Friday, November 7, 2014

Pleasure reading is (very) good for you at any age: The Sullivan-Brown report (2014)


Stephen Krashen

Sullivan and Brown (2014) examined determinants of performance on a vocabulary test given to 9,400 42-year olds in the UK. A great deal of data on these subjects gathered throughout their lifetimes was available, and the researchers were able to examine the impact of a wide variety of predictors.

Among the important results were these:

The amount of pleasure reading done (books):

Those who said they did more book reading for pleasure at age 42 had nigher levels of vocabulary development, even when social factors such as parental education and parental occupation were controlled, as well as the subjects' level of education and occupation.

Sullivan and Brown also reported that those who reported doing more reading as children and at age 16 also had higher vocabulary scores, but the amount of reading done at age 42 predicted vocabulary independent of earlier reading.

In other words, continuing to read as an adult counts. Language and literacy development is possible at any age.

 What kind of reading was done:

Those who read "high-brow" fiction had larger vocabularies than those who read middle-brow fiction, who in turn had larger vocabularies than those who read low-brow fiction. Low-brow readers did not have significantly higher vocabularies than those who read nothing. Reading fiction, at least middle-brow fiction, counts.

The same was true for "factual" books, except that low-brow factual book readers did slightly better than those who read nothing.

High-brow fiction readers did better than high-brow "factual" book readers, once again demonstrating the value of fiction.

It was not entirely clear what is high, middle or low-brow. Sullivan and Brown state that "crime, thrillers and mystery" are middle-brow fiction, "contemporary literary fiction" is high-brow, and "science and politics, economics and current affairs" are high-brown nonfiction. Other genres are not so defined.

Note (see table) that those who are regular readers, every day (3.5% increase) or several times a week (2.3%) of "middle-brow" fiction do as well as those who read "high-brow" books less frequently (eg. once a month = 1% advantage).

[e.g. middle brow (3.4%) + several times per week (2.3%) = 5.7%; high-brow (5.3%) + once a month (1%) = 6.3%]

A gap in the study is that only two kinds of reading were considered: Books and newspapers. There was no analysis of magazine reading, or reading articles and blogs from the computer.

Additional results

Newspaper reading: Those who read regular newspapers performed somewhat better on the vocabulary test. Those who read both regular newspapers and tabloids did not do better than those who read nothing, and those who read only tabloids did worse than those who read no newspapers of any kind.

Music: The results here are odd. Those who said they played a musical instrument at age ten did slightly better on the vocabulary test. Those who said they played an instrument at age 16 were no different than those who didn't play an instrument, and those who played an instrument at 42 did significantly better on the vocabulary test. Apparently, if you continue to play you might experience a decline but then later it pays off. The advantage at age 42, however, was modest, a advantage over those who didn't play an instrument at 42. In contrast, those who read books every day were 3.5% better than nonreaders, and those who read several times a week were 2.3% better than nonreaders.

This table presents Sullivan and Brown's results. The statistical technique used, multiple regression, enables the researcher to measure the impact of each predictor, holding the others constant, that is assuming that they do not influence each other.

As was the case in an earlier study of 16 year olds (Sullivan and Brown, 2013, summarized in http://skrashen.blogspot.com/2013/09/new-evidence-for-power-of-reading.html), social class of the parents was a strong predictor of vocabulary size when reading behavior was not considered. Once reading predictors were added to the analysis, parental social class variables were no longer significant predictors. This suggests that reading can help overcome at least some of the effects of poverty.

The vocabulary test had 20 items. The mean score was 63% correct (12.6/20). A 5% advantage means one item more was correct.

predictor
compared to


parent occupation
a routine job
% better
Significance
mangerial

0.1
Ns
intermediale

0.1
Ns
long-term unemployed

-1.5
Sig
parental ed
no exam taken


secondary schools exams not passed

1.2
very sig
secondary schools exams passed

0.7
Ns
degree

0.5
Ns
newspaers in home at 16
none


both

0.7
Ns
Regular newspapers only

0.4
ns
tabloids only

-0.7
Ns
Book reading at 16
rarely or never


more than once a week

1.8
very sig
once a week

0.9
very sig
less than once q week

0.3
very sig
child reading
never/hardly ever


often

3.5
very sig
sometimes

2 pt 5
very sig
plays musical instrucment
does ntt play


age 10

0.6
Sig
age 16

-0.6
Ns
age 18 exam score

3.5
very sig
prior vocabulary test score



age 5

0.1
very sig
age 10

0.2
very sig
age 16

0.3
very sig
education age 42
no exam taken


secondary school exams not passed

1.7
very sig
secondary  school exams passed

1.8
very sig
degree

1.9
very sig
elite degree

2.2
very sig
occupation age 42
routine


Mangerial

2.4
very sig
Intermediale

1.7
very sig
long-term unemployed

-1
Sig
Frequency of reading at 42
no reading


reads books every day

3.5
very sig
several times a week

2.3
very sig
at least once a month

1
0.14
every few months

1.1
0.12
at least once a year

0.4
ns
fiction at 42
no reading


low-brow

0.8
ns
middle-brow

3.4
very sig
high-brow

5.3
very sig
factual books at 42
no reading


low-brow

1.1
sign
middle-brow

2.3
very sig
high-brow

3
very sig
read newspapers at 42
no reading


Both

0.2
ns
regular newspapers only

1.2
sig
tabloids only

-1.3
very sig
musiclal instrument at 42
does not play
1.2
0.02
overall r2 = .56



very significant: ..01 or less



significant = .05 or less



not significant = larger than .05





Since the predictors are independent of each other, we can predict the best readers by adding them. Taking the maximum scores from each category:
Book reading at 16: 1.8
Read as child: 3.5
Elite degree: 2.2
Managerial occupation: 2.4
Reads books every day 3.5
High-brow fiction: 5
High-brow factual books: 3
Regular Newspaper: 1.2
Child reading: 3.5
Regular newspaper: 1.2
Plays musical instrument: 1.2
Total = 39.3
.393% of 20 = 7.86 point advantage

Those with all the advantages will thus score an average of nearly eight points higher (out of 20) than those with the least advantages.

Sources:

Sullivan, A. and Brown, M. 2013. Social inequalities in cognitive scores at age 16: The role of reading. London: Centre for Longitudinal Studies,
Institute of Education, University of London   www.cls.ioe.ac.uk
Sullivan, A. and Brown, M. 2014. Vocabulary from adolescence to middle-age. Centre for Longitudinal Studies
Institute of Education, University of London






1 comment: