ArticlePDF Available

Abstract and Figures

This study set out to develop a Chinese Author Recognition Test (CART) that might be used as a measure of objective print exposure for college students in Taiwan. We found that there is a linkage between print exposure and general reading achievement for college students. We also found that, among self-reported reading habits, comparative reading habits and CART, primary print knowledge scores within the CART family have the strongest prediction power for both the ‘General Scholastic Ability Test-Chinese’ and the ‘Department Required Test-Chinese’ beyond the joint contributions of vocabulary size and reading comprehension. By sharing the process of developing the instrument, we shed some light for researchers from regions other than English-speaking countries on how they might move forward in future investigations.
Content may be subject to copyright.
Developing a Chinese version of an
Author Recognition Test for college
students in Taiwan
Su-Yen Chen
Institute of Learning Sciences, National Tsing-Hua University, Taiwan
Sheng-Ping Fang
Department of Chinese Literature, National Tsing-Hua University, Taiwan
This study set out to develop a Chinese Author Recognition Test (CART) that might be
used as a measure of objective print exposure for college students in Taiwan. We found
that there is a linkage between print exposure and general reading achievement for
college students. We also found that, among self-reported reading habits, comparative
reading habits and CART, primary print knowledge scores within the CART family have
the strongest prediction power for both the General Scholastic Ability Test-Chinese
and the Department Required Test-Chinesebeyond the joint contributions of vocabu-
lary size and reading comprehension. By sharing the process of developing the instru-
ment, we shed some light for researchers from regions other than English-speaking
countries on how they might move forward in future investigations.
There is tremendous variation in literacy habits. Researchers in Western cultures have
linked such variation in college studentsprint exposure to their variability in
reading-related skills, such as lexical-decision latency (Chateau & Jared, 2000), spelling
ability (Burt & Fury, 2000; Stanovich & West, 1989), vocabulary (Martin-Chang & Gould,
2008; Stanovich, West & Harrison, 1995; West & Stanovich, 1991), verbal uency
(Stanovich & Cunningham, 1992), reading comprehension and reading rate (Martin-Chang
& Gould, 2008), and general measures of achievement in reading and reading-related
domains (e.g., the SAT reading and English subtests; Acheson, Wells & Macdonald,
2008). Because level of print exposure has been one of the key variables in individual
differences across many reading-related cognitive dimensions, developing a reliable
Chinese instrument for assessing relative levels of print exposure is the focus of this study.
Traditional research has used a self-reported measure of reading habits, which asks
college students or young adults to estimate the amount of time they spend reading.
Results of questionnaires and interviews have served to assess their variation in literacy
habits (e.g., Gallik, 1999; Guthrie, 1981; Walberg & Tsai, 1984). An alternative tool, the
Author Recognition Test (ART) for college students, was rst developed by Stanovich
and West (1989) and has been widely used since its creation. The ART has consistently
proven to be an ideal proxy measure for out-of-school print exposure. The test includes
the names of both popular authors and foils, and respondents are to indicate whether or
Copyright © 2013 UKLA. Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ,
UK and 350 Main Street, Malden, MA 02148, USA
Journal of Research in Reading, ISSN 0141-0423 DOI:10.1111/1467-9817.12018
Volume 00, Issue 00, 2013, pp 117
not they are familiar with the name of a particular author by putting a check mark next to
the name. According to Stanovich, Cunningham and West (1998), the advantages of the
ART over traditional methods include easy administration, low task demands, time-saving
and minimisation of the complications associated with social desirability involved in
self-reporting.
The rst version of the ART for college students was developed as a measure reecting
relative individual differences in print exposure and was not intended to measure absolute
levels of print exposure (Stanovich & West, 1989). It consisted of 50 popularauthors and
50 foils. Mostof the popular authors regularly appeared on ction and non-ction best-seller
lists. Major categories of ction included mystery/detective, romance/Gothic, spy/intrigue,
occult/supernatural, historical novels, Westerns, short stories and science ction. Non-ction
categories included sports, science, politics/current events, humour, religious, history,
biography and business/nance. Authors that appeared in the school curriculum were
avoided, as the ART was intentionally biased toward out-of-school reading experience. Foils,
that is, names taken from the editorial board of a professional journal that were unfamiliar to
all respondents in the study, were used to prevent respondents from simply checking all the
names. Scoring on the task was determined by subtracting the false-alarm rate from the hit
rate. The original version of the ART was later revised into a 25-item instrument containing
the names of 16 authors and 9 foils in West and Stanovichs (1991) study and into an 80-item
version containing 40 authors and 40 foils in several succeeding studies (e.g., Stanovich &
Cunningham, 1992; Stanovich et al., 1995). Although the Magazine Recognition Test
(MRT) and the Newspaper Recognition Test (NRT) have been frequently utilised, the ART
has remained the most sensitive measure of print exposure, probably because the writing
styles of individual authors inuence readersselection of reading materials and avid readers
usually have their favourite authors. Magazines and newspapers, on the other hand, represent
collections of works of many authors and therefore are of less distinctive value.
The ART was also revised or developed in several English-speaking countries other than
the United States. Chateau and Jared (2000), for example, generated a Canadian version of
the ART for college students and found that levels of print exposure were highly correlated
with the efciency of both orthographical and phonological word recognition. Burt and
Fury (2000) also revised the 1989 (Stanovich & West, 1989) and 1992 (Stanovich &
Cunningham, 1992) versions into a test for Australian college students. Authors found that
level of print exposure served as a better predictor of spelling ability than the joint
contributions of reading comprehension and vocabulary. Later, Masterson and Hayes
(2007) developed new versions of the ART and the Title Recognition Test (TRT) for adults
in the United Kingdom. Note that, in the past, the TRT was used mainly for children rather
than adults (e.g., Allen, Cipielewski & Stanovich, 1992; Cipielewski & Stanovich, 1992;
Cunningham & Stanovich, 1990; Ecalle & Magnan, 2008; McBride-Chang & Chang,
1995). Masterson and Hayesinitial list of authors and titles was equally distributed among
four sources: the top 100 most-loaned titles according to a library for 2002 and 2003, the
Nielscan Top 100 ction titles purchased before September 2004, the top-ranked list at an
online bookshop, and the titles or authors that had recently been awarded prizes. These lists
were then pilot-tested, and the nal lists were formed by deleting the items that were least
often correctly identied. The researchers found higher scores for the UKART than for the
UKTRT. Both measures served as good predictors for spelling ability, vocabulary and
reading rate, but only the UKART was related to comprehension.
Two recent studies might have elaborated our understanding of the ART. First, Martin-
Chang and Gould (2008) proposed the notion of dividing the ART into two dimensions:
2 CHEN and FANG
Copyright © 2013 UKLA
personal reading experience from primary print knowledge (PPK) and from secondary
print knowledge (SPK). They updated the original version of the ART (Stanovich &
West, 1989) by adding 30 contemporary best-selling authors suggested by avid
readers during pilot investigations, thus resulting in a nal checklist consisting of
75 widely recognised authors and 75 foils. In this version of the ART, in addition
to indicating whether or not they were familiar with the name of a particular author,
respondents were also asked to report if (1) they had heard of this author but not read
any of his or her books, (2) they had started a book by this author but not nished it,
or (3) they had read one or more of this authors books. Although scores for the PPK
were obtained by summing the number of authors that respondents successfully
recognised and reported that they had read one or more of this authors books
divided by the total number of real authors, scores for the SPK were calculated by
summing the number of authors that respondents successfully recognised but reported
that they had heard of this author, but not read any of his or her booksover the total
number of real authors. Foils that were incorrectly marked were ignored from both
measures of print exposure. Authors that respondents successfully recognised and
reported that they had started a book by this author but not nished itwere also
ignored. This study found that the PPK accounted for variance in all three criterion
variables (vocabulary, comprehension and reading rate) after the effects of the SPK
had been factored out.
Second, Acheson et al. (2008) measured print exposure not only with the revised
ART and MRT but also with several new self-reports on reading and writing habits:
time spent reading, time spent writing and comparative reading habits (CRH). In the
section of time spent reading, participants were asked to estimate the amount of time
they spent in a typical week reading certain types of material, for example, textbooks,
academic material, magazines, newspapers, emails, Internet material, ction and
non-ction. Similarly, they were asked to estimate the amount of time they spent
writing certain types of material, for example, for class, articles, personal, emails,
creative and job-related. In the CRH section, they were asked to compare their own
reading habits to those of other college students on a Likert scale consisting of ve
dimensions: time spent reading, complexity of reading material, reading enjoyment,
reading speed and reading comprehension ability. This study employed hierarchical
regression and found that when the CRH was entered before the ART/MRT composite,
the unique predictability of the ART/MRT was reduced to nonsignicance. The
researchers suggested that the CRH might be better equipped to capture a broader
range of reading experiences, including Internet-based reading. In addition, although
self-estimated reading time fails to differentiate slow readers from avid readers and
is therefore relatively unreliable, the CRH appears to be a useful supplement to other
assessments of print exposure.
In summary, previous studies from English-speaking countries have used
various measures of print exposure for college students: self-reported reading
habits, self-reported CRH and objective print exposure measures such as the ART.
