Promoting Mediation in Learning Error on Teaching Algorithms

Rui Ogawa

, Leticia M. Peres

, Alexander R. Kutzke

and Fabiano Silva

Federal Institute of Education, Science and Technology of Mato Grosso (IFMT), Canarana, MT, Brazil

Department of Computer Science, Federal University of Parana (UFPR), Curitiba, PR, Brazil

SEPT, Federal University of Parana (UFPR), Curitiba, PR, Brazil

Keywords:

Error Mediation, Social-interactionism, Teaching-learning, Teaching Algorithms.

Abstract:

This paper addresses the results of a research on the adoption of a computer platform for mediation of learning

errors in algorithm exercises for computer science students. The importance of the interaction between teacher

and learner is discussed as a way to reconsider learning errors. The need to evaluate tools for mediation of

learning errors is brought to the discussion. The essential activities a teacher should develop using a tool

that helps in the practice of algorithm teaching are listed, and based on them, a method that promotes the

interaction between teacher and learners is proposed as a mean to enhance the effectiveness in mediation of

learning errors.

1 INTRODUCTION

One of strategies for effective acquisition of knowl-

edge or skills is learning error mediation process,

which essencially consists of a presentation of a help

message for the learner, when a mistake is made

(Ramos, 2011). The use of tools to support the me-

diation of errors can help the learning-teaching pro-

cess (Moura and Peres, 2017). The FARMA-ALG

(Kutzke and Direne, 2014) is a tool based on error

mediation that, through analysis of relations and inter-

connections of learner errors, can capture the relevant

properties of this type of mediation.

With the availability of tools that allows mediation

of learning error on algorithm contents, it is identiﬁed

the necessity to study its use by teachers and learners.

Besides that, there is a lack of methods which guide

the creation of new ways of instructional project on

a social interactionist model, based on the interaction

by technology. The results of the efforts of teacher

may fall short expectations if there is not a guide

which orients him/her on the correct and effective use

of computational tools focusing on interactionism.

This paper presents the results of a research ap-

plying the FARMA-ALG tool for the observation of

learner-teacher interaction with error mediation on

ﬁrst algorithms course of computer science under-

graduate students. It was observed a reduction of

46% on the amount of erroneous attempts after learn-

ers had received some sort of intervention from the

teacher. There was an increase of 7.5% on the amount

of correct answers to the exercises. Despite these pos-

itive results, it was veriﬁed that the percentage of in-

correct answers, was still high, close to 70% of all

answers, even with interventions of the teacher. It

was also observed that the level of interaction was ex-

tremely low, reaching only 1.93% of the total. Con-

sidering the amount of submitted answers by the

learners that received some sort of interaction was

quite restrict, it was noticed that such behaviour oc-

curs, largely, due to the lack of a method that gives

orientations for the teacher to correctly use the inter-

action resources that FARMA-ALG provides.

Considering the above mentioned, this paper

has the objective of presenting the inherent aspects

in learning error mediation on teaching algorithms,

bringing to the discussion the need of evaluating a tool

that work on the context of learning error mediation,

and proposing a method which promotes interaction

at error mediation process.

This paper is organized as follows. Section 2

presents the conceptual basis of mediation of learn-

ing errors. Section 3 depicts an evaluation of learner-

teacher interaction. A method to guide the teaching

process to promote the interaction is proposed in Sec-

tion 4. Finally, Section 5 concludes with considera-

tions and future work.

358

Ogawa, R., Peres, L., Kutzke, A. and Silva, F.

Promoting Mediation in Learning Error on Teaching Algorithms.

DOI: 10.5220/0006776603580365

In Proceedings of the 10th International Conference on Computer Supported Education (CSEDU 2018), pages 358-365

ISBN: 978-989-758-291-2

2 LEARNING ERRORS

MEDIATION

Computer programming is a complex and high level

activity that demand a high cognitive load. This com-

plex scenario requires hard work from both teachers

and learners on the process of teaching and learning.

During this process, the learner faces difﬁculties that,

potentially, generate errors (incorrect answers to ex-

ercises).

