4 Results and Discussion
4.2 Teachers’ perspectives through interviews and the questionnaire
4.2.8 Factor analysis
Type 2 Receptive ways teachers learn
23 Methodology conferences never give English teachers any useful ideas.
24 I find professional journals useless
30 I cannot use the results of the latest research in my teaching.
26 Teacher’s books (of coursebooks) never provide good ideas.
25 I can use a lot of ideas from resource books (e.g. Recipe Book for Tired Teachers, 165 ideas).
28 I have learnt a lot from my students.
Type 2 statements seem to be more about outside sources of learning, ways that aim to give or transmit knowledge to teachers ready-made. This group of statements therefore can be assumed to describe transmissive, receptive or passive ways of learning.
The three statements below did not easily fit into either of the clusters, although they do have something in common, namely that all three statements refer to colleagues.
15 I have learnt a lot from my colleagues.
18 I have learnt a lot from my own previous teachers.
19 My mentor teacher/teacher trainer at school had no influence on my development.
percentages of the variance are shown below in Table 65. The four extracted factors accounted for 57.8 % of the total variance.
Table 65 Factor analysis results for knowledge domains
Component Matrix(a) Component
1 2 3 4
Command of E ,550 -,110 ,573 -,117
Comm of MT+ ,049 ,705 ,130 ,278
How to ,566 -,352 ,449 -,030
Edu System+ ,191 ,732 ,078 -,164
Communication ,505 ,026 -,386 -,306
Explaining ,465 -,121 -,152 ,708
Motivating ,648 -,163 -,128 -,408
Work with students ,542 -,309 -,480 ,217
Work with colleauges+ ,550 ,401 -,205 ,091
About the L ,347 ,005 ,477 ,229
Know the ss+ ,500 ,361 -,069 -,116
Extraction Method: Principal Component Analysis.
a 4 components extracted.
Communalities
Extraction
Command of E ,658
Comm of MT+ ,593
How to ,646
Edu System+ ,605
Communication ,498
Explaining ,755
Motivating ,629
Work with students ,668 Work with colleauges+ ,513
About the L ,401
Know the ss+ ,398
Extraction Method: Principal Component Analysis.
Rotated Component Matrix(a) Component
1 2 3 4
Command of E ,172 ,788 ,069 -,054
Comm of MT+ -,251 -,016 ,725 ,071
How to ,213 ,751 -,157 ,113
Edu System+ ,118 ,007 ,738 -,216
Communication ,691 -,039 ,101 ,095
Explaining -,010 ,180 ,075 ,847
Motivating ,739 ,286 -,034 ,012
Work with students ,493 -,010 -,159 ,632
Work with colleauges+ ,376 ,058 ,527 ,303
Know the ss+ ,399 ,143 ,462 ,071 Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a Rotation converged in 5 iterations.
Total Variance Explained
Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2,519 22,899 22,899 1,735 15,768 15,768
2 1,598 14,523 37,422 1,660 15,090 30,858
3 1,247 11,337 48,759 1,656 15,052 45,910
4 1,001 9,101 57,859 1,314 11,950 57,859
Extraction Method: Principal Component Analysis.
Factor 1, which accounts for approximately 23 % of the variance, received a loading larger than .7 from one statement, the one on the ability to motivate students. The second highest factor loading was that of the statement on the ability to communicate (.69). Other statements that loaded onto this dimension (Factor loading > .3) were the knowledge of the students and working with students and colleagues in order of higher loading to lower loading. This factor could be labelled Motivational and interpersonal skills. Factor 2, which accounts for about 15 % of the variance, received loadings higher than .7 from two statements: the knowledge of English and how to teach different aspects of the language. An acceptably high loading (Factor loading > .3) was identified in the case of knowledge about the language. This factor could therefore be labelled as Traditional subject matter knowledge.
Factor 3, which accounts for 11 % of the variance, received high loadings (Factor loading >
.7) from two statements: the knowledge of the mother tongue and knowledge of the educational system. Further statements that had acceptable loadings included knowing the students and working with colleagues. This factor could be labelled as Local contextual knowledge. Factor 4 explains only 9 % of the variance and received a loading higher than .7
from one statement: the ability to explain well. Statements that had acceptable loadings were working with students and working with colleagues. This factor could therefore be labelled as Collaborative skills.