The relationships between print exposure, and reading-related skills or general reading
performance have been well-established, especially between the ART and the reading-
related performance. Compared with studies from Western cultures, investigations of
college studentsprint exposure and its relationship with reading-related performance
in Taiwan have been limited. This paper represents one of the rst efforts dedicated to
this topic.
CHINESE AUTHOR RECOGNITION TEST 3
Copyright © 2013 UKLA
Construction of a Chinese version of the Author Recognition Test in Taiwan
Extant studies in Taiwan have almost exclusively employed self-reported reading habits by
using questionnaires to collect data concerning levels of print exposure. Chen (2007, 2012)
utilised self-reported reading habits as an indicator of print exposure and found that
Taiwanese college studentsfavourite reading materials were e-news, Bulletin Board
Systems, ction, magazines, newspapers and non-ction, among others, which is consis-
tent with recent ndings that reading Internet-based materials plays a signicant role in
contemporary college studentslives (e.g., Jolliffe & Harl, 2008; Liu & Huang, 2008).
Chen (2009) also found that a higher educational level is more strongly linked with reading
literature and professional non-ction than with reading popular ction and practical
non-ction, suggesting a relationship between education and reading taste. These results
generally corroborate studies related to the cultural consumption of reading carried out
by Dutch researchers: they found that educational level was a strong predictor of reading
highbrow, complex and prestigious books that have the unique features of being more
difcult and demanding, requiring more prior knowledge and being more important
and of higher literary value (e.g., Kraaykampt & Dijkstra, 1999; Van Eijck, 1999; Van
Rees, Vermunt & Verboord, 1999).
Few studies in Taiwan have explored the association between level of print exposure and
reading-related performance. Therefore, on the basis of the aforementioned literature, the
goal of this research was threefold: (1) to develop a Chinese Author Recognition Test
(CART) for college students in Taiwan; (2) to examine the overall correlations between
the CART and self-reported reading habits, self-reported CRH and reading-related
measures, in order to provide preliminary evidence for the CARTs construct validities;
and (3) to investigate the extent to which self-reported reading habits, self-reported
CRH and the CART help predict two criterion variables of general reading performance
(e.g., the General Scholastic Ability Test-Chinese [GSAT-Chinese] and the Department
Required Test-Chinese [DRT-Chinese]), beyond the joint contribution of vocabulary
size and reading comprehension, in an attempt to further validate the CART as a
measure for assessing individual differences in print exposure. By sharing the process of
developing the instrument, we shed some light for researchers from regions other than
English-speaking countries on how to proceed in future investigations.
Although a Chinese version of the TRT for children was used in McBride-Chang and
Changs (1995) study, neither the procedure for development nor the content of the
instrument was mentioned in the study. And no similar study has been conducted, probably
because the researchers found the results to be inconsistent with the ndings of Western
studies; that is, they found that print exposure did not signicantly help predict additional
variance in the reading comprehension measure once vocabulary was forced into a
hierarchical regression.
Before constructing the rst measure of objective print exposure for college students in
Taiwan, we had a number of decisions to make. First, the ART rather than the TRT, the
MRT or the NRT was chosen, because almost all extant studies have found the ART to
be the most sensitive instrument for college students. On the other hand, for children,
the TRT was found to be a more sensitive instrument than the ART by most studies,
because children may read many books but not take any notice of the author information.
In addition, because of the limited population and therefore the limited market size, there
are simply not enough magazine or newspaper titles available in Taiwan to generate a valid
NRT or MRT test.
4 CHEN and FANG
Copyright © 2013 UKLA
Second, unlike in English-speaking countries, where books originally written in English
might fairly well represent most readersrelative levels of print exposure, in Taiwan, most
readersreading experiences might consist not only of works written in Chinese but also of
those translated from many other languages. For example, according to eight available top-
ranked library loan titles and bookstore best-selling lists in 2010, 60% of titles were translated
works. Although developing an ART that consists of both Chinese authors and translated
names of foreign authors would seem to be an option for this study, it is rather unfeasible
to do so. The difculty is mainly due to the lack of a uniform system of translating names
of foreign authors. A particular foreign author may have several versions of translated name
in Taiwan, depending on the publisherspreferences. The average Taiwanese reader therefore
does not bother paying attention to the authors name at all when reading a translated work.
As a result, it was unfortunate that we had to give up on developing an instrument to measure
objective print exposure concerning reading experiences related to translated works.
Third, the original ART list by Stanovich and West (1989) intentionally included only
popular writers and avoided highbrow writers that were known only to the most academ-
ically inclined readers, but we decided to develop two CART sublists: CART-popular for
popular authors generated from top-ranked book lists and CART-highbrow for highbrow
authors suggested by avid readers. There were two reasons for this decision. The rst
was that we found that the initial list generated from the eight top-ranked library loans
and bookstores from 2010 in Taiwan was not very comprehensive as it did not include
many important Chinese writers. The second reason was that some previous studies have
proposed an association between higher educational level and reading both complex and
prestigious books; thus, we considered that it would be meaningful to explore the
difference between CART-highbrow and CART-popular.
The initial authors on the CART-popular sublist came from eight top-ranked lists in
Taiwan, including ve different types of university library loan schemes and the three most
popular bookstores. They were the following:
1. National Tsing Hua University Lending Library Top 100 books in 2010;
2. National Yang-Ming University Lending Library Top 100 books in 2010;
3. National Chiayi University Lending Library Top 100 books in 2010;
4. Feng Chia University Lending Library Top 100 books in 2010;
5. National Normal University Lending Library Top 100 Popular Science books in 2009;
6. KingStone Book Store 100 Best-selling books in 2010;
7. Eslite Book Store 100 Best-selling books in 2010 (winter only); and
8. Online Berkeley Book Store (http://www.books.com.tw/) 100 Bestselling books
in 2010.
These eight lists include a total of 440 writers, many of whom have multiple works on the
lists, or one particular work on multiple lists. In all, 177 writers out of the 440 (about 40%)
were categorised as Chinese writers; that is, their works on these lists were written in Chinese
regardless of whether they lived in Taiwan, Mainland China or elsewhere.
The initial authors on the CART-highbrow sublist came from three avid readers, one of
whom is a doctoral student majoring in Chinese literature; the other two are Masters
students who have won literary awards and have had works published. A highbrow writer
refers to an author whose work is relatively important, of the highest literary value, more
difcult to read and requires prior knowledge. These three avid readers collected highbrow
writersnames from their circles of friends and came up with an initial list of 92 authors.
CHINESE AUTHOR RECOGNITION TEST 5
Copyright © 2013 UKLA
Authors that appeared on both the popular and the highbrow lists were counted as
writers on the CART-highbrow and were deleted from the CART-popular. Following
Stanovich and Wests (1989) rule, authors that had appeared in any of the high-school
curricula were also deleted from the lists. Thus, the CART-composite list for the pilot test
consisted of 233 realauthors. We also added six foils, names of famous characters from
ction, on the list, to identify and weed out any random responses, yielding a total of 239
authors on the instrument for pilot testing.
Method
Development of the formal Chinese Author Recognition Test
We collected data for the pilot test in September 2011 from 525 college students taking
General Education courses at ve universities, with two in Northern Taiwan, one in Central
Taiwan, one in Southern Taiwan and one in Eastern Taiwan. The universities represent
different types of higher institutions in Taiwan in terms of public/private, institution size,
institution rank and combination of schools. On the instrument, we asked respondents to
indicate whether or not they were familiar with the name of a particular author by putting
a check mark next to the name. To prevent the subject from making the mark carelessly
and therefore contaminating our data, once any of the six foilswas checked, that partic-
ular respondents questionnaire was excluded from the data analysis. At the end, data from
248 respondents were categorised as valid. Among them, 104 (41.9%) were women and
144 (58%) were men; 26 (10.5%) were freshmen, 86 (34.7%) were sophomores, 79
(31.9%) were juniors and 57 (23%) were senior students.
According to the 248 respondents, the selection rate for each of the 233 realauthors
ranged from 0 to 245, with a mean of 36.11 (SD = 53.181), indicating that many authors
were selected by less than 10% of respondents. To be specic, at the two extremes, 156
authors were familiar to less than 10% of respondents, whereas only two authors were
familiar to 90% of respondents. Therefore, we took the remaining 75 authors with
recognition rates falling between 10% and 90% to be the nal list of real authorsin
our CART instrument. Among these 75 authors, 36 belong to the CART-popular list and
39 belong to the CART-highbrow list. They roughly represent the following categories:
literary ction and prose (28%), poetry and biography (8%), popular and light reading
(20%), Internet ction later published as print-based ction (8%), manga/graphic novels
(5%), professional non-ction (14%) and practical non-ction (15%). The distribution is
largely the same as that of the initial list. We added an equal number of 75 foils, names
taken from a list of school teachers on the Internet; we double-checked each name to be
sure that it was not the same as a writer of any kind, this time for the nal version. In this
nal version, in addition to indicating whether or not they were familiar with the name of a
particular author, respondents were also asked to report if (1) they had heard of this author
but not read any of his or her books, (2) they had started a book by this author but not
nished it, or (3) they had read one or more of this authors books. Therefore, three kinds
of scores were generated for the purpose of this study. First is the original CART score,
ranging from 75 to 75, calculated by taking the number of the correct items that were
checked and by subtracting the number of foils checked. The second one is the PPK-
composite score, ranging from 0 to 75, obtained by summing the number of authors that
respondents successfully identied and reported that (3) they had read one or more of this
6 CHEN and FANG
Copyright © 2013 UKLA
authors books. The PPK-composite score equals the PPK-popular score, ranging from 0 to
36, added to the PPK-highbrow score, ranging from 0 to 39. And nally, the SPK-composite
score, ranging from 0 to 75, obtained by summing the number of authors that respondents
successfully identied and reported that (1) they had heard of this author but not read any
of his or her books. The SPK-composite score also equals the SPK-popular score, ranging
from 0 to 36, added to the SPK-highbrow score, ranging from 0 to 39.
