1. Introduction

Constructivism is a philosophical framework for learning based on Piaget’s theory and formally introduced by von Clasersfeld during the 1970’s. Constructivism argues that knowledge is not passively received from the environment, but is actively constructed by the learner through a process of adaptation based on and constantly modified by the learner’s experience of the world ^[1]. The application of the ideas of constructivism in the teaching process has become very popular during the last decades, especially in school education. The steps of a typical framework for teaching which is based on those ideas are the following:

• Orientation (S₁): This is the starting step which connects the past with the present learning experiences and focuses student thinking on the learning outcomes of the current activities.

• Exploration (S₂): In this step students explore their environment to create a common base of experiences by identifying and developing concepts, processes and skills.

• Formalization (S₃): Here students explain and verbalize the concepts that have been explored and the instructor has the opportunity to introduce formal terms, definitions and explanations for the new concepts and processes and demonstrate new skills.

• Assimilation (S₄): In that step students develop a deeper and broader conceptual understanding and obtain more information about areas of interest by practicing on their new skills and behaviors.

• Assessment (S₅): This is the final step of the teaching process, where learners are encouraged to assess their understanding and abilities and teachers evaluate student skills on the new knowledge.

Depending on the student reactions in the classroom, there are forward or backward transitions between the three intermediate steps (S₂, S₃ and S₄) of this framework during the teaching process, the flow-diagram of which is shown in Figure 1.

Figure 1: The flow-diagram of the teaching process

In this article a Markov chain is introduced on the steps of the teaching process and through it interesting conclusions are derived for the teaching effectiveness.

2. Markov Chains

Roughly speaking a Markov chain (MC) is a stochastic process that moves in a sequence of steps (phases) through a set of states and has a one-step memory. In other words, the probability of entering a certain state in a certain step depends on the state occupied in the previous step and not in earlier steps. This is known as the Markov property. However, for being able to model as many real life situations as possible by using MCs, one could accept in practice that the probability of entering a certain state in a certain step, although it may not be completely independent of previous steps, it mainly depends on the state occupied in the previous step ^[2].

The basic concepts of MCs were introduced by Andrey Markov (Figure 2) in 1907 on coding literal texts.

Figure 2: A. Markov (1856-1922).

Since then the MC theory was developed by a number of leading mathematicians, such as A. Kolmogorov, W. Feller, etc. However, only from the 1960’s the importance of this theory to the natural, social and applied sciences has been recognized ^{[3][4][5][6][7][8]}.

2.1 Finite Markov Chains

When the set of states of a MC is a finite set, then we speak about a finite MC. For general facts on finite MCs we refer to Chapter 2 of the book ^[9].

Let us consider a finite MC with n states, say S₁, S₂, …, S_n, where n is a nonnegative integer, n 2. Denote by p_ij the transition probability from state S_i to state S_j , i, j = 1, 2,…, n ; then the matrix A= [p_ij] is called the transition matrix of the MC. Since the transition from a state to any other state (including its self) is the certain event, we have that

p_i1 + p_i2 + . + p_in  , i = 1, 2,., n            (1)

The row-matrix P_k = [p₁^(k) p₂^(k)… p_n^(k)], known as the probability vector of the MC, gives the probabilities p_i^(k) for the MC to be in state i at step k , for i = 1, 2,…., n and k = 0, 1, 2,…. Obviously we have again that

Using conditional probabilities on can show ([9], Chapter 2, Proposition 1) that for all nonnegative integers k is

P_k+1= P_k A (3).

Then, a straightforward induction on k gives that

P_k =P_o A^k  (4) .

Equations (3) and (4) enable one to make short run forecasts for the evolution of the various situations that can be represented by a finite MC. In practical applications we usually distinguish between two types of finite MCs, the absorbing MCs (AMCs) and the ergodic MCs (EMCs).

2.2. Absorbing Markov Chains

A state of a MC is called absorbing if, once entered, it cannot be left. Further a MC is said to be an AMC if it has at least one absorbing state and if from every state it is possible to reach an absorbing state, not necessarily in one step.

Working with an AMC with k absorbing states, 1 k < n, one brings its transition matrix A to its canonical (or standard) form A^* by listing the absorbing states first and then makes a partition of A^* to sub-matrices as follows

(5)

In the above partition of A*, I_k denotes the unitary k X k matrix, O is a zero matrix, R is the (n – k) X k transition matrix from the non-absorbing to the absorbing states and Q is the (n – k) X (n – k) transition matrix between the non-absorbing states.