In sum, four underlying dimensions appear to surface from the factor analysis:
motivational and interpersonal skills (Statements 8, 11, 12), traditional subject matter knowledge (1, 3, 4), local contextual knowledge (2, 7) and collaborative skills (10, 13, 14), but these labels might reflect just one interpretation of the results of the principal component analysis. The factor called local knowledge is a compound of only two statements, therefore, out of the four, this factor is the least substantiated one.
In addition, the aim of the principal component analysis was to extract fewer components from the data, but the two extracted factors accounted for only 37 % of the variance. In this model, command of the mother tongue and knowledge of the education system loaded onto one factor with outstanding values. Added to them were working with colleagues and knowing the students, with loadings above .5. All the other statements loaded onto the first factor. Since this emerging model did not explain as much of the variance as the previous one, and the first factor was not easily circumscribable, this model was dropped in favour of the four factor model.
As regards the second section of the opinionaire, after an initial run, when four factors had surfaced, no statements were to be eliminated from the investigation since no extraction values were under .3. This first extraction resulted in the emergence of four factors after the rotation converged in seven iterations. The component matrix, communalities, the rotated component matrix and the percentages of the variance are shown below in Table 66. The four extracted factors accounted for 51.8 % of the total variance.
Table 66 Factor analysis results for sources of learning
Component Matrix(a)
Component
1 2 3 4
Learn from colleagues ,432 -,050 -,556 ,077
Feedback ,632 -,357 ,093 ,081
Visits+ ,492 ,140 -,394 ,071
Ghosts ,359 ,095 -,167 ,612
Mentor+ ,405 ,145 -,317 ,393
S Relationship ,582 -,454 ,118 -,025
Reflection 1 ,502 -,275 ,354 ,183
Reflection 2 ,361 -,003 ,664 ,037
Meth Conf+ ,410 ,628 ,241 -,082
Prof journals+ ,256 ,742 ,131 -,018
Resource books ,467 ,333 -,187 -,329
Teacher's books+ ,139 ,484 ,275 ,218
Learn vocab from S+ ,501 -,128 -,053 -,440
New knowledge through
S ,630 -,258 -,147 -,331
Own teaching ,566 -,325 ,222 ,123
Research+ ,423 ,543 -,030 -,158
Extraction Method: Principal Component Analysis.
a 4 components extracted.
Communalities
Extraction
Learn from colleagues ,505
Feedback ,543
Visits+ ,421
Ghosts ,539
Mentor+ ,439
S Relationship ,559
Reflection 1 ,486
Reflection 2 ,573
Meth Conf+ ,628
Prof journals+ ,633
Resource books ,473
Teacher's books+ ,376
Learn vocab from S+ ,464 New knowledge through
S ,594
Own teaching ,490
Research+ ,500
Extraction Method: Principal Component Analysis.
Total Variance Explained
Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3,459 21,618 21,618 2,448 15,300 15,300
2 2,234 13,964 35,582 2,286 14,285 29,585
3 1,442 9,010 44,592 1,785 11,156 40,741
4 1,088 6,799 51,392 1,704 10,650 51,392
Extraction Method: Principal Component Analysis.
Rotated Component Matrix(a)
Component
1 2 3 4
Learn from colleagues ,007 -,088 ,381 ,593
Feedback ,647 -,065 ,260 ,230
Visits+ ,047 ,155 ,340 ,529
Ghosts ,223 ,132 -,224 ,650
Mentor+ ,099 ,147 ,020 ,638
S Relationship ,651 -,156 ,312 ,117
Reflection 1 ,691 ,048 ,009 ,074
Reflection 2 ,617 ,347 -,093 -,251
Meth Conf+ ,090 ,775 ,133 ,036
Prof journals+ -,112 ,782 ,038 ,091
Resource books -,027 ,397 ,538 ,158
Teacher's books+ ,064 ,558 -,237 ,065
Learn vocab from S+ ,262 ,054 ,625 -,029
New knowledge through
S ,377 -,050 ,655 ,142
Own teaching ,670 -,017 ,142 ,146
Research+ -,029 ,614 ,308 ,163
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a Rotation converged in 7 iterations.
