• 沒有找到結果。

Longitudinal Analysis of Language Assessment Scores Using ANCOVA 68

在文檔中 from This (頁 70-87)

6.2 ANALYSIS OF LANGUAGE ASSESSMENT DATA . 43

6.2.4 Longitudinal Analysis of Language Assessment Scores Using ANCOVA 68

One year teaching effect

For afl pupils in 2nd cohort 2nd administration

0.0 "BOTH

Longitudinal effect: November 1999 to May 2000

Analytical investigation of all the language assessment scores, however, revealed that there was no interaction of teaching mode and time for the writing scores, the listening or the coordinated oral and listening scores. However there was a significant finding for the oral scores.

In the second cohort, the teaching mode effect was interacting with the time effect for oral scores. The effect was significant (p - 0.003 Eta Squared = 0.063), but the magnitude was very small. This finding suggests that the influence of the NET was significant in producing higher oral assessment scores than those of local teachers or a combination of local teachers and NETs.

In other words, pupils taught by a NET scored significantly higher in the oral assessments than pupils taught by other teachers*

the language assessments. Whenever a significant effect of the analysis involves a teaching mode effect it is reported first, otherwise, the largest effect in the model is reported.

6.2.4.3 Findings

(A) Secondary (i) 1st cohort group

Dependent variables Coordinated Oral &

listening

Coordinated Oral Coordinated Listening Writing

ANOVA Model F-statistic

N/A 15.91

26.5 6.6

ANCOVA Model F-statistic

N/A 28.8

33:8

6.6

Box test p-value or Levene's test

< 0.001

< 0.001

< 0.001 0.002

Significant effect Interaction of School level and teaching mode 1 admin oral

1* admin listen School level

Effect, p-value 0.035

< 0.001

< 0.001

< 0.001

Effect Eta squared

value 0,024 0.225 0.191 0.135

Results in the above table suggest that after controlling the effect of the 1st administration, in the 2nd administration, the interaction effect of school level and teaching mode was significant in the analysis of the coordinated oral and listening scores. In the analyses of other dependent variables, no significant effect of teaching mode effect was observed.

School level

Low

Medium

High

Dependent variables

Oral Listening

Oral Listening

Oral Listening

Teaching effect p-value

0.027 0.001

0.036 0.799 0.003 0.082

Effect Eta squared value

0.06 0.1

0.055 0.007 0.082 0.069

Levene's test 0.077 0.004

0.679 0.344 0.396 0.013

Significant comparison Local to Local > NET to NET Local to Local > Local to NET

NET to NET > Local to Local NETtoNET>LocaltoNET NET to Local > Local to NET

NETtoNET>NETtoLocal NETtoNET>LocaltoNET NET to NET > Local to Local

None

NET toNET>NET to Local Ixx^toNET>NETtoLocal Local to Local > NET to Local NET to Local > Local to Local

p-value 0.009 0.011 0.001 0.004 0.042 0.008 0.034 0.006 0.005 0.013 0.001 0.017 Results in the above table show that after controlling the effect of 1st administration, in the 2nd

administration, the following two-year effects were suggested.

In the two-year period, 1998-2000, all students in low ability schools, who were taught by a local teacher for each of the two years had better oral scores than those taught by either a NET for two years or a local teacher in the first year and a NET in the second Those who were taught by a NET for two years had bettor listening scores than those taught by either local teachers or both local teacher and NETs. Those who were tau^t by a NET m

a local teacher in the second had better listening scores than those taught by a local teacher followed by a NET.

Students in medium ability schools who were taught by a NET for two years, had better oral scores than those taught by a NET and a local teacher, or by local teachers for two years.

All students in high-level schools who were taught by either NETs or local teachers for two years had better oral scores than those taught by a NET and a local teacher. Those who were taught by a NET and a local teacher had better listening scores than those taught exclusively by local teachers.

(IS) I"4 cohort group

Dependent variables Coordinated Oral &

listening

Coordinated Oral Coordinated listening Writing

ANOVA . Model

F-statistics N/A

63 28.7

0.9

ANCOVA Model F-statistics

N/A

37.1 44.9

1.8

Box test p-value or Levene's test

< 0.001

0.127 0.007 0.065

Significant effect Interaction of school level, form level &

teaching mode

Interaction of school level, form level &

teaching mode

Interaction of form level

& teaching mode 1* admin writing

Effect p-value

< 0.001

< 0.001 . 0.001

0.001

Effect Eta squared

value 0.031

0.031 0.029 0.147

Results in the above table shows that after controlling the effect of 1st administration, in the 2nd administration, the interaction effect of school level, fonn level and teaching mode was significant in the analysis of both the coordinated oral and listening scores and coordinated oral scores. The interaction effect of form level and teaching mode was significant in the analysis of the coordinated'listening scores.

