Document

The Long-Term Effectiveness of English
Language Instruction at Elementary
Schools
Uematsu, Shigeo
(Kyoto Sangyo University)
[email protected]
ICALLE 2014.10.17
Backgrounds
• This 7-year-study (2007-2013) was an
investigation of the subsequent effect of English
Learning at Elementary School (ELES) on both
English language skills and affective variables of
the Junior High School (JHS: lower middle school)
students .
• The participants were 3500 public JHS students .
Each year, 250 respectively from grades 7, 8, and
9 each year in Neyagawa, Osaka.
Backgrounds (Japan)
• Since April 2011, foreign language activity,
specifically ELES, has been implemented
nationwide in Japan once a week for 5th and
6th graders.
• The aim of the activity is to “form the
foundation of pupils’ communication abilities”
and to “foster a positive attitude toward
communication” through English instruction
(MEXT, 2009).
Is Elementary English activities
contributing to the lower-middle school
English learning?
• Are “English activities” fostering the
“foundation” of English communication?
Back grounds (Neyagawa City where
the research was conducted)
• Neyagawa where this research has been
conducted, English activities is one of the
countries’ research cities: started in 2005 as a
pilot, once a week at grade 5 and 6, and in
2006, the schools lowered the starting grade
to grade 1, teaching 10 hours in grades 1 & 2
and 20 hours in grades 3 & 4.
• Such a small amount!
Backgrounds (School)
• The JHS, which I conducted the research has
two local elementary schools in its’ catchment
area. ES-A and ES-B.
Who teach English in ESs?
Homeroom teachers (HRT) are expected to
teach English to pupils (from 2011 mandatory).
Some could, but most unwillingly. Because they
had no English background.
A native speaker of English Assistant Language
Teacher (ALT) and Japanese Assistant English
teacher (JAT) occasionally helped the homeroom
teacher conduct the lessons.
Budget for English Education
• Neyagawa city funded the English program well
(until 2010) and staffed 17 JATs and 12 foreign
teachers (ALTs). About 870 thousand USD a year
for ELES budget . (Japan only 2000 thousand USD)
• After 2010, the city changed the policy and
started to boost up ICT education. The budget
was sharply cut ( down to 50 thousand USD)and
only two JATs and two ALTs remained in 2011.
Literature Review Japanese ELES
context: Pros
•
•
•
•
•
•
•
•
•
Oller and Nagato (1974): reading & listening,
Megumi et.al (1996): listening,
Matsukawa (1997): interview & story-telling,
Shinohara (1999): speaking,
Yatsuka (2000):listening,
Butler & Takeuchi (2006): listening,
JASTEC (2007): listening & speaking,
Yukawa, Koyama, Sugimoto (2012):generally,
Uematsu (2012):speaking
Literature Review: Cons
•
•
•
•
•
Shirahata (2002): speaking,
Takada (2003,2004): all domain,
National Institute (2009): generally negative,
Hasegawa (2013): listening,
Fennelly et al. (2014):generally negative
Instrumentation
English Proficiency Test
• The JACE (Junior High School Assessment of
Communicative English) battery Level 1 through
Level 3 English proficiency tests were
administered to the grade 7, grade 8 and grade 9
students every year, except 2010.
• The test is made up of three subsections: (a)
vocabulary and grammar (22 items in 2 sections;
maximum 100 points), (b) reading (10 items in 3
sections; maximum 100 points), and (c) listening
(18 items in 2 sections; maximum 100 points).
Oral Interview Test
• The author and the school agreed that
participants from one class in grade 8 take
part in a ten-minute oral interview test.
• The interview test was made up of three
parts: (a) basic conversation, (b) four
questions and answers based on a picture,
and (c) a story-telling task based on the
picture.