Conducting the formal study
The formal study was conducted in October and November 2011. The sample consisted of
college students enrolled in a highly selective science/engineering research-oriented
university in Northern Taiwan. A total of 358 students were recruited. Among them, 212
(59.2%) were women and 146 (40.8%) were men; 92 (25.7%) were freshmen, 99
(27.7%) were sophomores, 96 (26.8%) were juniors and 71 (19.8%) were in their fourth
or fth (extended) year. Because the nal goal of this study was to validate the CART
by using hierarchical regression logic to partial out variance in vocabulary size and
reading comprehension before comparing linkages between various measures of print
exposure (e.g., CART, self-reported reading habits and CRH) and general reading
performance, assessed by the GSAT-Chinese and the DRT-Chinese, all respondents were
asked to take (1) the CART, (2) the Vocabulary Size Test and (3) the Reading Comprehen-
sion Test, and (4) to ll out a reading questionnaire regarding their self-reported reading
habits, CRH and their scores on the GSAT-Chinese and DRT-Chinese (if available) when
entering college.
Print exposure and experience measures
Self-reported reading habits. Respondents were asked to ll out the questionnaire according
to their own experiences of reading. Although earlier studies usually measured self-reported
reading habits by asking respondents to provide an overall time estimate, in this study, we
asked respondents to provide reading frequency of various kinds of materials on a four-point
scale ranging from 0 (never) to 3 (frequently), in addition, for comparison purposes, and then
computed an average frequency score, as suggested by Acheson et al. (2008). Furthermore,
we differentiated self-reported reading habits into two major categories: print-based and
Internet-based, as recent studies have suggested a new trend of reading Internet materials
among college students (e.g., Chen & Fang, 2012; Jolliffe & Harl, 2008; Liu & Huang,
2008). In other words, self-reported reading habits were measured not only by the overall
amount of time respondents estimated spending on extra-curricular reading in print-based
and Internet-based formats, respectively, but also on the frequency with which they engaged
in reading eight types of materials, some of them print-based (e.g., printed books ction,
printed books non-ction, newspapers and magazines) and the others Internet-based
(e.g., e-ction, e-news, blogs and Bulletin Board Systems), resulting in two average
scores of reading frequency.
Comparative reading habits. In the CRH section, respondents were asked to compare their
own reading habits with those of their peers for six dimensions: time spent reading, reading
enjoyment, variety of reading, reading speed, complexity of reading material and compre-
hension of reading material (Acheson et al., 2008). The comparisons were based on a
CHINESE AUTHOR RECOGNITION TEST 7
Copyright © 2013 UKLA
Likert scale ranging from 0 to 3, with higher numbers indicating a greater amount than
their peers. An average score was computed.
Chinese Author Recognition Test. The nal version of the CART described previously
was used. The instrument includes 75 real authors (e.g., 36 CART-popular and 39
CART-highbrow) and 75 foils. Seven kinds of scores were calculated: CART
score, PPK-composite, PPK-popular, PPK-highbrow, SPK-composite, SPK-popular and
SPK-highbrow. The Cronbachsαfor the CART score was .93; for PPK-composite, .91;
for PPK-popular, .83; for PPK-highbrow, .89; for SPK-composite, .80; for SPK-popular,
.73; and for SPK-highbrow, .70.
Reading ability measures
Vocabulary Size Test. Nearly half a century ago, the National Institute for Compilation and
Translation (NICT) collected characters from reading materials used in elementary schools
and rank-ordered these characters according to frequency of use from 1 to 4,708, with 1
being the most and 4,708 being the least frequently used (NICT, 1967). The ranking
largely holds nowadays. On the basis of the NICT norm, Hue (2003) developed several
Vocabulary Size Tests for college students in Taiwan. The Vocabulary Size Test we used
for this present study was revised from one of his tests. This test consists of 66 Chinese
characters sampled from a standard Chinese dictionary (Kao, 1985), which consists of
more than 10,000 character entries. These characters belong to four frequency levels: 16
characters from level 1, which ranks between 1 and 1,500 in the NICT character-frequency
norm; 14 characters rank between 1,501 and 3,000 (level 2); 15 characters rank 3,001 and
above (level 3); and 21 were characters not listed in the NICT norm but were listed in the
dictionary (level 4). Participants were rst asked to respond to each of the test characters by
writing down both its pronunciation, using the Mandarin Phonetic Alphabet, and its
meaning, by either giving a denition or using the character to compose a word or phrase.
Then, they were asked whether they had given that particular response because they (1)
had learned the character before, (2) had seen the character before, (3) guessed from the
radical or phonetic component contained in the character or (4) simply guessed. A correct
response to a test item with a rst or second type reason chosen was regarded as a correct
response, whereas a correct response with a third or fourth type reason chosen was
regarded as a correct guess. The number of characters that the participants knew was
estimated (Est) by rst subtracting the proportion of correct guesses from the proportion
of correct responses for each frequency level (x
1
,x
2
,x
3
,x
4
) and then extrapolating the
gure computed to the number of characters in the dictionary within the specied
frequency level, and nally adding up the numbers computed for the four frequency levels.
The formula for the calculation was as follows:
Est ¼
x1
16 1500 þ
x2
14 1500 þ
x3
15 1708 þ
x4
21 4708
Reading Comprehension Test. Because there is no standardised test for measuring reading
comprehension for college students in Taiwan, a Reading Comprehension Test developed
by Chen and Su (2010) with college students from a university in Northern Taiwan was
used for the purpose of this study. Participants had 30 minutes to read eight short vignettes
8 CHEN and FANG
Copyright © 2013 UKLA
and answer 30 corresponding comprehension questions. The questions cover both
literal (e.g., lexical access and parsing) and inferential (e.g., integration, summarisation
and analogy) comprehension. The internal consistency reliability reported in the test
manual is .76.
General Scholastic Ability Test-Chinese. This test is taken as a standard means of entry into
Taiwanese universities and colleges. In January of each year, senior high-school students
take the GSAT, which assesses their basic competence in Chinese, English, mathematics,
and the natural and social sciences. They then seek recommendations from their school or
make their own applications to institutions of their choice. All examinees should take the
whole set of tests. The score for the GSAT-Chinese subtest was used in this study as a gen-
eral measure of achievement in reading and reading-related domains. The GSAT-Chinese
is composed of two types of tasks: multiple-choice test items and short essay-writing tasks.
The two types of tasks share the same proportion of score: 54 to 54, with a total score of
108. Then, the raw scores are transformed into scores of ranking, ranging from level 1
to level 15, with a score of level 15 indicating being in the top 1% of examinees. In the
multiple-choice section, around half of the items are related to modern essays, and the
others are related to classical literature. You (2012) examined the GSAT-Chinese test items
related to modern essays from 2000 to 2011 with the framework of PISA as different
levels/processes of reading comprehension: retrieving information, forming a broad
understanding, developing an interpretation, and reecting on and evaluating the form of
a text. He found that about 32% of the items could be categorised as assessing students
abilities of forming a broad understanding, 64% as developing an interpretationand
the remaining 5% as relating to relevant literacy knowledge, a category not within the
framework of PISA. According to You, test items related to classical literature assess not
only studentsreading comprehension but also their prociency in ancient Chinese
language and culture. As for the short essay-writing tasks of GSAT-Chinese, Tseng
(2008) randomly selected examineeswriting scripts from 2002 to 2005, and the analyses
revealed that these studentsreading comprehension ability had a great impact on their
writing performance. Specically, failures in the writing tasks were mainly due to
misunderstandings of the rubrics or of the reading passages serving as prompts for writing.
To sum up, GSAT-Chinese measures various processes of reading comprehension,
knowledge of ancient Chinese language/culture and writing abilities.