Factor 1, which accounts for approximately 22 % of the variance, did not receive a loading larger than .7 from any single statement, but several statements loaded onto this dimension with a factor loading higher than .6: the statement on learning from one’s own teaching, the two statements on reflection, the statement on learning from feedback from the students and the statement on benefiting from the relationship with the students. Since all of these include a re examination of one’s teaching in the light of reality, this factor could be labelled as Reflective sources of learning. The only statement that perhaps stands out from the rest is the statement on benefiting from the relationship with the students, but it can be assumed that the relationship with the students entails interaction and an exchange of opinions and ideas with the students, which necessitates a level of reflection. Factor 2 accounts for 14 % of variance. Four statements loaded onto this dimension: the statements on learning from methodology conferences, the one on learning from professional journals, the one on learning from teacher’s books and the one on learning from research findings. As these statements describe the sources of learning that lie outside schools, this factor could be labelled as External sources of learning. Factor 3 explains 9 % of the variance and three statements loaded onto it: the statements on learning from resource books, learning new
vocabulary from students and acquiring new knowledge from students. Since the loading of the statement on resource books is not as high as that of the other two statements and the statement also loads onto factor 2, although with a lower factor loading, I decided to group the statement on resource books together with other external sources of learning in Factor 2.
Hence, the two student-related learning sources that are left constitute Factor 3, which could be labelled as Student-generated learning sources. Factor 4 accounts for approximately 7 % of the variance and comprises statements that are all related to the teacher’s relationship with other colleagues. Four statements load onto this dimension: the statement on learning from colleagues, the one on learning from visiting colleagues’ classes, the one on learning from one’s own previous teachers and the one on learning from one’s mentor teacher. Therefore, this factor could be labelled as Collegial sources of learning.
To recapitulate, four underlying dimensions appear to surface from the factor analysis of the second section of the questionnaire: reflective sources of learning (Statements 16, 20, 21, 22, 29), external sources of learning (23, 24, 25, 26), student-generated learning (27, 28) and collegial learning (15, 17, 18, 19), but the labels given might reflect just one interpretation of the results of the principal component analysis. Student-generated learning is a factor that only relies on responses given to two statements, so that appears to be the weakest factor.
The reliability of the eight scales was examined and the Cronbach alpha value was above the critical .6 level (Cohen et al., p. 506). Accordingly, the scales can be considered to be acceptably reliable (see Table 67).
Table 67 Reliability statistics concerning the scales
Cronbach's Alpha
Cronbach's Alpha Based
on Standardised
Items N of Items
.631 .666 8
Table 68 Descriptive statistics concerning the scales based on factor analysis
Locknow Intpers Tradsub Collab Reflect External Studgen Collegial
N Valid 251 251 251 251 251 251 251 251
Missing 0 0 0 0 0 0 0 0
Mean 3.2888 4.4681 4.1866 4.3851 4.2287 3.7428 3.7410 3.6790
Mode 3.50 4.67 4.00(a) 4.67 4.20 3.60(a) 3.50 3.75
a Multiple modes exist. The smallest value is shown
In Table 68 the factors about local contextual knowledge and student generated learning will have to be neglected since they are based on too few statements. However, an examination of the difference between mean scores for the different factors shows that the majority of the means differ significantly. More specifically, four out of the six comparisons proved significant differences.
Table 69 Paired samples t-test results concerning the mean scores of the scales
Tradsub Collab External Collegial
Intpers
t(251) = 7.569, p < .001 Not significant Tradsub
t(251) = -5.447, p < .001 Collab
Reflect
t(251) = 11,08, p < .001 t(251) = 13.320, p < .001 External
Not significant Collegial
Thus, from Table 68 and Table 69 one can conclude that the knowledge domains considered important by the teachers in the sample can be listed in the following order of importance:
1. Motivational and interpersonal skills and Collaborative skills, 2. Traditional subject matter knowledge.
As can be seen in Table 69, the difference between the importance of Motivational and interpersonal skills and Collaborative skills is not significant; therefore, their importance can be considered to be equal.
Similarly, from Table 68 and Table 69 one can conclude that the sources of learning considered important by the teachers in the sample can be listed in the following order of importance:
1. Reflective sources of learning,
2. Collegial sources and External sources of learning.
The difference between the importance of Collegial sources of learning and External sources of learning is not significant, thus, the extent to which they describe the sample can be considered to be equal (See Table 69).