During the teaching of complex knowledge ar-

eas, such as computer programming, it is essential to

overcome the common negative view of the error (for

which the error is only a sign of a learner problem). It

is necessary to make the error an integral part of the

teaching and learning process. In other words, it is

necessary to mediate the error.

Mediating the error means making it the subject of

an educational activity, i.e., problematizing it. Putting

it as part of a development process. An error indicates

a contradiction in thinking process, an inconsistency

not noticed by the learner. It is, then, up to the teacher

to point this contradiction. Making it the gateway to

a higher level awareness. However, unfavorable ma-

terial conditions faced in most educational environ-

ments can prevent or hinder the error mediation.

Studies conducted by (Vygotsky et al., 1934) in-

dicate that the stimulus of the cognitive development

of learner, to obtain solid and deep knowledge, is

closely related to the interaction between the indi-

viduals. Based on this assertions of Vygotsky, and

for the learner cognitive development, it is desirable

that there be sufﬁcient mediation to promote the great-

est possible interaction between those involved in the

teaching and learning process – teachers and their

learners.

It is necessary to review some considerations

about learning errors, since the errors should be seen

as part of the learning process. According to (Sforni,

2004), it is necessary to see the error as a moment

when a situation of conﬂict can lead to the learning

of a new concept. One negative consequence is that

the error can generate a feeling of incapacity in the

learner, leading him to the dropping of the algorithms

subjects and consequent possible evasion of computer

science courses. The teacher, once instrumented with

a computational tool to support the contexts of error,

may have the opportunity to contribute to the reduc-

tion of cases of dropouts, enrollment cancel and eva-

sions. Increasing the interaction between teachers and

learners is one of the strategies proposed by (Chicker-

ing and Ehrmann, 1996) when it comes to the use of

computational resources in education. Faced with an

error of the learner, the interaction with the teacher,

through comments, can contribute to the elucidation

of common problems and thus facilitate their under-

standing and resolution.

According to an interactionist approach, FARMA-

ALG

(Kutzke and Direne, 2014) is proposed to be

an instrument to promote the error mediation: an

implementation of a framework for computer pro-

gramming education which is intended to allow the

teacher to view and to manipulate answers records

and their relations. It presents the concept of simi-

larity graph between answers, over which the system

provides features such as: semi-automatic classiﬁca-

tion of answers, recommendation of answers for sim-

ilar learner groups and different types of data visual-

ization. With this instrument, teachers are capable of

analyzing concept formation process of the learner in

a more concrete way, in opposition to an immediate

and empirical vision. Thus, promoting the error me-

diation.

There are few researches on evaluation of tools in

the ﬁeld of learning error mediation. The research

conducted by (Moura and Peres, 2017) describes the

evaluation of a tool for promoting feedback. This

research consisted of analyzing the use of feedback

mechanism of errors and hits through methodological

referrals. The results demonstrate that the tool creates

a favorable environment for learners to question, ex-

press and collaborate. The authors also veriﬁed that

the feedback mechanism allowed the error measure-

ment and accuracy have effectively helped learners

and teachers identify mistakes and then act on them.

In the ﬁeld of studies on methodologies of use

of these tools, the research of (Garcia and Direne, )

have great relevance because it proposes a general-

ist methodology based on processes for the success-

ful accomplishment of learning sessions, which are

applicable for teachers and learners.

3 EVALUATION OF

INTERACTION IN LEARNING

ERROR MEDIATION

Although FARMA-ALG promotes the interaction be-

tween teachers and learners, there is a need for stud-

ies that allow us to state how the result of this inter-

action contributes to both. When evaluating FARMA-

ALG, it was veriﬁed that, although its effectiveness as

a tool that promotes the mediation of error has been

proven, the amount of interactions between teachers

and learners was extremely low. This is due to the

http://farma-alg.com.br

Promoting Mediation in Learning Error on Teaching Algorithms

359

lack of concise documentation and a speciﬁc method-

ology that guides teachers and learners about the re-

sources available in the tool. Among the resources,

the possibility of interaction between the teacher and

the learner in situations where the learner incurs an er-

ror during the resolution of an exercise is highlighted.