Form level

Forml

Form3 School

level Medium

High Low Medium

High

Dependent variables

Oral Listening

Oral T ^$teninci[

Oral Listening

Oral listening

Oral Listening

Teaching effect p-value

0.143 0.272

<0.001 0.855 0.004 0.234 0306 0.006 0.025 0.059

Effect Eta squared value

0.037 0.025 0338 0.004 0.146 0.04 0.022 0.091 0.086 0.062

Levene's test 0.128 0.485 0.741 0.016 0.536 0.636 0.477 0327 0.01 0581

Significant comparison None

None Local>NET Local > Both

None Local>NET

None None NET>Local

NET>Botii Bo!h>Local

None

p-value

< 0.001 0.003 0.003

0.03 0.005 0.025

Results in the above table show that after controlling the students* performance in the 1st

administration, in the 2nd administration, the following effects were suggested.

Fl students in high ability schools who were taught by a local teacher had better oral scores than those taught by a NET or by a combination ofNETs and local teachers. F3 students in low ability schools taught by local teachers had better oral scoresthan those taught by NETs. F3 students hi medium ability schools, who were taught by NET teachers, had better listening scores than those taught by local teachers or a combination of local teachers and NETs. F3

70

students in high ability school who were taught by a combination of local teachers and NETs had better oral scores than those taught by local teachers alone.

(B) Primary pupils (i) 1st cohort group

Coordinated Oral &

listening

Coordinated Oral Coordinated Listening Writing

ANOVA Model F-statistics

N/A 6.3 15.8

2.7

ANCOVA Model F-statistics

N/A 36.1 38.2 6.3

Box test p-vaiue or Levene's test

< 0.001

< 0.001

< 0,001 0.283

Significant effect Interaction of form level and teaching mode Interaction of both form level and school level with teaching mode 1st admin listen 1st admin writing

Effect p-value

0.01 0.011 0.012

<0.001 0.004

Effect Eta squared

value 0.031 0.022 0.022 0.206 0.143 Results in the above table suggest that after controlling the effect of the 1st administration, in the 2nd administration, the interaction effort of form level and teaching mode was significant in the analysis of both the coordinated oral and listening scores and coordinated oral scores. In addition, the interaction effect of school level and teaching mode was significant in the analysis of the coordinated oral scores

P4 P6

Dependent variables

Oral Listening

Oral Listening

Teaching effect p-value

0.071 0.008

< 0.001 0.784

Effect Eta squared value

0.026 0.055 0.234 0.003

Levene's test

< 0.001 0.162 0.203 0.002

Significant comparison none

Local to Local > NET to NET Local to Local > NET to NET NET to Local >NET to NET

None

p-value 0.008

< 0.001 0.001

In the above table, the results suggest that after taking account of the pupils* performance in the 1st administration, in the 2nd administration, P6 and P4 pupils who were taught by local teachers for two years had better oral and listening scores than those taught by NETs for the same period. However, P6 pupils taught by aNET in the first year and a local teacher in the second had better oral scores than those taught by a NET for two years

(ii) 2nd cohort group

Coordinated Oral &

listening

Coordinated Oral Coordinated Listening Writing

ANOVA Model F-statistics

9.1 10.1

5.4

ANCOVA Model F-statistics

383 14.4

•S3,

Box test p-value or Levene's test

0.046 0.151 0.03 0.022

Significant effect

1* admin. Oral l*admHLOral 1* admin. Oral School level

Effect p-value

< 0.001

< 0.001

< 0.001 0.049

Effect Eta squared

value 0.446 0.39 0.153 0.123 Results in the above table suggest that after controlling the effect of the 1st administration, in the 2nd administration, no significant effect of teaching mode was observed.

Conclusion

The e ffect o f different t caching m odes w as s hown t o h ave i nfluenced se condary s tudents' performance in the 2nd administration of the 1st and 2nd cohorts, and primary pupils' performance in the 2nd administration of the 1st cohort, when account was taken of their performance in the 1st administration.