The four knowledge domains and the four sources of learning correlate significantly (See Table 70). This is not surprising as the analysis aims at identifying components of the knowledge and skills of English teachers and the sources of their learning, which are interrelated, integrated and intertwined (Richards, 1998; Szesztay, 2004; Verloop et al., 2001, cited in Borg, 2003; Woods, 1996). It would indeed be a surprise if such integrated components did not correlate significantly.
Table 70 Intercorrelations between knowledge domains and sources of learning
Locknow Intpers Tradsub Collab Reflect External Studgen Collegial
Locknow PCorrelation 1 .133(*) -.021 .050 -.043 .112 -.018 .091
Sig. (2-tailed) .035 .742 .427 .500 .078 .778 .151
N 251 251 251 251 251 251 251 251
Intpers PCorrelation .133(*) 1 .257(**) .421(**) .434(**) .248(**) .348(**) .211(**)
Sig. (2-tailed) .035 .000 .000 .000 .000 .000 .001
N 251 251 251 251 251 251 251 251
Tradsub PCorrelation -.021 .257(**) 1 .253(**) .212(**) .044 .122 .197(**)
Sig. (2-tailed) .742 .000 .000 .001 .490 .053 .002
N 251 251 251 251 251 251 251 251
Collab PCorrelation .050 .421(**) .253(**) 1 .399(**) .164(**) .146(*) .263(**)
Sig. (2-tailed) .427 .000 .000 .000 .009 .020 .000
N 251 251 251 251 251 251 251 251
Reflect P Correlation -.043 .434(**) .212(**) .399(**) 1 .120 .427(**) .270(**)
Sig. (2-tailed) .500 .000 .001 .000 .057 .000 .000
N 251 251 251 251 251 251 251 251
External PCorrelation .112 .248(**) .044 .164(**) .120 1 .164(**) .291(**)
Sig. (2-tailed) .078 .000 .490 .009 .057 .009 .000
N 251 251 251 251 251 251 251 251
Studgen PCorrelation -.018 .348(**) .122 .146(*) .427(**) .164(**) 1 .278(**)
Sig. (2-tailed) .778 .000 .053 .020 .000 .009 .000
N 251 251 251 251 251 251 251 251 Collegial PCorrelation .091 .211(**) .197(**) .263(**) .270(**) .291(**) .278(**) 1
Sig. (2-tailed) .151 .001 .002 .000 .000 .000 .000
N 251 251 251 251 251 251 251 251
* Correlation is significant at the 0.05 level (2-tailed).
** Correlation is significant at the 0.01 level (2-tailed).
The statistical analysis of the responses to the questionnaire showed that the tool was a reliable one. Even though the section on knowledge domains included statements that were all considered to be very important and even the lowest ranked knowledge domain, the knowledge of the mother tongue, received a ranking of 3.03, the responses also revealed that the most important components of an English teacher’s knowledge were considered to be the command of English, the ability to transmit knowledge, communications skills, the ability to motivate and the ability to explain clearly. Similarly, the questionnaire included statements on professional learning sources that were all considered to be characteristic of English teachers with even the lowest scoring source, learning from colleagues, being 3.38 out of a total of 5.
The most frequently used sources of learning were the following: reflection, teachers’ own experience, visiting lessons and teachers’ relationships with their students. Minor differences were detected as regards different age groups, school location and gender.
An examination of correlations proved that learning from colleagues is associated with visiting their lessons, since those who rated either learning from colleagues or visiting colleague’s classes high, also rated the other similarly high. Learning from feedback is also ranked higher by those who ranked learning from students. Knowing the students and learning from student feedback were also given correlating rankings, just as well as learning from methodology conferences, learning from research reports and learning from professional journals. The rankings given to learning new vocabulary from students correlated with those given to acquiring new knowledge through the students.
The cluster analysis grouped the knowledge domain related statements into three distinct categories: professional skills, traditional skills and contextual knowledge. The
hierarchical cluster analysis run on the sources of learning categorised statements into two different sources of learning: interactive and receptive ways of learning.
The factor analysis yielded a four-factor model in terms of teacher knowledge domains: motivational and interpersonal skills, traditional subject matter knowledge, local contextual knowledge and collaborative skills appeared to surface from the data. A similar factor analysis run on the sources of knowledge resulted in a four-factor model, too: reflective sources of learning, external sources of learning, student-generated sources of learning and collegial sources of learning were differentiated.