In order to provide the best possible use of the re-

sources based on the social-interactionism proposed

by FARMA-ALG, a study was carried out on the ac-

tivities of the teacher when using the platform and,

based on this information, a methodology was pro-

posed that promotes the interaction .

A computer based environment that promotes me-

diation through error will only have its goals fully

achieved if teachers and learners effectively utilize

the interaction features offered by the tool. No mat-

ter how efﬁcient the mechanism of sending comments

and messages, if the teachers does not use them,

the social interactionist intention of the tool becomes

null. Seen in this terms, its use becomes restricted

only to the functional aspects of creating and making

available speciﬁc learning objects for teaching algo-

rithms. Even if the teacher have the resources to vi-

sualize and manipulate the answers through the Simi-

larity Graph, the real interaction with his learners will

only happen if there are messages and comments.

According to (Hannaﬁn, 1992), there is a lack of

a guiding structure on how the teacher should use

digital technology tools to achieve improvements in

the teaching-learning process. In the speciﬁc case of

FARMA-ALG use, there is no method or guide that

helps guiding teachers to make use of mediation fea-

ture that the tool provides. There is a need, therefore,

to propose and describe a sequence of structured pro-

cesses aimed to the teaching practice with the purpose

of guiding the teacher in order to achieve effectively

mediation.

3.1 Methodology of the Study

To support the development of an effective method,

it was necessary to deﬁne some objectives, practical

studies and analyzes. For the development of the re-

search, the following actions were foreseen:

• deﬁne the universe of research;

• perform critical analysis of interaction messages;

• perform statistical analysis of comments.

Data from answers of 229 students who submitted

8,887 responses, collected between 2015 and 2016

from seven classes of Algorithms and Data Struc-

ture I, from two universities, were analyzed. Three

groups were from the year 2015, totaling 118 students

from two universities who submitted 4,854 answers

to questions applied by two teachers. Four classes

were from the year 2016, with 105 students from

a single university, which generated 4,033 answers

to questions applied by four teachers. All classes

whose answers were collected are from the initial

phase of the computer science courses. The objective

was to evaluate the greatest possible number of an-

swers and interactions between teachers and students

who used FARMA-ALG. Relationships as percent-

ages of correct and incorrect answers, answers that re-

ceived interventions through teachers comments, stu-

dent scores and number of attempts were analyzed .

Through this analysis, the objective is to detect be-

haviors that can show how and at what level the use

of FARMA-ALG inﬂuences the results. Based on this

evaluation, enough information was obtained to ana-

lyze the effectiveness of FARMA-ALG as a tool to

support teaching practice in algorithm teaching.

Was tried to ﬁnd out, for example, in what types

of questions students ﬁnd greater or lesser degree of

difﬁculty and what are the most common mistakes.

An important item was the analysis of the relationship

between the answers that received some kind of in-

tervention by the teacher and the number of attempts

submitted by question. The interaction between stu-

dents and teachers was veriﬁed in only 82 answers

(1.93%) that received comments from the teacher, a

fact that may indicate a lack of time for teacher fol-

low up and send comments to all students who made

mistakes in their answers. Another explanation for

the low level of interaction was the nonexistence of

a method that guides both teacher and student during

the use of FARMA-ALG.

The answers that received most comments from

the teacher were those that presented some logic er-

ror, which totaled 27 answers, or 35.53 % of the to-

tal. Next, the questions that contained user dialogues

(unspoken in the statement) and output errors, with

18 answers (23.68%). Problems with syntax appear

in nine answers, representing 11.84% and as a con-

sequence, some of them caused ﬁve compilation er-

rors (6.58%). Some students submitted seven answers

(9.21%) whose solution was based on trial and error

mechanisms, with no guarantee that the answer would

be correct.

From the total of 82 answers to which the teacher

made some intervention, 71 were incorrect and 10

were correct. In 46 of them (64.78%) the students

made advances, being able to identify and correct the

errors. On the other hand, 25 responses (35.22%) re-

mained incorrect even with the teacher intervention.

Although the number of questions that were not com-

pleted was high – even with the teacher help – it is

noticeable the improvement in the performance of the

students who got help. From the 25 answers for which

CSEDU 2018 - 10th International Conference on Computer Supported Education

360

students did not continue the development of resolu-

tion, teacher interaction contributed to the fact that al-

most two-thirds of the answers that were wrong were

corrected.

