from a lecturer to a facilitator, from a dominator to a
helper. BL on-line platforms can further support class
activities like group discussions, deliverables
submissions, and feedbacks provisions. Varied BL
materials increase the availability and richness of
learning resources for students to interact with each
other and collaborate.
2.2 The Third-Generation Activity
Theory
Activity theory views that in the process of human
leaning activity, the role of mediations is very
important and it is the key factor for the learning
subject to change the objects and to realize the
cognition development (Vygotsky,1978). Activity
theory has evolved for generations. Vygostky in 1978
firstly proposed the first generation of activity theory
as part of the social cultural theory. The idea of
Vygostky is that human interactions are based on
mediations and interactions with the surrounding
environment. And in n1981, Leont forms the second
generation of activity theory based on the first one.
With the increased interest on individual and common
complexity, together with the existing of social
mediators, the second generation of activity theory
contains more levels of activities. In 1999, Engestrom
and Miettient further propose that human activities
are realized through practices and collaborations.
This is the third generation of activity theory and it
includes at least two sets of interaction activities.
Based on the third-generation activity theory, the
imbalance and mismatch between interactions is
dynamic (Kuutti,1996), and the resolving of conflicts
are helpful in realizing extended learning activity
design. That is to say, the conflicts in the activities are
conducive for the learning object to do extended
learning (Vygotsky,1978). Combining the theoretical
model of third-generation activity theory and the
classification of conflicts, BL design in Business
English Listening and Speaking courses should
consider the very basic activity system, that is the
offline teaching system, which includes the internal
conflicts like students’ learning motivations, teaching
resources, and the learning experience. What’s more,
the conflict between the basic activity system with the
new activity system, that is the BL system is very
obvious as well. Finally, the conflict between the
basic activity system and the on-line SPOC system,
like the integration with SPOC on-line tools with
offline class teaching. The first two conflicts are the
internal conflicts within the basic activity system, and
the third conflict is the conflict between the basic
activity system and the new activity system, the forth
conflict is the conflict in the integration of on-line and
off-line teaching activities. These conflicts are
dynamic due to the characteristics of the learning
subjects. The learning subject continuously interact
and compromise with the conflicts to develop and
evolve further (Kuutti, 1996). The third-generation
activity theory not only provide the theoretical bases
for the design of blended teaching activities, but also
provided the framework and guidance for blended
teaching evaluation model construction and
application.
2.3 PLS-SEM Formula
PLS-SEM is short for Partial Least Square-Structural
Equation Modelling. PLS-SEM includes
measurement scheme, modelling scheme, and the
weighting scheme. And the two main categories of
PLS-SEM are Reflective measurement and
Informative measurement, as it is shown is Figure 1.
Partial least squares (PLS) is an iterative estimation
that combines principal component analysis with
multiple regression. It is a causal modelling method,
which is used in Sweden, the European Union and the
United States. PLS method has many advantages:
there is no distribution requirement, while LISEREL
method, one of the other methods, assumes that the
joint distribution of explicit variables is multivariate
normal, and the goal is to predict according to the
block structure, internal relationship, and causal
prediction relationship. PLS estimation is consistent
and basically consistent when the sample size is large
and there are many explicit variables for each hidden
variable.
Figure 1: PLS-SEM Formula.
For Reflective measurement, it can be clearly seen
that the latent variable Y causes items, which are
expressed through indicators. For Informative
measurement, it is obvious that the latent variable Y
is caused by items, that is, the indicator forms the