It can be shown (^[10], Section 2) that the square matrix I_{n – k} - Q, where I_{n – k} denotes the unitary n-k X n-k matrix is always an invertible matrix. The fundamental matrix N of the AMC is defined to be the inverse matrix of I_n–k – Q. Therefore (^[11], Section 2.4). 

(6)

In equation (6) D (I_{n – k} – Q) and adj (I_n–k – Q) denote the determinant and the adjoin of the matrix I_{n – k} – Q respectively It is recalled that the adjoin of a matrix M is the matrix of the algebraic complements of the transpose matrix M^t of M, which is obtained by turning the rows of M to columns and vice versa. It is also recalled that the algebraic complement m_ij΄of an element m_ij of M is calculated by the equation

m_ij΄ = (-1)^i+jD_ij   (7)

In equation (7) D_ij denotes the determinant of the matrix obtained by deleting the i-th row and the j-th column of M.

It is well known (^[5], Chapter 3) that the element n_ij of the fundamental matrix N gives the mean number of times in state S_i before the absorption, when the starting state of the AMC is S_j , where S_i and S_j are non-absorbing states.

2.3. Ergodic Markov Chains

A MC is said to be an EMC, if it is possible to go between any two states, not necessarily in one step. It is well known (^[5], Theorem 5.1.1) that, as the number of its steps tends to infinity (long run), an EMC tends to an equilibrium situation, in which the probability vector P_k takes a constant price P = [p₁ p₂ …. p_n], called the limiting probability vector of the EMC. Therefore, as a direct consequence of equation (3), the equilibrium situation is characterized by the equation

P = PA, with p₁+ p₂+ ….+ p_n = 1      (8)

The entries of P express the probabilities of the EMC to be in each of its states in the long run, or in other words the importance (gravity) of each state of the EMC.

Let us now demote with m_ij the mean number of times in state S_i between two successive occurrences of the state S_j, i, j = 1, 2, …., n. It is well known (^[5], Theorem 6.2.3) then that

3. A Markov Chain Model for the Teaching Process

3.1. The Model

Here we introduce a finite MC having as states S_i, i = 1, 2,…, 5, the corresponding steps of the teaching framework described in our Introduction. From the flow-diagram of Figure 1 it becomes evident that this chain is an AMC with S₁ being its starting state and S₅ being its unique absorbing state. The minimum number of steps before the absorption is 4 and this happens when we have no backward transitions between the three middle states S₂, S₃ and S₄ of the chain. Denote by p_ij the transition probability of the MC from state Si to state Sj, for i, j =1, 2,…,5. Then the transition matrix A of the chain and its canonical form A* are the following:

Denote by I₄ the 4X4 unitary matrix. Then the fundamental matrix of the AMC is

Therefore, since S₁ is the starting state of the above AMC, it becomes evident that the mean number of steps before the absorption is given by the sum

    (10)

It becomes evident too that the bigger is T, the more are the student difficulties during the teaching process. Another factor of the student difficulties is the total time spent for the completion of the teaching process. However, the time is usually fixed in a formal teaching procedure in the classroom, which means that in this case T is the unique measure of the student difficulties.

3.2. A Classroom Application

The present application took place at the Graduate Technological Educational Institute of Western Greece for teaching the concept of the derivative to a group of fresher students of engineering. The instructor used the teaching framework that has been described in our Introduction as follows:

Orientation: The student attention was turned to the fact that the definition of the tangent of a circle as a straight line having a unique common point with its circumference does not hold for other curves (e.g. for the parabola). Therefore, there is a need to search for a definition of the tangent covering all cases and in particular of the tangent at a point of the graph of a given function.

Exploration: The discussion in the class led to the conclusion that the tangent at a point A of the graph of a given function y=f(x) can be considered as the limit position of the secant line of the graph through the points Α(a, f(a)) and Β(b, f(b)), when the point Β is moving approaching to Α either from the left, or from the right (Figure 3). But the slope of the secant line AB is equal to , therefore the slope of the tangent of the graph at A is equal to the limit of the above ratio when b tends to a.

