e 4.28, the mean of NLP was proved to be 143.5504 while Std Deviation turned out to be 11.64847. In addition, skewness and kurtosis regarding Neuro-Linguistic Programming were 1.047 and 7.920 respectively
Table 4.29
NLP Descriptives for Different Levels of Expert Teaching Style
Expert
Statistic
Std. Error
NLP
Low
Mean
132.7857
3.40520
95% Confidence Interval for Mean
Lower Bound
125.4292
Upper Bound
140.1422
5% Trimmed Mean
132.9286
Median
132.0000
Variance
162.335
Std. Deviation
12.74108
Minimum
110.00
Maximum
153.00
Range
43.00
Interquartile Range
14.00
Skewness
-.224
.597
Kurtosis
.317
1.154
Moderate
Mean
144.7345
1.02644
95% Confidence Interval for Mean
Lower Bound
142.7008
Upper Bound
146.7683
5% Trimmed Mean
144.4567
Median
143.0000
Variance
119.054
Std. Deviation
10.91118
Minimum
120.00
Maximum
210.00
Range
90.00
Interquartile Range
12.00
Skewness
1.753
.227
Kurtosis
10.632
.451
a. NLP is constant when Expert = High. It has been omitted.
Table 4.30
NLP Descriptives for Different Levels of Formal Authority Teaching Style
Formal authority
Statistic
Std. Error
NLP
Low
Mean
141.0313
1.00421
95% Confidence Interval for Mean
Lower Bound
139.0376
Upper Bound
143.0249
5% Trimmed Mean
141.3727
Median
142.0000
Variance
96.810
Std. Deviation
9.83918
Minimum
110.00
Maximum
163.00
Range
53.00
Interquartile Range
12.00
Skewness
-.582
.246
Kurtosis
1.317
.488
Moderate
Mean
150.8788
2.34147
95% Confidence Interval for Mean
Lower Bound
146.1094
Upper Bound
155.6482
5% Trimmed Mean
149.5051
Median
152.0000
Variance
180.922
Std. Deviation
13.45074
Minimum
135.00
Maximum
210.00
Range
75.00
Interquartile Range
14.00
Skewness
2.608
.409
Kurtosis
11.211
.798
Table 4.31
NLP Descriptives for Different Levels of Personal Model Teaching Style
Personal model
Statistic
Std. Error
NLP
Low
Mean
142.1635
.96592
95% Confidence Interval for Mean
Lower Bound
140.2478
Upper Bound
144.0791
5% Trimmed Mean
142.6068
Median
143.0000
Variance
97.031
Std. Deviation
9.85045
Minimum
110.00
Maximum
163.00
Range
53.00
Interquartile Range
12.50
Skewness
-.739
.237
Kurtosis
1.368
.469
Moderate
Mean
149.3200
3.25101
95% Confidence Interval for Mean
Lower Bound
142.6102
Upper Bound
156.0298
5% Trimmed Mean
147.4333
Median
148.0000
Variance
264.227
Std. Deviation
16.25505
Minimum
132.00
Maximum
210.00
Range
78.00
Interquartile Range
19.50
Skewness
2.184
.464
Kurtosis
7.358
.902
Table 4.32
NLP Descriptives for Different Levels of Facilitator Teaching Style
Facilitator
Statistic
Std. Error
NLP
Low
Mean
139.3750
1.69548
95% Confidence Interval for Mean
Lower Bound
135.9641
Upper Bound
142.7859
5% Trimmed Mean
139.6019
Median
142.0000
Variance
137.984
Std. Deviation
11.74666
Minimum
110.00
Maximum
163.00
Range
53.00
Interquartile Range
13.75
Skewness
-.365
.343
Kurtosis
.652
.674
Moderate
Mean
146.0247
1.21398
95% Confidence Interval for Mean
Lower Bound
143.6088
Upper Bound
148.4406
5% Trimmed Mean
145.4108
Median
145.0000
Variance
119.374
Std. Deviation
10.92586
Minimum
125.00
Maximum
210.00
Range
85.00
Interquartile Range
14.00
Skewness
2.434
.267
Kurtosis
13.657
.529
Table 4.33
NLP Descriptives for Different Levels of Delegator Teaching Style
Delegator
Statistic
Std. Error
NLP
Low
Mean
130.0000
2.99537
95% Confidence Interval for Mean
Lower Bound
123.0927
Upper Bound
136.9073
5% Trimmed Mean
129.8333
Median
130.0000
Variance
80.750
Std. Deviation
8.98610
Minimum
120.00
Maximum
143.00
Range
23.00
Interquartile Range
17.50
Skewness
.312
.717
Kurtosis
-.991
1.400
Moderate
Mean
144.5667
1.02304
95% Confidence Interval for Mean
Lower Bound
142.5410
Upper Bound
146.5924
5% Trimmed Mean
144.3981
Median
143.5000
Variance
125.592
Std. Deviation
11.20679
Minimum
110.00
Maximum
210.00
Range
100.00
Interquartile Range
14.00
Skewness
1.275
.221
Kurtosis
9.667
.438
4.2.3.3. Tests of Normality
Since the teaching styles are categorized into low, moderate, and high levels, each teaching style is considered as a nominal variable. Moreover, as the NLP is also on an interval scale, the choice of statistic to measure the relationship between one nominal variable and one interval variable is eta. However, since the fr
eq
uencies of some of the styles levels are very low, it was decided to choose non-parametric Kruskal Wallis and Mann Whitney tests to compare the levels of each style in terms of NLP scores. The reason for choosing non-parametric tests was that the test of normality results in Tables 4.34 to 4.38 indicated non-normality of the data (p .05).