Department Required Test-Chinese. Standard means of entry into Taiwanese universities
and colleges are held in July. Those who did not apply or get recommended for admission
to the institution in January can take a Department Required Test (DRT), depending on the
requirements of the particular college or university. This test is designated to assess more
advanced knowledge and ability than the GSAT and consists of 10 subjects, namely,
Chinese, English, Mathematics, History, Geography, Citizenship and Society, Physics,
Chemistry and Biology. The score for the DRT-Chinese subtest was also used in this
study as a general measure of achievement in reading and reading-related domains. The
DRT-Chinese is also composed of two types of tasks: multiple-choice test items and short
essay-writing tasks, with 55 points for the former and 45 points for the latter, totalling
100 points. In the multiple-choice section, around one-third of the items are related to
modern essays, and the others are related to classical literature. You (2012) examined
DRT-Chinese test items related to modern essays from 2000 to 2011, with the framework
of PISA, and found that about 6% of the items could be categorised as assessing students
CHINESE AUTHOR RECOGNITION TEST 9
Copyright © 2013 UKLA
abilities of retrieving information, 25% as forming a broad understanding, 50% as
developing an interpretationand 19% as reecting on and evaluation of the form of a
text. Similar to GSAT-Chinese, the test items of DRT-Chinese related to classical
literature assess not only studentsreading comprehension but also their prociency in
ancient Chinese language and culture. As for the writing tasks of DRT-Chinese, Pan
(2009) pointed out that it places more emphasis on the abilities of integration and
expression than the GSAT-Chinese. To sum up, DRT-Chinese also measures various
processes of reading comprehension, plus knowledge of ancient Chinese language/culture
and writing abilities. However, for the purpose of screening students more apt for the
Humanities and Social Sciences, DRT-Chinese is designed to assess more advanced
knowledge and ability than the GSAT-Chinese with the following three features: (1) a
higher proportion of multiple-choice items related to classical literature; (2) modern
literature-related multiple-choice items measuring higher levels of reading comprehension;
and (3) more difcult writing tasks.
Results
Table 1 presents the N, the range of scores, the means and the SDs of the primary measures
taken in this study. Table 2 presents a matrix displaying correlations among all the
variables investigated. Among the seven CART measures, the CART score and three
PPK measures (popular, highbrow and composite) were signicantly correlated with most
of the self-reported reading habits, with almost all of the CRH measures and with most of
the reading measures, whereas hardly any of the three SPK measures were signicantly
correlated with the other variables. In addition, negative correlations were found between
the PPK and the SPK scores. Among the PPK measures, although the PPK-popular score
was signicantly associated with reading frequency on both print-based and Internet-based
materials, the PPK-highbrow score was signicantly linked to reading frequency only for
print-based materials. Furthermore, the PPK-highbrow score appeared to have higher
correlations with CRH and reading measures than the PPK-popular score.
Among the four self-reported reading habits, we found that reading of print-based
materials had signicant correlations with CRH, PPK and most reading performance
measures, whereas reading of Internet-based materials had little or no signicant
correlations with these variables. Regarding CRH, the composite score was found to have
signicant correlations with PPK and most reading performance measures.
To further explore these relationships, a factor analysis was performed. Table 3 provides
the factor loadings of a principal component analysis after varimax rotation for the mea-
sures used in the present study. Six factors were extracted, using both the Scree test and
Kaisers rule of eigenvalues greater than 1. The combination of the six factors extracted
accounted for 73.77% of the variance in the measures of respondentsreading performance
and print exposure. CART-total, PPK-composite, PPK-popular and PPK-highbrow
clustered under the rst factor; SPK-composite, SPK-popular and SPK-highbrow clustered
under the second factor; two self-reported reading habits, for print-based materials and
CRH, clustered together under the third factor; both general measures of reading
achievement (i.e., GSAT-Chinese and DRT-Chinese) clustered under the fourth factor;
two specic reading measures (vocabulary size and reading comprehension) clustered un-
der the fth factor; and nally, two self-reported reading habits related to Internet-based
materials clustered under the sixth factor. In other words, the factor analysis indicated that
10 CHEN and FANG
Copyright © 2013 UKLA
there are several dimensions along which reading performance and print exposure can be
measured, all of which seem to capture slightly different aspects of reading.
Tables 4 and 5 present two sets of hierarchical regression analyses examining the
relative extent to which various kinds of print exposure measures predict college students
two general reading performances. GSAT-Chinese assesses studentsbasic competence of
reading comprehension (i.e., forming a broad understanding and developing an interpreta-
tion), knowledge of ancient Chinese language/culture and writing abilities. In Table 4,
Model 1 reveals that vocabulary size and reading comprehension (i.e., lexical access,
parsing, integration, summarisation and analogy) scores altogether helped predict 4.7%
of the variance in performance on the GSAT-Chinese test. Model 2 shows that in addition
to vocabulary size and reading comprehension, self-reported reading habits still contribute
to a signicant increase in the overall model t. Model 3, however, reveals that
when CRH was entered, the unique contribution of self-reported reading habits was
reduced to non-signicance. Moreover, when objective print exposure level measured with
the PPK-composite score was entered, as in Mode1 4, the contribution of both self-reported
reading habits and CRH was reduced to non-signicance (R
2
= .11).
Table 1. Mean scores (with SDs) of reading and print exposure measures.
N
Min.
possible
Max.
possible
Obtained
range Mean SD
Self-reported reading habits
(reading frequency of ction, non-ction,
newspapers and magazines)
352 0 3 03 1.53 .53
CRH-total 354 0 3 03 1.64 .55
CART score 358 75 75 068 35.87 12.59
CART-correct identications 358 0 75 069 39.33 13.34
CART-foils checked 358 75 0 57 to 0 3.46 6.54
PPK-composite 358 0 75 057 17.72 10.50
PPK-popular 358 0 36 033 7.28 4.89
PPK-highbrow 358 0 39 026 10.44 6.92
SPK-composite 358 0 75 044 15.54 7.43
SPK-popular 358 0 36 022 7.89 4.44
SPK-highbrow 358 0 39 024 7.65 4.29
Vocabulary Size Test total 338 0 9416 05604 4365.16 579.09
Frequency level 1 338 0 1500 01500 1464.50 135.65
Frequency level 2 338 0 1500 01500 1427.09 154.14
Frequency level 3 338 0 1708 01366 836.82 212.96
Frequency level 4 338 0 4708 01345 636.47 262.89
Reading Comprehension Test 358 1 30 829 23.37 3.26
GSAT-Chinese 331 1 15 915 13.22 1.16
DRT-Chinese 186 0 100 3988 67.65 9.42
Notes: CRH, comparative reading habits; CART, Chinese Author Recognition Test; PPK, primary print knowl-
edge; SPK, secondary print knowledge; GSAT-Chinese, General Scholastic Ability Test-Chinese; DRT-Chinese,
Department Required Test-Chinese.
CHINESE AUTHOR RECOGNITION TEST 11
Copyright © 2013 UKLA
Table 2. Correlations among self-reported reading habits, comparative reading habits, objective print exposure and reading measures.
Self-reported reading habits Comparative reading habits (CRH) Objective print exposure (CART measures) Reading measures
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
1
2 .22
**
3 .39
**
.11
*
4 .06 .15
**
.22
**
5 .41
**
.04 .48
**
.11
*
6 .31
**
.04 .39
**
.06 .64
**
7 .32
**
.00 .28
**
.05 .59
**
.49
**
8 .35
**
.02 .43
**
.09 .59
**
.47
**
.43
**
9 .24
**
.00 .26
**
.16
**
.42
**
.42
**
.36
**
.46
**
10 .21
**
.09 .15
**
.08 .33
**
.33
**
.43
**
.37
**
.35
**
11 .41
**
.04 .45
**
.12
*
.81
**
.77
**
.75
**
.75
**
.69
**
.62
**
12 .13
*
.03 .35
**
.22
**
.36
**
.18
**
.22
**
.27
**
.18
**
.12
*
.30
**
13 .18
**
.08 .30
**
.30
**
.25
**
.12
*
.10 .25
**
.23
**
.06 .21
**
.54
**
14 .26
**
.06 .40
**
.09 .45
**
.27
**
.30
**
.38
**
.26
**
.19
**
.42
**
.63
**
.57
**
15 .26
**
.07 .40
**
.20
**
.41
**
.23
**
.24
**
.36
**
.27
**
.15
**
.38
**
.67
**
.84
**
.92
**
16 .09 .05 .08 .02 .02 .01 .02 .09 .09 .02 .03 .33
**
.18
**
.10 .15
**
17 .04 .04 .01 .11
*
.04 .08 .04 .06 .07 .04 .07 .19
**
.06 .21
**
.17
**
.45
**
18 .08 .05 .05 .05 .01 .04 .01 .09 .09 .04 .06 .30
**
.14
**
.19
**
.19
**
.86
**
.85
**
19 .15
**
.01 .16
**
.03
*
.11
*
.13
*
.03 .14
*
.12
*
.01 .13
*
.23
**
.20
**
.27
**
.26
**
.00 .00 .00
20 .06 .02 .07 .12 .08 .09 .02 .16
**
.09 .10 .12
*
.20
**
.09 .12
*
.12
**
.07 .07 .08 .27
**
21 .15
**
.03 .15
**
.01 .20
**
.12
*
.16
**
.11 .19
**
.12
*
.20
**
.21
**
.24
**
.27
**
.29
**
.07 .08 .09 .17
**
.16
**
22 .21
**
.01 .23
**
.04 .14 .11 .11 .15
*
.15
*
.11 .14 .03 .24
**
.30
**
.31
**
.04 .04 .00 .32
**
.13 .43
**
Notes: 1, Time spent reading Print-based; 2, Time spent reading Internet-based; 3, Reading frequency of ction, non-ction, newspapers and magazines; 4, Reading frequency of e-ction,
e-news, blogs and Bulletin Board Systems; 5, comparative reading habits (CRH) time spent reading; 6, CRH variety of reading; 7, CRH complexity of materials; 8, CRH reading
enjoyment; 9, CRH reading speed; 10, CRH comprehension of materials; 11, CRH total; 12, Chinese Author Recognition Test (CART) score; 13, PPK-popular; 14, PPK-highbrow;
15, PPK-composite; 16, SPK-popular; 17, SPK-highbrow; 18, SPK-composite; 19, Vocabulary Size Test; 20, Reading Comprehension Test; 21, GSAT-Chinese; 22, DRT-Chinese.