The analysis were also performed using descrip-

tive and inferential techniques. The data of interest

for the statistical analysis were:

• the number of attempts until the ﬁrst intervention;

• the number of attempts since the ﬁrst intervention,

until the correct response;

• the time the teacher took to complete the ﬁrst in-

tervention;

• the time, from the ﬁrst intervention, to the correct

response.

In the descriptive analysis, for the aforementioned

datasets, means, standard deviations and coefﬁcients

of variation were compared. In the inferential analy-

sis, the same data were submitted to statistical tests to

compare means to validate or not, the results observed

in the descriptive analysis.

A reduction of 46% in the number of attempts

was detected after the ﬁrst teacher intervention. In

contrast, there was an increase in the time spent by

the student to respond correctly after receiving help,

which suggests that, properly instructed and guided,

he went from an earlier state of “trial and error”, to

more grounded answers. Considering the results of

the studies, was veriﬁed the need to prepare documen-

tation, with a concise and replicable method, that can

help teachers and students that use FARMA-ALG, so

that all their social interacionist potential is effectively

achieved.

4 METHOD PROPOSAL

The use of a tool such as FARMA-ALG requires or-

ganization and planning from teacher, so that classes

can be efﬁcient and produce the desired result. The

functions and activities that the teacher must perform

for the effective use of the tool, and are part of the

proposed method, are listed below:

1. Identify which course content will be taught;

2. Teach the classes using the prefered method;

3. Elaborate subjects related to Learning Objects,

containing: a) Name and b) Description;

4. Edit Learning Objects, creating: a) Introduction;

b) Exercises; c) Questions; d) Test cases;

5. Correct exercises that was proposed in FARMA-

ALG;

6. Identify learners with difﬁculty levels who require

intervention;

7. Establish interaction with these learners through

comments and messages;

8. Follow-up evolution of learners after intervention.

The activities listed above are essential for the ef-

fective use of FARMA-ALG. However, they are not

yet organized into a structured way that permits a

methodology that can be replicated. Special attention

should be given to corrective activities, identiﬁcation

of learners with difﬁculties, interaction and follow-

up. In this section, the processes of the methodology

must contain to contemplate the needs to guide the

teacher in a successful implementation of interaction-

ist model, implemented by FARMA-ALG platform,

are described.

The activities are distributed among four phases,

as follows:

1. Planning;

2. Elaboration and providing;

3. Evaluation;

4. Interaction and mediation.

In this model, activities of each phase must be

completed in order to advance to the next, according

to description of Figure 1. The structure provides an

iterative mechanism, in a manner that the teacher and

learners (professor and students, as in the ﬁgure) can

go back to previous phases and use the information

to improve activities and achieve the proposed objec-

tives.

The objectives of each phase and the products to

be delivered are listed in Table 1 and are detailed in

steps in the following sections.

4.1 Planning

The main goal of the planning phase is to deﬁne what

are the learning objectives that the subject should

offer. These objectives describe what concepts and

skills learners (students) should know and understand

to be able to analyze problems, propose solutions and

implement them. This phase is subdivided into three

stages, in which the teacher (professors) will identify

the needs and requirements, deﬁne what the learning

objectives will be, and analyze how the use of the

available computerized structure should be.

Step 1: Identiﬁcation of Needs and Requirements.

1. The skills that learners must acquire in complet-

ing the proposed activities should be identiﬁed. A

good basis for the choices is the syllabus of the

course, usually provided by the course coordina-

tion, as well as its pedagogical project.

Promoting Mediation in Learning Error on Teaching Algorithms

361

Figure 1: Phases of method.

Table 1: Objectives and products of phases.

Phase Phase objective Phase product

Phase 1 Deﬁnition of subject objectives Teaching plan

Phase 2 Learning Objects Elaboration and providing in

FARMA-ALG

Learning Objects containing exercises and ques-

tions

Phase 3 Evaluation, correction and identiﬁcation of difﬁ-

culties

Report pointing out the main difﬁculties of learn-

ers

Phase 4 Interaction and mediation with learners Intervention with learners and appropriateness of

exercises

2. The programming languages to be used must be

the same as those which FARMA-ALG supports.