JACE test score for G7
Hours amount:70→70→90→120→130
65.00
60.00
55.00
VG
Read
50.00
Listen
45.00
40.00
2007
2008
2009
2011
2012
Analysis of JACE test (G7)
One-way ANOVA was conducted and revealed
• (vocab/grammar) F(4, 1034) = 2.26, p = .06)
• (reading) F(4, 1034) = 1.87, p = .11)
• (listening) F(4, 1034) = 5.34, p = .00)
Bonferroni post hoc test detected
• No significant difference was detected for V/G
• No significant difference was detected for reading
• For listening, 2011 students were outperformed by
2007 students (p = .02, d = -.36 ), 2008 students (p
= .02, d = -.26 ), 2009 students (p = .00, d = -.52 )
JACE test score for G8
Hours amount:35→70→70→110→120
55.00
50.00
45.00
VG
Read
40.00
Listen
35.00
30.00
2007
2008
2009
2011
2012
Analysis of JACE test (G8)
One-way ANOVA was conducted and revealed
• (vocab/grammar) F(4, 1059) = 1.57, p = .18)
• (reading) F(4, 1059) = 2.61, p = .03)
• (listening) F(4, 1059) = 4.70, p = .00)
Bonferroni post hoc test detected
• No significant difference was detected for V/G
• For reading, 2012 students were outperformed by
2011 students (p = .02, d = .-29).
• For listening, 2012 students were outperformed by
2008 students (p = .01, d = -.36), 2011 students (p = .00,
d = -.40 )
Oral Interview Test Score(G8)
16.00
14.00
12.00
10.00
8.00
conv
6.00
story
4.00
2.00
0.00
2007
2008
2009
2011
2012
Analysis of Oral Interview Test (1)
For conversation test, one-way ANOVA was
conducted and revealed
(conversation) F(4, 164) = 5.06, p = .001)
Bonferroni post hoc test detected
2012 students were outperformed by 2008
students (p = .00, d = -1.04) , 2009 students(p
= .04, d = -.61)
Analysis of Oral Interview Test (2)
For story-telling test, one-way ANOVA was
conducted and revealed
• (story-telling) F(4, 164) = 11.83, p = .000)
Bonferroni post hoc test detected
• 2007 students were outperformend by 2008
students (p = .00, d = -1.25), 2009 students (p
= .00, d = -1.39), 2011 students (p = .00, d = -1.36),
2012 students(p = .00, d = -.85)
JACE test score for G9
Hours amount:10→35→70→90→110
75.00
70.00
65.00
VG
Read
60.00
Listen
55.00
50.00
2007
2008
2009
2011
2012
Analysis of JACE test (G9)
One-way ANOVA was conducted and revealed
• (vocab/grammar) F(4, 1016) = 1.18, p = .32)
• (reading) F(4, 1016) = 1.81, p = .12)
• (listening) F(4, 1016) = .63, p = .64)
Bonferroni post hoc test detected
• No significant difference was detected for V/G
• No significant difference was detected for reading
• No significant difference was detected for
listening
Summary of 2007-2009 JACE test
• The 2009 G7 cohort which started the ELES
program earliest, from grade 4, was the most
successful in developing their overall English
proficiency measured by JACE test scores for
vocabulary and grammar, reading, and
listening when compared with 2008 and 2007
students, but they did not reach a statistically
significant level.
• However, this advantage lasted just one year.
Summary of 2007-2009 Oral Interview Test
• Speaking test scores for 2007, 2008, 2009
participants improved to statistically
significant levels each year as the amount of
exposure increased in grade 8. ELES can exert
a powerful effect on fostering the foundation
of communication skills in English.
Summary of 2011-2012 JACE test
• Grade 7 students of year 2011 received the
total amount of 120 hours from grade 2
scored the worst in vocab/grammar, reading,
and listening test scores. (Lower to statistically
significant levels).
Summary of 2011-2012 Oral Interview Test
• 2012 G8 students, who received the most
instruction (120 hours from grade 2) failed to
achieved the best score. They couldn’t break
the plateau, implying that English speaking
ability cannot be improved by just lowering
the starting age.
Summary of Motivational, attitudinal
Questionnaire
• Only one statistically significant difference was
found between 2007 students and 2008
students (p = .05, d = .26). The answers were
very stable for the entire research period.
• The endorsements for F3 Respect for Self and
Others (five items) were lower than others,
implicating that this Factor items were misfit.
What about other variables?
Differences in gender(2012, G7)
62.00
61.00
60.00
59.00
58.00
57.00
56.00
55.00
54.00
53.00
52.00
51.00
男子
女子
VG
Read
Listen
Analysis of gender difference
T-test was conducted and revealed
• (Vocab & Grammar) (t= -.64, df=208, p=.52)
• (Reading) (t= -.94, df=208, p=.35)
• (listening) (t= -2.27, df=208, p=.03)
• Female students scored higher to a statistically
significant level.