Figure 3: Tangent at a point of the graph of a given function

Formalization: Based on what it has been discussed at the step of exploration, the instructor presented the formal definition of the derivative number f΄(a) at a point Α(a, f(a)) of a given function y=f(x) as the limit (if there exists) of when , and of the tangent of the graph of y=f(x) at A as the straight line through A with slope f΄(a). Some examples followed of calculating the derivative at a given point of a function and the tangent of its graph at this point. Then the definition of the derivative function y΄ = f΄(x) of the function y=f(x) was given and suitable examples were presented to show that its domain is a subset of the domain of y=f(x).

Assimilation: Here the fact that the derivative y΄ = f΄(x) expresses the rate of change of the function y=f(x) with respect to x was emphasized and its physical meaning was also presented connected to the speed and the acceleration at a moment of time of a moving object under the action of a steady force. The fundamental properties of the derivatives followed (sum, product, composite function, etc.) as well as a list of formulas calculating the derivatives of the basic functions and applications of them.

It has been observed that the student reactions during the teaching process led to 2 transitions of the discussion from state S₃ (formalization) back to state S₂ (exploration). Therefore, since from state S₂ the chain moves always to S₃ (Figure 1), we had 3 in total transitions from S₂ to S₃. The instructor also observed 3 transitions from S₄ (assimilation) back to S₃. Therefore, since from state S₃ the chain moves always to state S₄ (Figure 1), we had 4 in total transitions from S₃ to S₄. In other words we had 3+3 = 6 in total “arrivals” to S₃, 2 “departures” from S₃ to S₂ and 4 “departures” from S₃ to S₄. Therefore p₃₂ = 2/6 and p₃₄ = 4/6. In the same way one finds that p₄₃ = 3/4 and p₄₅ = 1/4. Replacing the above values of the transition probabilities to equation (10) one finds that the mean number T of steps before the absorption of the MC is equal to 14. Consequently, since the minimum number of steps before the absorption is 4, the students faced significant difficulties during the teaching process. This means that the instructor should find ways to improve his teaching procedure of the same subject in future.

3.3. An Important Remark

In certain cases it is possible to develop either an AMC or an EMC model for representing the same situation. In case of the teaching process, for example, the flow diagram of Figure 1 could be revised by assuming that, when the teaching process of a subject matter is integrated, then a new process starts by the instructor for teaching the next subject of the course. That means that the teaching process is transferred back from step S₅ to S₁ to be repeated from the beginning again. The revised flow diagram of the teaching process, therefore, takes the form of Figure 4.

Figure 4: Revised flow diagram of the teaching process

In this case the resulting MC on the steps of the teaching process is obviously an EMC.  Since S₁ is the starting state of the EMC it becomes evident that the sum m=m₁₅+m₂₅+m₃₅+m₄₅ calculates the mean number of steps of the EMC between two successive occurrences of the state S₅. Therefore, the mean number of steps for the completion of the teaching process will be m+1, since it includes also the step S₅. With the help of equation (9) one finds that

(11)

The value of the limiting probability p₅  is calculated with the help of the matrix equation (8). In this equation the transition matrix of the EMC differs from the corresponding matrix A of the AMC of section 3.1only with respect to the last row, where 1 is transferred from the fifth to the first column and its other entries are 0. Performing the necessary calculations, equation (8) leads to a linear system of five equations with respect to the p_i’s, i = 1, 2, 3, 4, 5. Adding by members the first four of those equations one finds the fifth one. Thus, replacing the fifth equation with the equation p₁+p₂+p₃+p₄+p₅ = 1 and solving the resulting 5X5 linear system one finds the required value of p₅ and with the help of equation (11) calculates m (for more details see ^[9], pp. 62-63). It becomes evident that, the greater is the value of m the more the student difficulties during the teaching process. Concerning the classroom application of section 3.2 , after performing all the necessary calculations one finds that m+1 = T.

4. Conclusion

In the paper at hands a mathematical representation of the teaching process based on the principles of constructivism for learning was developed with the help of the theory of MCs. This representation enables the instructor to evaluate the student difficulties during the teaching process, which is very useful for reorganizing the plans for teaching the same subject in future. Although the theoretical development of the MC model is quite laborious, its final application in practice is too easy and straightforward. It requires only to count the backward movements   among the three middle steps S₂, S₃ and S₄ of the teaching process. The theory of MCs can be used for several other applications in Education (e.g. see Chapter 3 of ^[9]) and this is one of the main targets of our future research.