Based on this information, the teacher should in-

dicate to learners a reference guide of the adopted

language in order to solve any basic syntax errors.

3. As the subject requires high level of abstraction

and cognition skills of aspects about logic, it is

interesting to carry out a pre-evaluation in order

to diagnose eventual discrepancies of knowledge.

Step 2: Deﬁning Learning Objectives. Based on

results observed in diagnostic evaluation, in previous

step, the following steps should be performed:

1. Taking the list of competencies obtained in Step

1, identify the skills that the learner must acquire

at the end of the assessment.

2. According to the skills identiﬁed in the previous

step, the contents that must be contained in the

Learning Objects are deﬁned.

3. Write, based on the previous steps, a formal list of

learning objectives.

Step 3: Analysis of the Feasibility to Achieving

Learning Objectives using FARMA-ALG.

1. Evaluate all learning objectives and identify, ac-

cording to the complexity level of each one, which

can be transformed into Learning Objects in order

to make it available in FARMA-ALG.

2. Objectives that do not ﬁt the above possibility

should be treated with an alternative platform ap-

proach.

4.2 Elaboration and Providing

At this stage the teacher (professor) will, based on the

subject syllabus, develop Learning Objects that will

be available as activities in FARMA-ALG. These ac-

tivities are compounded by algorithm questions, or-

ganized into lists of exercises whose difﬁculty level

gradually increases. At the three steps that comprise

this phase, the teacher will elaborate the activities,

verify the adherence of the activities to the proposal

and resources offered by FARMA-ALG, and deﬁne

the work plan with the schedule of sessions.

Step 1: Choice of Exercises.

1. According to what was achieved in the deﬁnition

of learning objectives in Step 2 of the previous

phase, select the content that will be worked on.

2. Establish how many exercises will be required to

achieve the goal, according to the results of the

pre-assessment. Deﬁne which exercises should

be included in the ﬁrst list, based on the pre-

evaluation.

3. Deﬁne in how many lists the other exercises will

be divided and how will be the distribution among

the lists, according to the difﬁculty of the exer-

cises. Each list should contain exercises whose

difﬁculty allows learners (students) to be able to

CSEDU 2018 - 10th International Conference on Computer Supported Education

362

solve them, even if they need some kind of help.

Step 2: Elaboration of Learning Objects.

1. Turn each exercise into a Learning Object com-

patible with the FARMA-ALG prerequisites.

2. Distribute the exercises into lists according to the

amount of levels the teacher has deﬁned as neces-

sary for the learner to successfully complete a list

and be prepared for the next one.

Step 3: Schedule Elaboration.

1. Elaborate a schedule of online availability of each

list of exercises, according to the progress of the

subject in the classroom.

2. Deﬁne the accomplishment of activities according

to the order of lists of exercises, their duration and

the due date in which they should be answered. To

do so, it is necessary to revise the availability of

time of teachers, monitors and learners.

3. Establish deadlines for resolution of proposed ex-

ercises.

4. Indicate who will be the instructors of the subjects

and inform the class.

4.3 Evaluation

At evaluation phase, the teacher (professor), together

with the tutor, will correct the answers of exercises

submitted by learners (students) in FARMA-ALG.

During correction, it is critical to identiﬁed the main

difﬁculties faced by learners. At the end of this phase,

a report should be generated indicating which aspects

were presented as difﬁculties in solving the exercises.

Step 1: Preparing for the Session.

1. Compose a FARMA-ALG operating instructions

manual and make it available on the platform,

preferably on the home screen. It should state

the technical hardware, software, and network re-

quirements for users to have access to.

2. Publish the periods in which each exercise list will

be available.

3. Notify learners and tutors, via e-mail, the address

to access the platform, as well as the necessary

procedures to register and enroll in the course.

4. Deﬁne how learners will get support in case of

problems with the platform.

5. Publish Learning Objects so the enrolled learners

can access the lists of exercises and solve the pro-

posed questions.

Step 2: Session Development.