*p<.05; ** p<.01; ***p<.001.
12 CHEN and FANG
Copyright © 2013 UKLA
The DRT-Chinese is designed to assess more advanced knowledge and ability than the
GSAT in the sense that it measures higher levels of reading comprehension (e.g., reecting
on and evaluating the form of a text), knowledge of ancient Chinese language/culture and
Table 3. Principal components factor analysis after varimax rotation.
Factor
123456
Time spent reading Print-based .034 .046 .757 .280 .022 .143
Time spent reading Internet-based .138 .146 .197 .130 .260 .712
Reading frequency of ction, non-ction,
newspapers and magazines
.334 .055 .729 .011 .147 .070
Reading frequency of e-ction, e-news, blogs
and BBSs
.244 .044 .014 .166 .224 .793
CRH-total .220 .026 .763 .011 .065 .022
CART score .780 .299 .218 .018 .104 .020
PPK-popular .831 .123 .039 .207 .025 .183
PPK-highbrow .836 .213 .283 .123 .020 .080
PPK-composite .923 .193 .199 .176 .025 .035
SPK-popular .036 .854 .007 .110 .004 .123
SPK-highbrow .086 .825 .051 .079 .025 .063
SPK-composite .070 .991 .024 .024 .016 .041
Vocabulary Size Test .146 .006 .216 .300 .599 .010
Reading Comprehension Test .025 .024 .003 .024 .844 .000
GSAT-Chinese .138 .081 .078 .796 .025 .029
DRT-Chinese .202 .058 .111 .763 .215 .003
Notes: BBSs, Bulletin Board Systems; CRH, comparative reading habits; CART, Chinese Author Recognition
Test; PPK, primary print knowledge; SPK, secondary print knowledge; GSAT-Chinese, General Scholastic
Ability Test-Chinese; DRT-Chinese, Department Required Test-Chinese.
Table 4. Hierachical regressions of vocabulary size, reading comprehension and print exposure measures on
the General Scholastic Ability Test-Chinese.
Model 1 Model 2 Model 3 Model 4
Vocabulary Size Test .135
*
.118
*
.114
*
.075
Reading Comprehension Test .135
*
.132
*
.122
*
.115
*
Self-reported reading habits
(Reading frequency of ction,
non-ction, newspapers and magazines)
.121
*
.058 .003
Comparative reading habits
(comparative reading habits-total)
.136
*
.091
Objective print exposure
(primary print knowledge composite)
.205
**
R
2
Change .047
**
.014
*
.014
*
.032
**
R
2
.047 .061 .075 .107
*p<.05; ** p<.01; ***p<.001.
CHINESE AUTHOR RECOGNITION TEST 13
Copyright © 2013 UKLA
more sophisticated writing skills. In Table 5, Model 1 reveals that vocabulary size and
reading comprehension scores altogether helped predict 10.5% of the variance in
performance on the DRT-Chinese test, with vocabulary size found to be the only
signicant predictor. Model 2 shows that in addition to vocabulary size and reading
comprehension, self-reported reading habits again contribute to a signicant increase
in the overall model t. Model 3 reveals that CRH failed to make an additional
contribution. As far as objective print exposure is concerned, Model 4 reveals that when
the PPK-composite was entered, it contributed to a signicant increase in the overall model
t(R
2
= .16). Models 41 and 42 were designed to probe the relative extent to which
PPK-popular and PPK-highbrow predict DRT-Chinese. Similar results were found for
these two models.
Discussion
In this study, we constructed a CART-popular subtest as well as a CART-highbrow
subtest, on the basis of the literature that indicates an association between higher ed-
ucational level and cultural consumption (e.g., Chen, 2009; Kraaykampt & Dijkstra,
1999). Three kinds of scores were generated from CART in this study: the original
CART score, the PPK and the SPK. In a previous study, Martin-Chang and Gould
(2008) found that SPK was a signicant predictor of both vocabulary and reading
comprehension, but PPK had more predictive power than SPK because it accounted
for variance in the criterion variables after the effects of SPK had been factored
out. In addition, they found that PPK and SPK were not correlated with each other.
Table 5. Hierachical regressions of vocabulary size, reading comprehension and print exposure measures on
the Department Required Test-Chinese.
Model 1 Model 2 Model 3 Model 4
Model
41
Model
42
Vocabulary Size Test .295
***
.255
**
.254
**
.221
**
.223
**
.225
**
Reading Comprehension Test .082 .078 .078 .090 .084 .090
Self-reported reading habits
(Reading frequency of ction,
non-ction, newspapers and
magazines)
.163
*
.150 .086 .109 .093
Comparative reading habits
(CRH-total)
.027 0.02 .014 .003
Objective print exposure
(PPK-composite)
.202
*
Objective print exposure
(PPK-popular)
.169
*
Objective print exposure
(PPK-highbrow)
.178
*
R
2
Change .105
***
.025
*
.001 .032
*
.025
*
.024
*
R
2
.105 .129 .130 .162 .155 .154
Notes: CRH, comparative reading habits; PPK, primary print knowledge.
*p<.05; ** p<.01; ***p<.001.
14 CHEN and FANG
Copyright © 2013 UKLA
However, in this present study, negative correlations were found between the PPK and
the SPK scores. Moreover, the SPK scores were found to have no correlations with
vocabulary size, reading comprehension, GSAT-Chinese or DRT-Chinese. In other
words, these results appeared to suggest that SPK can be removed from the CART
for college students in the future. In contrast, we found that the PPK scores were sig-
nicantly correlated with all four reading performance measures and that they had the
strongest predictive power for both GSAT-Chinese and DRT-Chinese beyond the joint
contribution of vocabulary size and reading comprehension among the self-reported
reading habits, CRH and objective print exposure variables. These results are consis-
tent with previous ndings that the ART serves as a better index than self-reported
reading habits for providing a relative level of print exposure for college students
(e.g., Burt & Fury, 2000; Chateau & Jared, 2000; Martin-Chang & Gould, 2008;
Stanovich & Cunningham, 1992; Stanovich & West, 1989; Stanovich et al., 1995;
West & Stanovich, 1991). At the same time, these results corroborate previous nd-
ings in the linkage between print exposure and reading achievement, providing more
empirical evidence from Taiwanese college students. Finally, a difference between
CART-popular and CART-highbrow subtests was not supported in the present study
because the results from the regression analysis revealed that they have about the same
predictive power for DRT-Chinese beyond vocabulary size, reading comprehension and
other reading habit variables.
Acheson et al. (2008) proposed a CRH index for print exposure and found it to have a
stronger association with general reading performance than the ART. They suggested that
CRH is a more effective instrument than the ART, as the ART is unable to capture a broad
range of reading experiences, especially with regard to Internet-based materials. Neverthe-
less, our results suggested that there is no correlation between CRH and Internet reading time
and frequency. Our results also suggested that CRH had little or no predictive power for
GSAT or DRT. Future study is therefore needed in this area.
Compared with the results of studies from English-speaking countries, the magnitude
of correlation coefcients among CART measures and reading ability measures, and the
R
2
for the joint contribution of vocabulary size, reading comprehension and various print
exposure measures were quite moderate in this study. There are two probable explana-
tions. First, this study suffered from limitations because of the lack of suitable measure-
ments for assessing the reading-related skills of college students in Taiwan. And second,
tests derived in a Western culture may not be suitable for a population from an Eastern
culture. An ART is bound to be book market-dependent and therefore culture dependent.
Because the CART developed in this study could not cover reading experiences related to
translated books, a major source of reading materials for Taiwanese readers, a comple-
mentary instrument might be needed. Although the application of the CART constructed
in this study should be conned to Taiwanese college students, the process of developing
such an instrument can be shared with other Chinese-speaking regions. Similar proce-
dures can be adopted for other age-appropriate norms or for the general public. Future
investigations across cultures are encouraged.
Acknowledgements
The authors would like to thank the National Science Council of the Republic of China for
nancially supporting this research under Contract No. NSC 100-2420-H-007-001-MY3.
CHINESE AUTHOR RECOGNITION TEST 15
Copyright © 2013 UKLA
References
Acheson, D.J., Wells, J.B. & MacDonald, M.C. (2008). New and updated tests of print exposure and reading
abilities in college students. Behavior Research Methods, 40(1), 278289.