6.2.5 Alternative Analysis of Oral and Listening Scores

6.2.5.1 Paired Samples T-tests

*

Paired samples t-tests were performed on both the first and second cohort of secondary students* oral and listening assessment scores. The aim was to establish any possible gain between the pre and the post test scores and if so to establish any possible patterns in teaching modes over the given time period. The analysis was done by school level (high, medium and low), with the following results:

(A) Secondary

(i) First cohort group Listening Assessment Form 1 students High level schools F1-F2 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year!

NET Local NET Local

n

32 53 11 6

Pre-test

32.57 26.98 30.95 34.75

s.d (pre test) 6.849 6.533 6.547 4.782

Post-test

35.85 31.45

• 32.72 40.75

s.d (post

test) 8.013 8.107 7.281 7.789

Pre-post TestR2

0.565 0.529 0.945 0.880

t value (*ig.) -2.648 p=0.013*

-4.500 p=<0.001*

-2.448 p=0.034*

-3.464 p=0.018*

72

Medium level schools Fl - F2 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

n

20

67 10 45

Pre-test

22.50 22.72 18.75 21.93

s.d (pre test) 4,793 6.097 3.952 3.579

Post-test

2L22 24.69 22.75 20.82

s.d (post

test) 3.715 6.944 5.029 5.739

Pre-post TestR2

0.093 0.50 -0.459 0.263

t value

0.986 p=0.337 -2.458 p=0.017*

-1.644 p=0.134

1.261 p=0.214

* significant at p = < 0.05

Low level schools Fl - F2 Listening

teaching mode Yearl NET NET Local Local

teaching mode Yearl NET Local NET Local

n

34

23 11

12

Pre-test

18.57 14.86 15.68 15.87

s.d (pre test) 5.514 4.973 2.795 5.214

Post-test

21.75 16.43 13.90 14.00

s.d (post

test) 7.397 9.747 3.555 3.908

Pre-post TestR2 0.496 0.010 -0.257 0.187

t value («gO -2.770 p=0.009*

-0.689 p=0.498

1.163 p=0.272

1.101 p=0.295

* significant at p = < 0.05

In lower secondary schools (F1-F2), there was significant gain in the mean scores of the listening assessment hetween the pre and post tests in five cases. For the higji level schools, student taught by either a teaching transition of NET to NET, Local to NET or Local to Local had higher scores in the post test than in the pie test For the medium level schools, student taugjht by a teaching transition of NET to Local had M^o: scores in the post test than in the pre test. For the low level schools, student taught by a teaching transition of NET to NET had higjier scores in the post test than in the pre test

Form 3 and Form 4 students High level schools

F3 -F4, F4 - F5 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

n

43

Pre-test

27.95

s.d (pre test) 4.545

Post-test

31.00

&d (post test) 4.472

Pre-post TestR2

0.650

t value (sig.) -5.295 p=<0.001*

No relevant data 22

14

30.77 30.14

3.544 4.435

34.45 29.21

3.081 9.463

0346 0357

-4.533 p=«0.001*

0390 p=0.703

* significant at p = < 0.05

Medium level schools F3 - F4, F4 - F5 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year!

NET Local NET Local

n

24 28 35 2

Pre-test

20.41 21.53 20.22 15.50

s.d (pre test) 3.034 3.834 5.303 4.949

Post-test

23.37 20.50 21.71 23.50

s.d (post

test) 5.156 5.821 6.871 3.535

Pre-post TestR2

0.384 0.419 0.728 -1.00

t value (sigO -2.972

^=0,007*

L002 p=0.325 -L861 p=KX071 -1333 p=0.410

* significant at p = < 0.05 Low level schools

F3-F4, F4-F5 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

n

53

Pre-test

17.03

s*d (pre test) 3.546

Post-test

19.81

s.d (post

test) 3.858

Pre-post TestR2

0.535

t value . <«g*) -5.637 p=<0.001*

No relevant data No relevant data

11 17,54 5.768 17.72 6.117 0.733 -0.138 p=0.893

* significant at p = < 0.05

In upper secondary schools (F3-F4), there was significant gain in the mean scores of the Estening assessment in four cases. For high-level schools, student taught by a teaching transition of NET toNET or Local toNEThadtogfrerscoiramtte

For medium-level schools, student taught by a teaching transition of NET to NET had higher scores in the post test than in the pre test For low-level schools, student taught by a teaching transition of NET to NET had higher scores in the post test than in the pre test.