1. Before actually releasing the platform for use,

perform connection tests to diagnose any prob-

lems in a timely manner to solve them.

2. Grant access to FARMA-ALG to learners, only to

the list of exercises whose content was previously

developed in the classroom.

3. Periodically, monitor learner participation, check-

ing if they are getting access to questions and sub-

mitting their answers without any kind of prob-

lem.

Step 3: Evaluation of Responses During Submis-

sion Period.

1. On the FARMA-ALG main panel, check the fol-

lowing items:

• Write down the number of incorrect answers.

• Check for comments and messages sent by

learners, directed to the teacher or tutors.

2. In the ’Graph Manipulation’ Menu, with the ap-

propriate use of the available ﬁlters, perform the

following actions:

• Set up the ﬁlters in order the search carry out

for a question, of a class and that is in a certain

category of error. i.e.: Question: Triângulo de

Pascal (Pascal’s Triangle); Class: Alg-2016/1-

ci055-B; Category: Saída (Output). See exam-

ple in Figures 2 and 3.

• Identify the incorrect answers with the highest

degree of similarity between them.

Figure 2: Search using ﬁlters.

3. Using the ’Timeline’ Menu, perform a procedure

similar to the previous step, in order to identify

responses with the highest degree of similarity as

possible.

4. Additionally, use the ’Search’ features to ﬁnd er-

ror patterns.

5. Use, if necessary, the ’Tags’ feature, which allows

you to select responses with certain types of er-

rors, as shown in Figure 4.

Promoting Mediation in Learning Error on Teaching Algorithms

363

Figure 3: Answers found.

Figure 4: Search result for error tags.

4.4 Interaction and Mediation

At the fourth and last phase, interaction occurs be-

tween learners (students), tutors and teachers (profes-

sors). As questions are being answered, teachers and

tutors should identify learners who have encountered

difﬁculties. The observation of similarity in incorrect

answers, carried out in the previous phase, should be

used to subsidize this task. Then, teacher and tutor

should promote interaction with the learners by send-

ing speciﬁc comments about the most common mis-

takes made in the attempts to solve the exercises. If

necessary, in addition to talking with the learner, in

order to understand their difﬁculty and to assist them,

the appropriateness of the exercise may be necessary.

Step 1: Identify Learners who Need Support.

1. Read the messages and comments sent by learn-

ers. If they have sent them before the teacher, they

are asking for clariﬁcation.

2. Identify learners who have made mistakes, ac-

cording to the following criteria:

• Level of progress in solving questions of the

current exercise list. The lower the progress

bar, the greater the possibility that the learner

is experiencing difﬁculties and in need of help.

• Number of attempts per Learning Object. Ex-

ercise lists with large numbers of attempts may

indicate that learners are not sure their answers

are correct.

• learners who made mistakes very similar to

those of other colleagues. The data obtained

in Step 3 of the previous step are used for this

purpose.

Step 2: Identify Questions that Learners had Most

Difﬁculty Solving.

1. Identify the questions that got the most incorrect

answers.

2. Categorize the incorrect questions by types of er-

rors.

Figure 5: Graph and similarity result details.

Step 3: Intervention with Learners.

1. Respond to messages and comments from learn-

ers who have already submitted them.

2. Send messages or comments to groups of learners

with similar difﬁculties on certain questions.

3. Send individual messages or comments to learners

who have a speciﬁc difﬁculty.

4. Monitor learner performance after intervention.

5. Complement the intervention with additional

messages and comments, if necessary.

The evaluation of interaction in learning error me-

diation in previous section made it possible to observe

that FARMA-ALG is effective in its purpose of reme-

dying an erroneous answer through the teacher inter-

vention. Yet, the number of recorded interactions was

low. The method proposed in this section was devel-

oped based on the observations of this study.

5 CONCLUSIONS

For FARMA-ALG evaluation, descriptive and infer-

ential techniques and methods were used. A reduction

CSEDU 2018 - 10th International Conference on Computer Supported Education

364

of 46% in the number of attempts after the learner-

teacher interaction was conﬁrmed. The intervention

of teacher is, therefore, a great contribution so that the

learners can solve their doubts and progress in solving

the problems proposed.