Allen, L., Cipielweski, J. & Stanovich, K.E. (1992). Multiple indicators of childrens reading habits and attitudes:
Construct validity and cognitive correlates. Journal of Educational Psychology, 84(4), 489503.
Burt, J.S. & Fury, M.B. (2000). Spelling in adults: The role of reading skills and experience. Reading and Writing:
An Interdisciplinary Journal, 13, 130.
Chateau, D. & Jared, D. (2000). Exposure to print and word recognition processes. Memory & Cognition, 28(1),
143153.
Chen, S.Y. (2007). Extracurricular reading habits and reading interests of college students in Taiwan: Findings
from two national surveys. Journal of Adolescent & Adult Literacy. 50(8), 642655.
Chen, S.Y. (2009). Adult Taiwanese book readerspractices: The role of age, gender, education and reading for
enjoyment, knowledge, relaxation and social conversation. Paper presented at AERA Annual Meeting, San
Diego, CA, 1317 April.
Chen, S.Y. & Fang, S.P. (2012). College studentsreading practices and proles in both print-based and Internet-
based format. National Science Council, Taipei, Taiwan: Report for project NSC 100-2420-H-007-001-MY3.
Chen, J.-L. & Su, Y.-F. (2010). The effects of different reading strategy instructions on low reading ability college
studentsreading strategy performance: Evidences from think aloud data. Paper presented at the 20th annual
meeting of the Society for Text & Discourse, Chicago, IL.
Cipielewski, J. & Stanovich, K.E. (1992). Predicting growth in reading ability from childrens exposure to print.
Journal of Experimental Child Psychology, 54, 7489.
Cunningham, A.E. & Stanovich, K.E. (1990). Assessing print exposure and orthographic processing skill in
children: A quick measure of reading experience. Journal of Educational Psychology, 82(4), 733740.
Ecalle, J. & Magnan, A. (2008). Relations between print exposure and literacy skills: New evidence from Grade
15. British Journal of Developmental Psychology, 26(4), 525544.
Gallik, J.D. (1999). Do they read for pleasure? Recreational reading habits of college students. Journal of
Adolescent & Adult Literacy, 42(6), 480489.
Guthrie, J.T. (1981). Reading in New Zealand: Achievement and volume. Reading Research Quarterly, 17, 627.
Hue, C.-W. (2003). Number of characters a college student knows. Journal of Chinese Linguistics, 31, 300339.
Jolliffe, D.A. & Harl, A. (2008). Texts of our institutional lives: Studying the reading transitionfrom high school
to college: What are our students reading and why? College English, 70(6), 599617.
Kao, S.F. (1985). Guomin Biaozhun Zidian [Guo-Min Standard Character Dictionary]. Taipei, Taiwan:
Cheng-Chung.
Kraaykamp, G. & Dijkstra, K. (1999). Preferences in leisure time book reading: A study on the social
differentiation in book reading for the Netherlands. Poetics, 26, 203234.
Liu, Z. & Huang, X. (2008). Gender differences in the online reading environment. Journal of Documentation,
64(4), 616626.
Martin-Chang, S.L. & Gould, O.N. (2008). Revisiting print exposure: Exploring differential links to vocabulary,
comprehension and reading rate. Journal of Research in Reading, 31(3), 273284.
Masterson, J. & Hayes, M. (2007). Development and data for UK versions of an author and Title Recognition Test
for adults. Journal of Research in Reading, 30, 212219.
McBride-Chang, C. & Chang, L. (1995). Memory, print exposure, and metacognition: Components of reading in
Chinese. International Journal of Psychology, 30(5), 607616.
National Institute for Compilation and Translation (1967). A study on the high frequency words used in Chinese
elementary school reading materials. Taipei, Taiwan: Chung Hwa Book Co., Ltd.
Pan, L.I. (2009). Review and reection on the design of non-multiple-choice test of the Department Required Test
Chinese. Available from: http://www.cere.edu.tw/CeecMag/Articles/177/177-3.htm#.
Stanovich, K.E. & Cunningham, A.E. (1992). Studying the consequences of literacy within a literacy society: The
cognitive correlates of print exposure. Memory and Cognition, 20, 5169.
Stanovich, K.E. & West, R.F. (1989). Exposure to print and orthographic processing. Reading Research
Quarterly, 24, 402433.
Stanovich, K.E., Cunningham, A.E. & West, R.F. (1998). Literacy experiences and the shaping of cognition.
In S. Paris & H. Wellman (Eds.), Global prospects for education: Development, culture and schooling.
(pp. 253288). Washington, DC: American Psychological Association.
Stanovich, K.E., West, R.F. & Harrison, M.R. (1995). Knowledge growth and maintenance across the life span:
The role of print exposure. Developmental Psychology, 31, 811826.
16 CHEN and FANG
Copyright © 2013 UKLA
Tseng, P.-F. (2008). The impact of examineesreading comprehension ability on their performance in the General
Scholastic Ability Chinese test. Bulletin of Testing and Assessment,5,5387.
Van Eijck, K. (1999). Socialization, education, and lifestyle: How social mobility increases the cultural
heterogeneity of status group. Poetics, 26, 30928.
Van Rees, K., Vermunt, J. & Verboord, M. (1999). Cultural classications under discussion latent class analysis
of highbrow and lowbrow reading. Poetics, 26, 34965.
Walberg, H.J. & Tsi, S. (1984). Mathew effects in education. American Educational Research Journal, 20,
442451.
West, R.F. & Stanovich, K.E. (1991). The incidental acquisition of information from reading. Psychological
Science, 2(5), 325330.
You, S.H. (2012). The reading tasks on Chinese modern essays of College Entrance Examination (20022011).
Bulletin of Testing and Assessment, (10), 3169.
Su-Yen Chen has a PhD degree from the University of Texas at Austin in the United States and
currently is a Professor at the Institute of Learning Sciences at National Tsing Hua University in
Taiwan.
Sheng-Ping Fang has a PhD degree from the University of California at Riverside in the United
States and currently is a Professor at the Department of Chinese Literature at National Tsing Hua
University.
Received 24 July 2012; revised version received 23 July 2013.
Address for correspondence: Su-Yen Chen, Institute of Learning Sciences, National
Tsing-Hua University, Taiwan. E-mail: suychen@mx.nthu.edu.tw
CHINESE AUTHOR RECOGNITION TEST 17
Copyright © 2013 UKLA
... Reading comprehension at Wave 2 Children's reading comprehension ability was assessed using two 45-min Chinese reading comprehension tasks, which included narrative and non-narrative passages, all previously used in studies (Chen & Fang, 2015;Tong & McBride, 2017). Each task consisted of four passages, accompanied by questions designed to assess proficiency in word interpretation, main idea comprehension, context inference and summarization. ...
... One point was awarded for each correct answer, and the maximum score for this session was 59 (Cronbach's α = .78). For the second task, children read four passages adopted from Chen and Fang (2015). Two passages were around 500 characters, and the other two were about 700 characters. ...
Article
Full-text available
Background and Aims This study examined how parenting styles influence children's language skills and reading comprehension. Materials and Methods Six hundred and eighty‐five Chinese‐speaking third graders ( M age = 9.23 years, SD = .66; 341 girls) were randomly recruited from eight primary schools. We measured students' primary caregivers' parenting styles, parental education, family income, migration status, number of children's extracurricular books in the home and family cohesion at Wave 1 (i.e. grade 3). We also assessed students' reading motivation, language skills (i.e. vocabulary knowledge and syntactic awareness) and reading comprehension at Wave 2 (i.e. grade 4). Results The structural equation model analysis revealed that parenting styles indirectly affect language and reading comprehension development, with authoritative and authoritarian parenting exerting different influences on the two outcomes. Specifically, authoritative parenting was positively related to the number of children's books, which in turn was directly, or indirectly through reading motivation, associated with children's language skills and reading comprehension. In contrast, authoritarian parenting was negatively related to family cohesion, which was associated with children's reading motivation, and consequently, their language skills and reading comprehension. The multigroup analysis showed that the indirect pathways varied slightly across parental migration statuses. Discussion and Conclusion These findings enhance the global understanding of the pathways linking parenting styles to children's language skills and reading comprehension, suggesting that educators and researchers should not overly emphasize the direct role of parenting styles in children's academic performance.
... As we can see in this brief overview, vocabulary knowledge and reading appear to be in a relationship. While there are author recognition tasks designed for English (Acheson et al. 2008), Spanish (De la Garza, in preparation), Chinese (Chen and Fang 2015), Korean (Lee et al. 2019), and Dutch (Brysbaert et al. 2020), such a task for Turkish has not yet been developed. Similarly, although vocabulary size tests were developed for various other languages, Turkish does not yet have a vocabulary size test. ...