Eight out of the above nine groups involve the deployment of NET teachers (and the ninth was made up of only a very small sample size of 6). In fact, five of these groups involved students who were taught by NETs in two consecutive years. Whilst advising caution in the gmeralization^of paired sample t-tests (which are fixed point in time measurements), there is some evidence that there is a NET effect in helping to improve the listening skills of both upper and lower secondary students.

74

Oral Assessment Form 1 students High level schools Fl-F2Oral

teaching mode Year!

NET NET Local Local

teaching mode Year2 NET Local NET Local

n

77 64 17 20

Pre-test

19.75 20.15 2035 20,50

s.d (pre test) 5.171 4.677 4.636 6.202

Post-test

19.37 19.46 21.17 22.30

s.d (post

test) 5.125 4.918 4.720 6.300

Pre-post TestR2

0.419 0.596 0.754 0.886

t value

<*ig.) 0.606 p=0.546

L274 p=0.207 -1.034 p=0316 -2.698 p=0.014*

* significant at p = < 0.05

Medium level schools Fl-F2Oral

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET

Local

n

2 75 12 71

Pre-test .

14.50 17.18 18.50 16.50

s.d (pre test) 2.121 5.082 4.295 6.765

PosWest

18.00 1738 15.66 16.03

s4 (post

test) 2.828 4.434 5.069 5.041

Pre-post . TestR2

-1.000 0.571 0363 0.658

t value (sig.) -1.000 p=0.500 -0377 p=0.707

L844

^=0.092 0.775 p=O.441

Low level schools Fl-F2Oral

teaching mode Yearl NET NET Local Local

teaching mode Yearl NET Local NET

Local

n

42 26 12 14

Pre-test

12.83 12.84 15.75 8.28

s.d (pre test) 5.954 4.397 4.515 4.631

Post-test

11.42 14.51 12.41 15.71

s.d (post

test) 4.644 3.853 4.481 3.220

Pre-post TestR1

0.629 -0280 0.347 0.815

lvalue (sig.) 1.930 p=0.061

-1.291 p=0.208 2246 p=0.046*

-10.153 p=<001*

* significant at p = < 0.05

In lower secondary schools (F1-F2), there was significant gain in the mean scores of the oral assessment between the pre and post tests in three cases. For the higji level schools, student taught by a teaching tn^

pre test For the low level schools, student taught by either a teaching transition of Local to Local or Local to NET had higher scores in the post test than in the pre test.

Form 3 and Farm 4 students High level schools

F3-F4,F4-F5Oral

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

n

81 34 27 26

Pre-test

19.32 19.23 25.48 21.38

s<d (pre test) 4.826 5.199 4371 2.926

Post-test

21.29 19.85 25.51 23.11

s.d (post

test) 4.417 6.243 3.714 3.993

Pre-post TestR2

0.640 0.512 0.481 0.533

t value (sig.) -4.516 j»=<001*

-0.629

^=0.534 -0.046

^=0.963 -2.543 p=0.018*

c significant at p = < 0.05 Medium level schools F3—F4,F4-~F5Oral

teaching mode

* Yean NET NET Local Local

teaching mode Yearl NET Local NET Local

n

20 42 47 24

Pre-test

17.65 16.76 16.65 12.54

s.d (pre test) 5.050 4,673 5.506 5.200

Post-test

19.85 20.14 19.38 18.16

s.d (post

test) 4.826 5.089 5.198 5.700

Pre-post TestR2

0.246 0.328 0.503 0.742

t value (rig*) -1.622 p=0.121 -3.755 p-0.001*

-3.494 p=0.001*

-6.988 p=<0.001*

' significant at p = < 0.05 Low level schools

F3~F4fF4-F5Orat

teaching mode Yearl NET NET Local Local

teaching

mode Yearl NET Local NET Local

n

82

Pre-test

1432

s.d (pre test) 5.998

Post-test

15.84

sA (post

test) 5.186

Pre-post TestR1

0.556

t value (sig.) -2.585 p=0.012*

no relevant data no relevant data

22 11.18 3.775 15.09 4.770 0.710 -5.426

p=<0.001*

*significaniatp = <0.05

In upper secondary schools (F3-F4), there was significant gain in the mean scores of the oral assessment in seven cases. For the high level schools, student taught by either a teaching transition of Local to Local orNET to NET had higher scores in the post test than in the pre test.