Table 4.34
Tests of Normality Regarding Expert Style
Expert
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
NLP
Low
.240
14
.028
.890
14
.080
Moderate
.101
113
.007
.872
113
.000
a. Lilliefors Significance Correction
b. NLP is constant when Expert = High. It has been omitted.
Table 4.35
Tests of Normality Regarding Formal Authority Style
Formal authority
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
NLP
Low
.087
96
.068
.964
96
.009
Moderate
.203
33
.001
.752
33
.000
a. Lilliefors Significance Correction
Table 4.36
Tests of Normality Regarding Personal Model Style
Personal model
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
NLP
Low
.090
104
.039
.958
104
.002
Moderate
.160
25
.098
.787
25
.000
a. Lilliefors Significance Correction
Table 4.37
Tests of Normality Regarding Facilitator Style
Facilitator
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
NLP
Low
.109
48
.200*
.960
48
.105
Moderate
.126
81
.003
.823
81
.000
*. This is a lower bound of the true significance.
a. Lilliefors Significance Correction
Table 4.38
Tests of Normality Regarding Delegator Style
Delegator
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
NLP
Low
.200
9
.200*
.860
9
.095
Moderate
.109
120
.001
.882
120
.000
*. This is a lower bound of the true significance.
a. Lilliefors Significance Correction
4.2.3.4. Final Results
Tables 4.39 to 4.43 present the results on the comparison of total NLP scores across the categories of teaching styles. Evidently, the categories of all teaching styles except the Personal Model in terms of NLP are significantly different from one another. In other words, except for the Personal Model there is a significant relationship between different teachers’ styles and NLP. A closer look at the descriptive statistics of these teaching styles reveals that the moderate category of the above teaching styles are of higher NLP in comparison to their low categories. This indicates that there is a positive relationship between teachers’ Expert, Formal Authority, Facilitator, and Delegator styles and NLP.
Table 4.39
Comparing NLP across Categories of Expert
Table 4.40
Comparing NLP across Categories of Formal Authority
Table 4.41
Comparing NLP across Categories of Personal Model
Table 4.42
Comparing NLP across Categories of Facilitator
Table 4.43
Comparing NLP across Categories of Delegator
4.2.4. Testing the Third Null Hypothesis
H03: There is no significant relationship between teachers’ autonomy and NLP (Neuro-linguistic programming).
4.2.4.1. Assumption of Linearity
In order to test above null hypothesis, correlational measures needed to be employed. Pearson product moment correlation and Spearman rho were two options; however, to choose between these two measures, some assumptions needed to be checked in advance. The first of these the linearity of the relationship between NLP and autonomy, which was done by drawing the scatter graph (Figures 4.1 to 4.3). As the figures display, it seems that the two variables’ data are approximately aligned along a straight line; however, the lines are not very diagonal. Therefore, despite the linearity of the relationship, a low correlation coefficient is expected between the two variables.
Figure 4.1. General Autonomy Scatter Plot
Figure 4.2. Curriculum Autonomy Scatter Plot
Figure 4.3. Total Autonomy Scatter Plot
4.2.4.2. Assumption of Normality
The next assumption is to do with the normality of the data, which was investigated employing Kolmogorov-Smirnov and Shapiro-Wilk tests of normality, whose results in Table 4.44 indicate that the data were not normally distributed (p .05).
Table 4.44
Tests of Normality
Tests of Normality
Kolmogorov-Smirnova
Shapiro-Wilk
Statistic
df
Sig.
Statistic
df
Sig.
General Autonomy
.115
129
.000
.971
129
.007
Curriculum Autonomy
.123
129
.000
.961
129
.001
Total Autonomy
.100
129
.003
.979
129
.042
NLP
.100
129
.003
.905
129
.000
a. Lilliefors Significance Correction
4.2.4.3. Final Results
With regard to the fact that data were not normally distributed, the choice of statistic became Spearman rho, whose results in Table 4.45 show that there is almost no significant relationship between the two variables. In fact, NLP is only significantly and positively correlated with General autonomy with small to medium effect size (p 05). In other words, the null hypothesis is mainly supported; that is to say, except for General autonomy, there is no significant relationship between teachers’ Total and Curriculum autonomy and NLP (Neuro-Linguistic Programming).
Table 4.45
Correlations among Curriculum, General and Total Autonomy and NLP
NLP
General Autonomy
Curriculum Autonomy
Total Autonomy
Spearman’s rho
NLP
Correlation Coefficient
1.000
.205*
-.028
.103
Sig. (2-tailed)
.
.020
.757
.246
N
129
129
129
129
General Autonomy
Correlation Coefficient
.205*
1.000
.245**
.807**
Sig. (2-tailed)
.020
.
.005
.000
N
129
129
129
129
Curriculum Autonomy
Correlation Coefficient
-.028
.245**
1.000
.728**
Sig. (2-tailed)
.757
.005
.
.000
N
129
129
129
129
Total Autonomy
Correlation Coefficient
.103
.807**
.728**
1.000
Sig. (2-tailed)
.246
.000
.000
.
N
129
129
129
129
*. Correlation is significant at the 0.05 level (2-tailed).
**. Correlation is significant at the 0.01 level (2-tailed).
4.2.4. Testing the Fourth Null Hypothesis
H04: There is no significant difference between EFL teachers’ teaching styles and NLP in predicting autonomy?
In order to test this hypothesis, multiple regression analysis was employed three times for the three Total, General, and Curriculum autonomy scores. The variables whose predictive powers are supposed to be examined are the 5 teaching styles and NLP. Employing multiple regression requires checking several assumptions which are initially checked in the following.
4.2.4.1. Assumption of Multicollinearity
The first assumpti
on is