Article
Full-text available
This article reports the development of two novel research tools for Turkish, the Turkish Author Recognition Task (TART) and the Turkish Vocabulary Size Test (TurVoST). Such tools have been readily available for English, Spanish, Korean, Dutch and Chinese but not for Turkish. These tools help researchers to identify the print exposure levels of L1 speakers and an approximation of L1 speakers’ receptive vocabulary knowledge, respectively. Measuring print exposure is important as it is an important driver of L1 development from a usage-based perspective (e.g., Dąbrowska in Cognition 178:222–235, 2018), which influences vocabulary, grammar, and collocation knowledge. The findings show that the TART and TurVoST are significantly correlated at 0.47 and the TART accounts for almost 18% of the variance in vocabulary knowledge. Reliability (Cronbach’s alpha) scores were found to be 0.99 and 0.74 for two tests respectively. In light of similar previous studies of various ARTs and vocabulary size tests, the TART and the TurVoST are found to be reliable research instruments with correlations and reliability scores within the range of what has been reported in the literature. Potential uses of these two instruments are discussed. All data, R codes, and research instruments are publicly available at https://osf.io/u6t8m/?view_only=63cf706c381a4214950984dae5470df6.
... In their seminal work on the ART, Stanovich and West (1989) did not opt for canonical authors in selecting the real author names, but looked at recent sales figures. This has become the standard in later tests developed by other researchers (Acheson et al., 2008;Chateau & Jared, 2000;Chen & Fang, 2015;Grolig et al., 2020;Masterson & Hayes, 2007). Stanovich and West argued against selecting "highbrow" authors, expecting them to be known only by the "most highly educated or academically inclined readers" (Stanovich & West, 1989, p. 408). ...
Article
Full-text available
Reading is a key skill for university students. The Author Recognition Test (ART) and the Title Recognition Test (TRT) have both been used for decades to measure print exposure which correlates with reading and other linguistic skills. Given the available evidence for interindividual differences in reading skills, this study addresses three open issues. First, to what extent do ART and TRT scores correlate with individual differences regarding students’ study programs? Second, how do these results correlate with the self-reported time spent reading different types of text genres (e.g., fiction, nonfiction) per week? And third, this study compares ART and TRT to one another. We tested students from six study programs in the humanities and (medical) sciences which vary in the amount and kind of reading material required for study success. We found that students perform significantly differently in the ART and the TRT depending on their field of study. Students in a study program focusing on fiction and literature perform best overall. We also replicated the well-known effect of age on ART and TRT scores: older students have better scores. We did not find reliable effects of reading time on test performance, whereas individual creative writing habits did positively predict ART/TRT test results. These results raise a number of important questions regarding the ART/TRT in general and regarding interindividual differences in personal reading and writing habits and the change in reading habits in times of media convergence.
... On the other hand, there is evidence suggesting a potential positive relationship between digital reading habits and comprehension. Some studies evidenced positive associations between digital reading frequency and reading comprehension in adolescents (Botello-Peñaloza & López-Alba, 2014;Islam et al., 2020;Jackson et al., 2011;Johnson, 2013;Kong et al., 2022;Wu & Peng, 2017) as well as in college students (Amiama-Espaillat & Mayor-Ruiz, 2017;Chen & Fang, 2013;Mežek, 2013). This could be partially explained because digital reading presents exclusive features that we cannot encounter in print materials, such as interactivity, nonlinearity, immediacy in accessing information, and the convergence of text, images, audio, and video (Chen & Chen, 2014). ...
Article
Previous research has evidenced a strong positive relationship between leisure print reading habits and reading comprehension across the lifespan. The rapid evolution of new forms of leisure digital reading could modify such a relationship. This meta-analysis extends previous research by analyzing the relationship between leisure digital reading habits and reading comprehension. We analyzed 40 effect sizes using multilevel analysis. Data involved 469,564 participants from studies published between 2000 and 2022. The average effect size reflects a small significant effect on reading comprehension (r = .055), which contrasts with the medium size effects found in the literature related to print reading habits and comprehension. This relationship is significantly moderated by the reader’s educational stage. At early stages (primary and middle school) negative relationships are observed between leisure digital reading and text comprehension, while at later stages (high school and university) the relationship turns positive. We highlight the different contributions that reading modalities and technological contexts have on our reading comprehension, especially across the lifespan. In sum, leisure digital reading does not seem to pay off in terms of reading comprehension, at least, as much as traditional print reading does.
... Following the original development of an ART reflecting largely English/American literature, versions tailored to speakers of other dominant languages have emerged. These include the version of the ART in Mexican/Spanish (Rodrigo et al., 1996), Chinese (Chen & Fang, 2015), Korean (Lee Brysbaert et al., 2020), and German/European languages (GART; Grolig et al., 2020). To complement our current study of the ART (McCarron & Kuperman, 2021), we also include publicly available item response data from the DART and GART in our investigation. ...
Article
Item response theory (IRT) analyses are often used to evaluate measurement error in educational and psychological test instruments. In such contexts, the latent traits/proficiencies are typically assumed normally distributed and a cumulative normal/logistic measurement link function is applied. Such choices are consistent with constructs that are viewed as bipolar in nature and play a critical role in defining the latent proficiency metric against which the measurement error in the test is evaluated. Recently, alternative models that portray the construct as unipolar have been highlighted as being more appropriate for certain psychopathology and personality constructs. In this paper we extend consideration of unipolar IRT models for a recognition task measure, using several example datasets from various versions of the Author Recognition Test (ART), a measure of print exposure. We show how the decision between unipolar versus bipolar IRT modeling has substantial implications for the quantification and interpretation of measurement error in the ART. In sharp contrast to prior bipolar IRT analyses of the ART, under unipolar IRT measurement error in the ART is minimized at low levels of latent print exposure, and increases as latent print exposure increases. Implications for consideration of unipolar IRT with other constructs and measures (e.g., vocabulary, specialized forms of knowledge) that reflect a similar type of response process are considered in the discussion.
Article
Objective To establish convergent and discriminant validity for a combined measure of print exposure (i.e., Author Recognition Test and Magazine Recognition Test [ART/MRT]) and assess its potential utility for estimating premorbid cognitive functioning. Method Community-dwelling older adults (N = 84; 95% non-Hispanic White) completed the ART/MRT, Test of Premorbid Functioning (ToPF), Dementia Rating Scale – 2nd Edition (DRS-2), Hopkins Verbal Learning Test – Revised (HVLT-R-DR), and select subtests from the Delis-Kaplan Executive Functioning System (D-KEFS) as measures of executive functioning (i.e., D-KEFS-EF) and processing speed (i.e., D-KEFS-PS). Pearson correlations and linear regressions were used to examine the relationships between the ART/MRT, cognition, and demographics. Results Cognitive scores, with the exception of HVLT-R-DR, were positively correlated with ART/MRT score such that better cognitive performance was associated with greater print exposure (range r = 0.39–0.49). ART/MRT score was positively correlated with years of education and negatively correlated with age. ToPF and DRS-2 differentially and uniquely predicted ART/MRT score beyond the other cognitive and demographic variables and beyond each other. Conclusions Findings indicate that measures of print exposure reflect crystallized knowledge but may also capture fluid abilities that may be more vulnerable to age-related decline or neurodegeneration. Assessment of print exposure may offer an alternative to word reading measures that may be inappropriate for translation into other languages and for use with individuals with certain language difficulties.
Article
Full-text available
The paper reviews the methods for assessing different components of reading skills in adults with reading difficulties, along with functional reading skills. We are particularly interested in the assessment methods available to researchers and practitioners, developed predominantly in the research context, and not available solely in English. We discuss the large-scale international study, PIAAC, as an example of a framework for such assessments. Furthermore, we cover the following types of assessment tools: (1) self-assessment questionnaires, probing into comprehension difficulties and reading habits; (2) measures of print exposure, such as author recognition tests, correlating with other reading-related skills; (3) measures of word recognition and decoding, including reading aloud of words and pseudowords, as well as silent lexical decision tasks; (4) fill-in-the-blank tasks and sentence reading tasks, measuring predominantly local comprehension, entangled with decoding skills; (5) comprehension of longer reading passages and texts, focusing on functional texts. We discuss comprehension types measured by tests, text types, answer formats, and the dependence problem, i.e., reading comprehension tests that can be solved correctly without reading. Finally, we tap into the new ideas emerging from the AI systems evaluation, e.g., using questions generated from news articles or Wikipedia or asked directly by search engines users. In the concluding section, we comment on the significance of incorporating background information, motivation, and self-efficacy into the assessment of adult literacy skills.
Article
Background It is well evident that individuals' levels of print exposure are significantly correlated with their reading ability across languages, and an author recognition test is commonly used to measure print exposure objectively. For the first time, the current work developed and validated a Simplified Chinese Author Recognition Test (SCART) and examined its role in explaining Chinese online reading. Methods In Study 1, we constructed the SCART for readers of simplified Chinese and validated the test using data collected from 203 young adults in Mainland China. Participants were measured on the SCART and three self‐report tasks about their reading experiences and habits. Study 2 recruited additional 68 young adults in Mainland and measured their print exposure (with the same tasks used in Study 1), reading‐related cognitive ability (working memory, rapid automatic naming, Chinese character reading, and vocabulary knowledge), and Chinese online reading via an eye‐tracking passage reading task. Results Results of Study 1 support the high reliability and validity of the SCART. Results of Study 2 indicate that SCART scores significantly predicted participants' online reading processing while controlling for subjective reading experiences and habits, and reading‐related cognitive abilities. Across two studies, we found converging evidence that the in‐depth recognition of the authors (i.e., participants have read the books written by these authors) appears to be a better indicator of print exposure than the superficial recognition of the author names. Conclusions Taken together, this work filled in the gap in the literature by providing an evidence‐based, objective print exposure measure for simplified Chinese and contributes to a broader understanding of print exposure and online reading processing across different writing systems.
