you look at the specific titles, you see there is a very strong pull-forward effect. For example in First Grade, there are nine books that the 10% best readers are reading that don't show up among the list of twenty for the average readers. However, seven of those nine books are among the top twenty for the average readers in the next grade - the top readers are pulling their reading material in from the same popular books, just sooner. This pull-forward effect accounts for most of the difference between the lists for the 10% and their average compatriots at each grade level. Incidentally, most of the books being pulled forward are from series. Clearly young readers are getting hooked on something they like and are raiding the later inventory of books.

47% of all books across all grades are parts of a series.

Series serve as a wonderful way of keeping a reading mind stoked; if they liked one in the series they are likely to enjoy the others. The series which show up among the top twenty titles across the grades include:

"Biscuit" series

"Cat in the Hat" series

"David" series

"Little Critter" series

"If You . . "series

"Henry and Mudge" series

"Black Lagoon" series

"Jimmy' Boa" series

"Officer Buckle" series

"Click, Clack" series

"Magic Tree House" series

"Miss Nelson" series

"Arthur" series

"Alexander" series

"Captain Underpants" series

"Junie B." series

"Alexander" series

"Amelia Bedelia" series

"Harry Potter" series

"Lemony Snicket" series

"Shiloh" series

"Hatchet" series

"Chronicles of Narnia" series

"Inheritance Trilogy" series

"Clique" series

"Twilight Saga" series

"Sisterhood" series


I was amazed at the absence of the Hardy Boys.

No comments:

Post a Comment