Testing Lecture Comprehension Through Listening-to-summarize Cloze Tasks

5.5.2.4 Students

According to statistics generated from student questionnaires, Cronbach’s Alpha
(see Table5.10) here in the reliability statistics is 0.915, also a very highfigure
showing satisfactory internal reliability of the questionnaire.
The current KMO measure (see Table5.11) for the student questionnaire is
0.856, also quite suitable for satisfactory main factor analysis. In the same table, the
Bartlett’s Test of Sphericity has the corresponding P << 0.05, which again shows
the variables are correlated and the data can be run for exploratory factor analysis.
The Table of Total Variance (see Table5.12) here shows that thefirst factor
accounts for 26.101% of the total, which means thefirst extracted factor can
actually explain more than one fourth of the variance. The following factors in turn
are second factor accounting for 8.929%, the third 6.832%, the fourth 5.126%, the
fifth 4.381%, the sixth 3.800%, the seventh 3.475%, the eighth 3.005% and the last
one 2.872%. Altogether the nine factors explain 64.521% of the total while all the
remaining factors are insignificant.
Apparently, the nine-factor model with items sporadically loaded on different
factors is too complex and there are too many dimensions for interpretation. It is
worth noticing that thefirst four factors already account for 46.988% of the total
variance (see Table5.13). In addition, the Scree plot which shows the eigenvalues
of the correlation matrix in descending order of magnitude helps visualize the
relative importance of factors. In the following Scree Plot (see Fig.5.2), thefirst
four factors present a sharp drop in the plot and a noticeable“bend”occurs at the
fourth factor while the subsequent ones can be ignorable. On the other hand, when
the number of factors is restricted to 8, 7, 6 and 5 respectively, the corresponding
Rotated Component Matrix does not demonstrate that all the extracted factors are
substantially represented by their items, in another word, there are not enough items
heavily loaded on their corresponding factors, so, after the number of factors is
fixed to 4, each of the four factors turns out to be satisfactorily represented by a
number of items and the Factor Matrix is greatly simplified, which in turn is
advantageous to the interpretation of latent factors.

Table 5.10 Reliability
statistics of student
questionnaire

Reliability statistics

Cronbach’s Alpha N of Items

0.915 35

Table 5.11 KMO and
Bartlett’s test of student
questionnaire

KMO and Bartlett’s test

Kaiser-Meyer-Olkin measure of sampling

adequacy

0.856

Bartlett’s test of

sphericity

Approx.

Chi-square

3123.872

Df 595

Sig. 0.000

56 5 Exploring Task Demands of TEM 8 Mini-Lecture...