Article
Full-text available
Linguistic experience plays a clear role in accounting for variability in sentence comprehension behavior across individuals and across sentence types. We aimed to understand how individual differences in reading experience predict reading behavior. Corpus analyses revealed the frequencies with which our experimental items appeared in written and spoken language. We hypothesized that reading experience should affect sentence comprehension most substantially for sentence types that individuals primarily encounter through written language. Readers with more text exposure were faster and more accurate readers overall, but they read sentence types biased to written language particularly faster than did readers with less text exposure. We see clear effects of text exposure on sentence comprehension in ways that allow explicit links between written and spoken corpus statistics and behavior. We discuss theoretical implications of effects of text exposure for experience-based approaches to sentence processing.
Article
Full-text available
The Chinese people have great regard for those who read widely, yet little is known of the extracurricular reading behaviors of Chinese students. This study drew on data from two national surveys to investigate the amount of time Taiwanese college students spend on extracurricular reading. Findings are interpreted in relation to prior research on the reading habits of college students internationally. The study found that Female students on average did not spend more time on extracurricular reading than males Students from public institutions, who generally have higher academic competence, did not spend more time reading than students from private institutions Education majors spent the least time on extracurricular reading Newspapers, magazines, and bestsellers were the most popular reading materials, but manga (graphic novels) had medium popularity Crosscultural differences might have an impact on the amount of time spent reading and on reading interests
Article
Full-text available
To test the hypothesized cumulative advantages of educative factors, the science-achievement scores on a 69-item test of science knowledge of 1,284 young adults, ages 26 to 35, surveyed by the National Assessment of Educational Progress (NAEP) in 1977, were regressed on three composite independent variables: motivation and prior and current educative experiences. The test scores were related significantly to prior experience-embodied variables, such as parental socioeconomic status, respondent education, and specific scientific training, as well as to motivation to learn and current amount and intensity of information acquisition, such as news media exposure and reading. Early educative experience predicts current educative activities and motivation; and all three factors contribute significantly and independently to the prediction of achievement.
Article
Full-text available
In this paper, we will report on our study of how social mobility affects patterns of cultural consumption. The main questions concern the extent to which social mobility leads to more heterogenous consumption patterns within status groups, and the manner in which the consumption of culture, both popular and highbrow, is shaped by the joint impact of a person's current status and his or her status of origin. Since consumption is indeed determined by both parental schooling and respondent's level of education, the results show that, with respect to legitimate culture (reading literature, visiting museums and attending performing arts), cultural heterogeneity among the higher educated can be explained to a substantial degree by differences in family background. For popular culture (reading romantic fiction, attending soccer games and watching commercial television) the consequences of parental background are less clear-cut. The results strongly suggest that the influx of upwardly mobile persons among the higher educated diminishes the association between social status and cultural consumption.
Article
The authors discuss a survey of reading practices that they administered to students at their home institution, the University of Arkansas, as well as logs that students at the school kept of their daily reading acts. An important finding was that, contrary to possible belief, students at this university are reading quite a bit, although they are not spending much time on materials assigned in their courses. The authors propose some methods for boosting students’ interest in academic texts, and they call for other institutions to conduct similar studies.
Article
Does a relationship exist between academic achievement and the time spent in recreational reading? A significant connection was found between achievement and the time these college students spent reading for pleasure during vacations.
Article
Reading comprehension of students in two age groups, 14 and 18, is compared for three countries. Based on IEA data reanalyzed by Purves, the following rank order occurs: New Zealand, United States, Iran. The concept of reading volume is proposed as an indicator of the quantity of reading among adults and students to account partially for these differences among countries. A metric is devised for reading volume, based on the number of copies (of newspapers, magazines, and books) read per person per month. Data derived from six national and international surveys, comparable in method, purpose, and sampling of the populations, are used to calculate reading volumes, with the result that New Zealand was higher than the United States which was higher than Iran. In addition, reading volumes of subgroups of New Zealanders divided by age, sex, occupation, TV viewing, and recreational interest are presented. These findings are placed in an educational framework in which desired levels of reading achievement for students depend on reading demands of society which are (minimally) estimated by reading volume. The symbiosis of achievement and volume is discussed in the contexts of educational planning and assessment./// [French] On a comparé dans trois pays la compréhension de lecture d'éléves de deux groupes d'âge: 14 et 18 ans. Basé sur les données du IEA réanalysées par Purves, l'ordre suivant a lieu: Nouvelle Zélande, Etats-Unis et Iran. On a proposé le concept de volume de lecture comme un indicateur de qualité de lecture parmi les adultes et les éléves afin d'expliquer en partie ces différences parmi les pays. On a dispoé un métrage pour le volume de lecture, basé sur le nombre de copies (de journaux, magasines et livres) lues par personne chaque mois. On a utilisé des données provenant de six enquêtes nationales et internationales, comparables en méthode, but et échantillon de populations, afin de calculer les volumes de lecture. Les résultats obtenus sont les suivants: La Nouvelle Zélande a un volume plus grand que les Etats-Unis, le volume de cuex-ci dépasse à son tour celui de l'Iran. De plus, on a présenté les volumes de lecture des sousgroupes de Nouveaux Zélandais divisés en âge, sexe, occupation, choix de programmes télévisés, et loisirs. Ces découvertes ont été placées dans un cadre éducatif selon lequel les niveaux désirés d'accomplissement de lecture des éléves dépendent des demandes sociales de lecture qui sont (de manière minime) estimées par volume de lecture. La symbiose d'accomplissement et de volume est discutée dans les contextes de planification et évaluation d'enseignement./// [Spanish] Se comparan comprensión de lectura de dos niveles de edad, 14 y 18 años, en 3 países. Basado en los datos de la IEA (International Education Association), analizados de nuevo por Purves, el siguiente orden de rangos resultó: Nueva Zelanda, Estados Unidos, Irán. Se propone el concepto de volumen de lectura como un indicador de la cantidad de lectura de adultos y estudiantes, para justificar parcialmente las diferencias entre países. Se desarrolla un sistema de medición para el volumen de lectura, basado en el número de periódicos, revistas y libros leídos por mes y por persona. Datos derivados de 6 encuestas nacionales e internacionales - comparables en metodología, objetivos, y muestreo de población-se utilizan para calcular volumen de lectura, con el resultado de que el de Nueva Zelanda era mayor que el de los Estados Unidos, que al mismo tiempo era mayor que el de Irán. Además, se describen los volúmenes de lectura de subgrupos neozelandeses divididos por edad, sexo, ocupación, hábito de mirar la televisión, y actividades durante las horas de ocio. Estos resultados se insertan en una estructura pedagógica en la que niveles de destreza de lectura de alumnos, sujetos a la demanda de la sociedad, se calculan (mínimamente) por medio de volumen de lectura. Se discute la simbiosis de destreza y volumen en el contexto de planificación pedagógica y medición.
Article
In a study of ninety college seniors, measures of exposure to print predicted individual differences in vocabulary and cultural literacy after differences in general ability (SAT scores) had been statistically controlled, but did not predict differences in grade point average or knowledge of the field of major (psychology). Measures of exposure to nonprint media (television and films) predicted variance in cultural literacy, but not in vocabulary, after cognitive ability had been controlled. The data challenge the view that knowledge acquisition is determined only by the efficiency of cognitive components that encode and store information. Instead, the results indicate that differences in exposure to information are a significant independent contributor to differences in knowledge across individuals. The investigation further demonstrated the convergent validity of two briefly administered indicators of individual differences in exposure to print.
Article
This study examined the relative contributions of visual and verbal memory, metacognition about reading, and print exposure to reading comprehension among 100 Chinese fifth graders. The four concurrently measured componential skills were substantially associated with reading comprehension and reliably distinguished between good and poor readers. In a hierarchical regression, after controlling for the effects of verbal intelligence, both verbal and visual memory abilities predicted unique variance in reading comprehension. The distinctness of each memory skill underscores the existence of two unique memory processes in Chinese reading. Future research should examine the longitudinal contributions of these cognitive abilities to reading comprehension.
Article
There are estimates of the number of characters a Chinese college student knows, however, none of the estimates was obtained from empirical studies. In the present study, four groups of Chinese characters were stratified sampled from two dictionaries in accord with the characters' linguistic frequencies. One hundred and twenty college students were recruited from two universities, and they were asked to write down the meanings and the pronunciations for a group of sampled characters. For each meaning and pronunciation responded, the subjects were asked to write down the reason that the response was generated. The results of the study indicated that a college student knows about 5150 characters and a set of strategies to guess the pronunciation and the meaning of an unknown character.