دانلود پایان نامه ارشد درمورد learning، not، Regression

دانلود پایان نامه ارشد

le. As displayed in Table 4.10, the visual style entered the regression model as the best predictor of learning strategy (R = .823, R2 = .675). That is to say the visual learning style can predict 67.8 percent of learning strategy.
The group style entered the regression model as the second best predictor of learning strategy (R = .845, R2 = .714). That is to say the visual and group learning styles can predict 71.4 percent of learning strategy.
Table 4.10: Model Summary, Regression Analysis Predicting Learning Strategy by Using Subcomponents of Learning Style Preferences
Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.823a
.678
.675
12.531
2
.845b
.714
.710
11.852
a. Predictors (Constant): Visual
b. Predictors (Constant): Visual, Group
c. Dependent Variable: Learning Strategy

The results of the ANOVA tests of significance of the regression model (Table 4.11) indicated that the regression model including visual style (f = 306.75, P .05) and both visual and group styles (F = 180.55, P .05) enjoy statistical significance. Thus, the sixth null hypothesis as EFL learners’ learning style preferences do not predict their use of language learning strategies is rejected. The visual and group styles are the only significant predictors of learning strategy.

Table 4.11: ANOVA Test of Significance of Regression Model Predicting Learning Strategy by Using the Subcomponents of Learning Style Preferences
Model
Sum of Squares
df
Mean Square
F
Sig.
1
Regression
48166.367
1
48166.367
306.758
.000b

Residual
22924.572
146
157.018

Total
71090.939
147

2
Regression
50723.564
2
25361.782
180.556
.000c

Residual
20367.375
145
140.465

Total
71090.939
147

a. Dependent Variable: Learning Strategy
b. Predictors: (Constant), Visual
c. Predictors: (Constant), Visual, Group

Table 4.12 displays the learning styles that did not enter the regression model due to their non-significant contribution to learning strategies (P .05).
Table 4.12: Excluded Variables of Learning Style Preferences
Model
Beta In
t
Sig.
Partial Correlation
Collinearity Statistics

Tolerance

Individual
.147c
1.789
.076
.147
.287

Auditory
-.085c
-.765
.446
-.064
.159

Tactile
-.044c
-.490
.625
-.041
.243

Kinesthetic
.162c
1.848
.067
.152
.254
a. Dependent Variable: Learning Strategy
b. Predictors in the Model: (Constant), Visual, Group

Unfortunately the regression model has not satisfied the assumptions of linearity although the assumption of homoscedasticity is met. As displayed in Figure 4.9, the spread of dots do not form a rectangular. Rather, the dots have formed a horizontal line. These results suggest that the assumption of linearity is violated. Fortunately, the dots do not show a funnel shape (narrow at one and wide at the other end). Thus, it can be concluded that the assumption of homoscedasticity is met. It should be mentioned that in case these assumptions are violated it does not create a problem. As suggested by Field (2009) if the assumptions of regression analysis are not met the results can only be generalized to the population from which the sample is selected.

Figure 4.9: Scatter Plot of Testing Assumptions of Linearity and Homoscedasticity

4.3.7 Testing the Seventh Null Hypothesis

H 07: EFL learners’ use of language learning strategies does not predict their creativity.
A multiple regression analysis was run to predict EFL learners’ creativity by using the six components of the language learning strategies. As displayed in Table 4.13, the regression model converged in four steps. The social strategy was entered into the model on the first step predicting 79.9 percent of scores on creativity (R = .894, R2 = .799). The cognitive strategy increased the predictive power to 82.1 percent (R = .906, R2 = .821). The affective strategy added up the percentage of prediction to 82.6 percent (R = .909, R2 = .826). Finally, the metacognitive strategy entered the model on the last step which increased the regression model’s prediction to 93.2 percent (R = .912, R2 = .832).
Social strategy was the best predictor of creativity R (r = 0.89, r2 = 79). That is to say social strategy predicts 79 percent of creativity, while cognitive, affective and metacognitive learning strategies add only 4 percent to the r-squared.

Table 4.13: Model Summary; Regression Analysis Predicting Creativity by Using Subcomponents of Language Learning Strategies

Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.894a
.799
.798
6.158
2
.906b
.821
.819
5.832
3
.909c
.826
.823
5.767
4
.912d
.832
.828
5.687
a. Predictors: (Constant), Social
b. Predictors: (Constant), Social, Cognitive
c. Predictors: (Constant), Social, Cognitive, Affective
d. Predictors: (Constant), Social, Cognitive, Affective, Metacognitive
e. Dependent Variable: Creativity

The results of the ANOVA test of significance of the regression model [F (4, 143) = 177.56, P .05, ω2 = .827 it represents a large effect size] (Table 4.14) indicated that the regression model was statistically significant. Thus the seventh null hypothesis as EFL learners’ use of language learning strategies does not predict their creativity is rejected.