Differences in Schools:
ES-A and ES-B
62.00
60.00
58.00
56.00
A校
54.00
B校
52.00
50.00
48.00
VG
Read
Listen
Analysis of School Difference
T-test was conducted and revealed
• (Vocab & Grammar)(t= 2.27, df=197, p=.03)
School A outperformed School B to a statistically
significant level.
• (reading) (t= 1.11, df=197, p=.27)
• (listening) (t= 1.32, df=197, p=.19)
Out-of-school English learning: When to start?
(2012, G7)
70.00
65.00
60.00
55.00
50.00
45.00
VG
40.00
Read
35.00
Listen
30.00
なし(82人)
6年(37人)
5年(30人)
4年(15人)
3年以前(43人)
Analysis of starting age difference
One-way ANOVA was conducted and revealed
• (Vocab & Grammar) F(4, 202) = 9.53, p = .00)
• (reading) F(4, 202) = 5.64, p = .00)
• (listening) F(4, 202) = 8.52, p = .00)
Bonferroni post hoc test detected
• For V/G, there were statistically sig. difs. between None and grade
6(d = -.82 ) , grade 5(d = -.80 ), grade 4(d = -.97 ) , grade 3(d = -.92 )
• For reading, there were statistically sig. difs. between None and
grade 6(d = -.61 ) , grade 5(d = -.50 ) , grade 4(d = -.74 ) , grade 3(d
= -.78 )
• For listening, there were statistically sig. difs. between None and
grade 6(d = -.73 ) , grade 5(d = -.59 ) , grade 4(d = -.89 ) , grade 3(d
= -.98 )
Out-of-school English learning:
How often a week? (2012, G7)
80.00
70.00
60.00
50.00
VG
40.00
Read
30.00
Listen
20.00
10.00
0.00
0回(80)
週1回(72) 週2回(48)
週3回(7)
以上(3)
Analysis of frequency difference
One-way ANOVA was conducted and revealed
• (Vocab & Grammar) F(4, 205) = 8.78, p = .00)
• (reading) F(4, 205) = 7.23, p = .00)
• (listening) F(4, 205) = 7.71, p = .00)
Bonferroni post hoc test detected
• For V/G, there were statistically sig. dif. between None and Once
a week (p = .00, d =-.66 , Twice a week (p = .00, d =-1.01 )
• For reading, there were statistically sig. dif. between None and
Once a week(p = .01, d =-.60 , Twice a week (p = .00, d =-.87 )
• For listening, there were statistically sig. dif. between None and
Once a week(p = .00, d =-.66 ) , Twice a week (p = .00, d =-.91 )
Findings
1)The subsequent effects of ELES could be
identified up to grade 8 in speaking and to grade
7 in listening.
2)Out-of-school English learning exerted great
influence on the vocabulary / grammar, reading,
and listening proficiency measured by JACE test.
3)Gender difference and school difference were
observed to a statistically significant level.
Findings(2)
4) Motivational, attitudinal questionnaire
results were almost the same over years
• If it is carefully administrated (same time,
same date, same school, following manuals)
and is carefully analyzed (preliminary analysis,
such as outlier handlings, using Rasch logitized
scores), extreme differences are unlikely to
emerge.
Findings (3)
• 5) Students’ English proficiency measured by
JACE test and Oral interview test results were
sharply dropped in 2011 probably because the
city drastically cut the budget for English
education.
Conclusion (quantitative)
One can hardly find the subsequent
effectiveness of ELES on JHS students’ skill
domain as a benefit of ELES called “fostering the
foundation,” as long as ELES is administered only
once a week.
What about qualitative analysis?
If one can hardly find a clear
subsequent effectiveness of ELES
on students’ skill domain as a
benefit of ELES, qualitative
approach might shed light.
Interview method
• From 2011, I stated to interview 10 Junior high
school students (from each grade), asking them
reflect on ELES (semi-structured interview).
• I also interviewed several junior high school
teachers, elementary school teachers and
administrators.
• For cording interview data, SPSS Text Analytics for
Surveys (ver. 4.01)was employed .