Regarding the effectiveness of FARMA-ALG, it

was veriﬁed that its use promoted an increase in the

number of correct answers by about 7.5%, consider-

ing the performance of learners, from one semester

to the next. However, it was found that even so, the

percentage of incorrect answers remained high, close

to 70% of all answers. Most of the incorrect answers

occurred in questions of the ﬁrst exercise lists, which

indicates possible lack of mastery of the basic prin-

ciples of the programming languages used. This as-

sumption is based on the fact that most of the errors

detected were of compilation and unexpected output.

It was observed that the interaction level, veriﬁed

by the amount of messages and comments sent by the

teachers was extremely low, reaching only 1.93% of

the total responses. Despite the small number of inter-

actions, it was veriﬁed that they were essential for the

learner to progress in solving the exercises. However,

it was found that the high number of attempts on the

most basic questions can be reduced by anticipating

the intervention.

The questions that received comments and mes-

sages from teachers were carefully checked. A large

number of consecutive attempts were made by the

same student for a single question, which indicates

a possible “trial and error” situation. Faced with this

picture, it was observed that the teacher comments al-

ways aimed to guide the learners about where they

were making the mistakes and how to circumvent

them. The teacher intervention was successful in ver-

ifying that the students who received a recommenda-

tion succeeded in about 65% of the total number of

questions in which they had made mistakes.

There was a reduced amount of interactions, due

in large part to the lack of a method that guides the

teacher in his activities with FARMA-ALG. The sim-

pler and easier the use of a platform, the more effec-

tive the results of its adoption. However, even if the

aspects of usability are visible and favorable, there are

others that, although very important, are not explic-

itly stated and documented in FARMA-ALG. As an

example, there is the error mediation feature through

interaction, one of the main differentials of the tool.

The lack of a method for interaction, as seen from

the results of the evaluation of interaction in learn-

ing error mediation, can compromise the breadth and

effectiveness of the proposed interaction actions. For-

malizing and making available a method that works

as a guide in all phases of the teaching process and

considers interaction resources, can beneﬁt teachers

and learners. Therefore, it was proposed that all the

social interactionist characteristics of FARMA-ALG,

can be fully exploited. Considering that this is an on-

going research, there are currently no results on the

effectiveness of the proposed methodology. There is,

therefore, a need to conduct new researches that eval-

uates the outcome of the proposed methodology im-

plementation.

REFERENCES

Chickering, A. W. and Ehrmann, S. C. (1996). Implement-

ing the seven principles: Technology as lever. AAHE

bulletin, 49:3–6.

Garcia, L. F. U. and Direne, A. Metodologia para imple-

mentação de estratégias colaborativas mediadas por

ferramentas de interação síncrona. Tecnologias, So-

ciedade e Conhecimento, 1(1).

Hannaﬁn, M. J. (1992). Emerging technologies, isd, and

learning environments: Critical perspectives. Ed-

ucational Technology Research and Development,

40(1):49–63.

Kutzke, A. R. and Direne, A. (2014). Mediação do erro

na educação: um arcabouço de sistema para a instru-

mentalização de professores e alunos. In Anais do

Simpósio Brasileiro de Informática na Educação, vol-

ume 25, pages 737–746. In Portuguese.

Moura, V. A. B. and Peres, L. (2017). Avaliação do impacto

da retroação na aprendizagem apoiada por uma ferra-

menta educacional. REVISTA BRASILEIRA DE IN-

FORMÁTICA NA EDUCAÇÃO, 25:60. In Portuguese.

Ramos, G. S. (2011). Detecção e remediação de

erros na generalização de conceitos matemáti-

cos por meio de sistemas tutores inteligentes.

http://acervodigital.ufpr.br/handle/1884/25051. Mas-

ter Thesis. PPGINF/UFPR. In Portuguese.

Sforni, M. S. d. F. (2004). Aprendizagem conceitual e or-

ganização do ensino: contribuições da teoria da ativi-

dade. Araraquara: JM Editora. In Portuguese.

Vygotsky, L., Hanfmann, E. E., and Vakar, G. E.

(1962[1934]). Thought and language.

Promoting Mediation in Learning Error on Teaching Algorithms

365