Table 4.14: ANOVA Test of Significance of Regression Model; Predicting Creativity by Using Subcomponents of Language Learning Strategies

Model
Sum of Squares
Df
Mean Square
F
Sig.

Regression
22973.276
4
5743.319
177.566
.000e

Residual
4625.285
143
32.345

Total
27598.561
147

a. Dependent Variable: Creativity
b. Predictors: (Constant), Social
c. Predictors: (Constant), Social, Cognitive
d. Predictors: (Constant), Social, Cognitive, Affective
e. Predictors: (Constant), Social, Cognitive, Affective, Metacognitive

As displayed through figure 4.10 the assumption of homoscedasticity is not met. The spread of dots seems to be narrow on one end and wide on the other end and shows a falling-and-rising pattern.

Figure 4 10: Predicting Creativity by Using Subcomponents of Language Learning Strategies

4.3.8 Testing the Eighth Null Hypothesis

The eighth hypothesis states that:
H 08: EFL learners’ learning style preferences do not predict their creativity.
A multiple regression analysis was run to predict EFL learners’ creativity by using the six subcomponents of the learning style preferences. As shown in Table 4.15, the regression model converged in one step. The kinesthetic style was the only variable entering the model predicting 93.1 percent of scores on creativity (R = .865, R2 = .931).

Table 4.15: Model Summary of Regression Analysis; Predicting Creativity by Using Subcomponents of Learning Style Preferences

Model
R
R Square
Adjusted R Square
Std. Error of the Estimate
1
.965a
.931
.930
3.622
a. Predictors: (Constant), Kinesthetic
b. Dependent Variable: Creativity

The results of the ANOVA test of significance of the regression model [F (1, 146) = 1957.38, P .05, ω2 = .930 it represents a large effect size] (Table 4.16) indicated that the regression model was statistically significant. Thus, the eighth null hypothesis as EFL learners’ use of learning style preferences does not predict their creativity is rejected.

Table 4.16: ANOVA Test of Significance of Regression Model; Predicting Creativity by Using Subcomponents of Learning Style Preferences

Model
Sum of Squares
df
Mean Square
F
Sig.

Regression
25682.892
1
25682.892
1957.385
.000b

Residual
1915.669
146
13.121

Total
27598.561
147

a. Dependent Variable: Creativity
b. Predictors: (Constant), Kinesthetic

As displayed through Figure 4. 11 the assumption of homoscedasticity is not met. The spread of dots seems to be narrow on one end and wide on the other end and shows a falling-and-rising pattern.

Figure 4.11: Predicting Creativity by Using Subcomponents of Learning Style preferences

4.4 Construct Validity

A factor analysis through the varimax rotation was carried out to probe the underlying construct of the subcomponents of learning style preferences, language learning strategies and creativity. It should be mentioned that, the sample size was adequate for running the factor analysis (KMO = .85 .60). Moreover, the correlation matrix employed to run the factor analysis, is an appropriate one (Bartlett’s χ2 = 3201.63, P .05).

Table 4.17: Sampling Adequacy and Sphericity Assumptions
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
.857
Bartlett’s Test of Sphericity
Approx. Chi-Square
3201.633

df
78

Sig.
.000

As shown in Table 4.18, the SPSS extracted two factors, which accounted for 84.47 percent of the total variance.

Table 4.18: Total Variance Explained
Component
Initial Eigenvalues
Extraction Sums of Squared Loadings
Rotation Sums of Squared Loadings

Total
% of Variance
Cumulative %
Total
% of Variance
Cumulative %
Total
% of Variance
Cumulative %
1
9.874
75.957
75.957
9.874
75.957
75.957
6.355
48.885
48.885
2
1.108
8.521
84.478
1.108
8.521
84.478
4.627
35.593
84.478
3
.571
4.393
88.871

4
.485
3.729
92.600

5
.302
2.322
94.922

6
.194
1.492
96.413

7
.154
1.188
97.601

8
.094
.723
98.324

9
.080
.616
98.940

10
.059
.456
99.396

11
.042
.322
99.718

12
.024
.182
99.900

13
.013
.100
100.000

Finally, Table 4.19 displays the factor loadings of the questionnaires under the two extracted factors. The tactile and creativity load on the first factor and compensation loads on the second factor. The other variables

پایان نامه
Previous Entries دانلود پایان نامه ارشد درمورد learning، 148، Correlation Next Entries دانلود پایان نامه ارشد درمورد learning، CR، [r