For me medium level schools, student taught by either a teaching transition of Local to Local, NET to Local or Local to NET had higher scores in the post test than in the pre test For the low level schools, student taught by eimer a teaching transition of Local to Local or NET to NET had higher scores in the post test than in the pre test.

Although there was a significant gain in the mean scores of the oral assessments between the pre and the post tests, there does not seem to be any discemable pattern with regard to the deployment of NETs. Although NETs were involved in five of me above ten groups it is not

76

possible on the evidence of this analysis to attribute any NET effect to the gains. In lower secondary schools (F1-F2), on balance local teachers were involved in all three significant groups, whilst in upper secondary schools (F3-F4), both NETs and Local teachers were involved in an equal number of groups. This paired samples t-test indicates that there is no evidence of any NET effect in helping to improve the oral skills of either upper or lower secondary students.

(ii) 2nd Cohort Group Oral Assessment

Form 1 students High level schools Fl pre -Fl post Oral

teaching mode NET LOCAL BOTH

n = 61 30 37

mean orall

17.74' 17.20 18.24

mean ora!2 17.34 22.23 17.51

mean difference -0.40 +5.03 -0,73

correlation orall/ora!2 0.760*

0.360 0.628*;

paired samples p value p- 0.390 p = 0.051*

p = 0.296

* significant at p = < 0.05

Medium level schools Fl pre -Fl post Oral

teaching mode NET LOCAL BOTH

n*=

32 67 12

mean orall

17.06 17.06 12.08

mean 0ra/2 17.87 18.91 14.75

mean difference 40.81 +1.85 +2.67

correlation orall/oral2 0.908*

0.578*

0.003*

paired samples p value

^=0.062 p- 0.001*

p = 0.004*

* significant at p = < 0.05 Low level schools Fl pre -Fl post Oral

leaching mode NET LOCAL BOTH

n = mean

orall

mean ora!2

mean difference

correlation orall/oral2

paired samples p value Insufficient data

Insufficient data

83 12.32 13.49 +1.17 0.775* p = 0.014*

1 significant at p = < 0.05

In the nine categories of Fl students above, there was insufficient data for analysis in two cases:

NET, low level; LOCAL, low level. Of the remaining seven categories, the paired sample t-tests showed significant gain in oral assessment between orall and oral2 in four categories:

LOCAL, high level; LOCAL, medium level; BOTH medium level; BOTH low level. In the remaining three categories there was no significant difference in oral assessment between orall apd oral2. We can cautiously conclude ftom these pared sample t-tests that students taught by both local teachers and those taught by a combination of both local and NET show significant gain in their speaking scores, whilst those students taugfrt only by NET teachers do show gain but not at a significant level.

Form 3 Students High level schools

F3 pre test - F3 post test Oral

teaching mode

NET Local Both

n 20 40 60

Pre-test average 27-05 21.05 16.00

Post-test average

2530 20.22 19,82

Post-test average difference

-1.75 -0.83 +3.82

Pre-post Correlation value

0.548 0.810*

0.388*

Paired samples p-value p = 0.024*

p = 0.080 p< 0.001*

* significant at p = < 0.05

Medium level schools

F3pre test-F3post test Oral

teaching mode NET Local Both

n 33 57 52

Pre-test average

16.80 16.78 16.96

Post-test average

19.57 18.45 17.40

Post-test average difference

+2,77 +1.67 +0.44

Pre-post Correlation value

0.724*

0.606*

0.472*

Paired samples p-value p< 0.001*

p = 0.003*

p< 0.001*

* significant at p = < 0.05

Low level schools

F3 pre test - F3 post test Oral

teaching mode

NET Local Both

n 50 28

Pre-test average 1622 1721

Post-test average

16.90 20.46

Post-test average difference

+0.68 +3.25

Pre-post Correlation value

0.848*

0.395

Paired samples p-value p = 0.161 p ~ 0.005*

Insufficient Data

* significant at p = < 0.05

In the nine categories of F3 students above, there was insufficient data for analysis in two cases:

NET, low level; LOCAL, low level. Of the remaining seven categories, the paired sample t-tests showed significant gain in oral assessment between orall and oral2 in four categories:

LOCAL, higji level; LOCAL, medium level; BOTH medium level; BOTH low level In the remaining three categories there was no significant difference in oral assessment between orall and ora/2. We can cautiously conclude from these paired sample t-tests that Fl students taught by both local teachers and those taught by a combination of both local and NET show significant gain in their speaking scores, whilst those students taught only by NET teachers do show gain but not at a significant level.