• Data was analyzed based on Grounded theory
(Creswell, 2006, Strauss & Corbin, 1998)
Interesting findings from the student
interview
Q3 What is the optimal combination of
teaching?
• English native speaker plus Japanese teacher
who handles English well. (11 students)
• Above plus homeroom teacher (2 students)
• English native speaker only (3 students)
Q4 From when to start, how many times a
week?
•
•
•
•
•
•
Once a week from grade 1
Twice a week from grade 1
Twice a week from grade 3
3 times a week from grade 3
Once a week from grade 5
3 times a week from grade 5
1
1
3
8
1
3
6%
6%
18%
48%
6%
18%
Early start was supported by many, however, there
were considerable numbers of ‘mid’ or ‘late’ start.
For example,
• UC (G7, female): I think starting English at the
grade 3 is best. Everybody can enjoy English
activity then. Stuff taught at grades 1 and 2, I
couldn’t just remember. Now my brother in
the grade 3 is preparing his presentation. We
help him practice English. Grade 3 is in my
understanding, that developmental stage.
Q5 Have you ever experienced any difficulty in
leaning English since you entered JHS
•
•
•
•
•
None, thanks to ELES
None
Grammar
Others
Spelling
6
4
2
2
1
40%
27%
13%
13%
7%
Who should teach and how?
(multiple)
• Native speaker of English
15
• Japanese teacher who can translate difficult stuff
8
• Homeroom teacher as ‘observer’, ‘facilitator’ 5
• Homeroom teacher should!
5
• All English is no thank you
3
• My homeroom teacher travelled around the
world and told us stories in English. Very fun! 1
You still like English now?
• Like, very much
• Can’t say which
• Don’t like
16(80%)
4(10%)
4(10%)
Elementary school teachers(1)
• In elementary school, there is no evaluation
for English. So, they were relaxed and enjoyed
English. I feel very happy to hear that
graduates have positive images toward ELES.
The interview scripts of the students made all
my time and work worthwhile. (51, female)
• I feel happy that students understand the
differences between ELES and grammar
learning in JHS.(29, male)
Elementary school teachers (2)
• Actually, I learned that they often encounter
the situation of “I’ve got it now” in JHS (42,
female)
• I really appreciate the opportunity of reading
the transcripts and know that they have a
positive feeling toward ELES and remember
what they did then. Interview transcripts give
me a source of energy for the betterment of
the class, texts and realias (39, female)
JHS school teachers (1)
• When students get stuck in learning English in
JHS, he or she is too serious about studying
English. In order to troubleshoot this kind of
cases, we JHS teachers have to know more
about the curriculum and class activities of
ELES. (25, male)
JHS school teachers (2)
• English taught only in the form of sounds was
finally understood through JHS English classes
and made meaning, wow! (29, male)
• We need to know what was taught in ELES and
what we need to teach.(49, male)
• I think English education should start from
grade 1, but have to pay attention so it
doesn’t create ‘English phobia’ (36, female)
Final Conclusion
• We definitely need ongoing effort and
expense for education. It takes time to bear
fruit.
• Students (Pupils) like to be taught by native
speakers of English with appropriate
assistance through their L1.
• Only once-a-week English can retain as
pleasant memories when there is no
evaluation.
MEXT’s plan toward the 2020 Olympic
• MEXT issued a new plan to boost up English
proficiency of Japanese (2014)
• ELES starts in the grade 3 as an English activity
(with quasi-textbooks and no evaluation) 1-2
times a week.
• In grades 5 & 6, English will be implemented as a
subject (with textbooks and evaluation) 3 times a
week.
Related information
• Uematsu, S. (Forthcoming). "Long-Term Effects
of Learning English: Experiences from
Japanese Elementary Schools." Springer.
• Uematsu, S. (2012).“The Effect of English
Learning in Elementary School on
Students‘ English Language Development in
Junior High Schools." The Journal of Asia
TEFL ,9(4), pp.113-133.
Thank you for listening!
• This research was aided by MEXT KAKENHI
Grant Number (2007-2010):19520530, and
JSPS KAKENHI Grant Number (2011-2013):
23520769
• Contact: [email protected]
• You can download the PPT from
• http://uematsu-shigeo.net/page.do?id=17