(B) Primary Schools (i) First Cohort Groups

Paired sample t-tests were also conducted for both the first and second cohort of pupils, according to teaching mode. For the first cohort, there were two such modes (Le. the type of teaching (NET, or Local) experienced in year 1 and year 2) but due to ^

not possible to conduct the analysis on all possible group combinations. Results are shown in the tables below:

78

Listening Assessment P3 Listening

teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

number

100

Pre-test

1337

Post-test

15.12

Pre-post TestR2

0.528

Pre-post test difference

1.75

t value (sig.) -5.267 p=<0.001*

data not available data not available

40 14.65 15.70 0363 L05 -2371

p=0.023*

P5 Listening teaching mode Yearl NET NET Local

teaching mode Year2 NET Local Local

number

69 13 63

Pre-test

1639 18.84 14.74

Post-test

1736 1830 17.00

Pre-post testR2

0.596 0.815 0.582

Pre-post test difference 0.97 -0.53 2..2S

t value (sig«) -2.452 p=0.017*

-1.203 p=0.252 -6.096 p=<0,001*

The results in the above table show that in listening, on average four of the five groups that were analysed showed significant positive gain, as follows, but it is not possible to ascribe gain in listening to any particular teaching mode. The gain however does lend support to the reliability of the instruments, since significant gain over time would be predicted.

Oral Assessment

P3toP4Oral

teaching mode Yearl NET NET Local Local

teaching mode

Yearl NET Local NET Local

number

133

Pre-test

1528

Post-test

15.51

Pre-post TestR1

0.616

Pre-post test difference 0.23

lvalue (sig.) -0.593 p=0.555 data not available

data not available

75 12.89 17.13 0.687 4.27 -11399

p=<0.001* _

P5toP6Oral

Teaching mode Yearl NET NET Local Local

teaching mode Year2 NET Local NET Local

number

94 16

Pre-test

1537 1432

Post-test-16.14 1631

Pre-post TestR2

0.693 0/744

Pre-post test difference 0.79 1.99

t value (sig-) -2289 p=0.024*

-2,271 p-0.038*

data not available — —

78 14.95 1836 0.597 3.41 .7.831

pO.OOl*

The above results show that in speaking, on average, four of the five groups that were analysed showed significant positive gain. From these results, again it is not possible to ascribe gain in

speaking scores as measured by our instruments to any particular teaching mode. As in the listening scores analysis the results again support the reliability of the instruments used, since significant gain over time would be predicated.

(ii) Second Cohort Group Listening Assessment High Level Schools P3 Listening

teaching mode Both

number 12

Pre-test 17.83

Post-test 17.83

Pre-post test R2.

0.785

Pre-post test difference 0.000

lvalue (sig.) 0.000 P=l

P5 Listening

Teaching mode NET Both

number 22 8

Pre-test 19.32 19.37

Post-test 18.27 .18.62

Pre-post test R2

0.682 0.684

Pre-post test difference -1.05 -0.75

t value (sig-) 2.383 p=0.027*

L53 p=0.170

Medium Level Schools P3 Listening

teaching mode Both

number 74

Pre-test 13.53

Post-test 15.48

Pre-post test R2

0.285

Pre-post test difference 1.95

t value (sig.) -4.102

p-<0.001*

P5 Listening

teaching mode Local Both

number 21 46

Pre-test 14.43 16.69

Post-test 15.95 1730

Pre-post test R2

0.071 0.49

Pre-post test difference

1.52 0.61

t value (sig.) -1.657

p=0.113 -1.287

p=0.205

The above results show that in listening, P 3 pupils taught by NET in higji-level school showed significant negative fall, P 3 pupils taught by a combination of NET and Local teacher in medium-level school showed any significant positive gain- However, from these results it is not possible to ascribe gain in listening scores as measured by our instruments to any particular teaching mode.

80

在文檔中 from This (頁 70-87)