Effects of Learning Style and Training Method on Learning by va23823

VIEWS: 0 PAGES: 40

									  What about Teaching Learners in
  Accordance of Their Aptitude?
  A Study Based on DICE TDD System in
  Learning to Programming


Authors:
Li-Ren Chien, Daniel J. Buehrer, Chin-Yi Yang and
Chyong-Mei Chen
Affiliation:
National Chung Cheng Univeristy, Hsin Kuo High
School, Providnce University
                                                    1
                   Introduction
   What is DICE?
   Training Method: DICE with TDD
   Adaptive Learning
       Consider Individual Difference




                                         2
                     DICE
 An assessment system (grader).
 On-line

 Java RMI based

 Server client in core

 Parse tree based

 For a Programming Language Course now

 Typed Mind Maps as data model

 With Test Driven Development model

 With a adaptive model by Kolb Learning Style
The Architecture of the DICE System
   Server                      Teacher Job
                                                                Problem
             Parse Tree               Grading                  Management
            Management
             C Parser                             TDD Model

                                                                                            XML
            Java Parser              Plagiarism                  Score
                                     Detection                 Management
        Parse Tree
        Displayer
                                 Client Sessions
                                      Chat Room                   Judgment
            Node Editor                                            Thread
                                   User Information
                                                                Authorization
                                                                                    Typed
              Clients                                                               Mind      JDBC
              Monitor                  Command                   Keep Alive
                                         Pool                                       Maps




                                  Server Interface (RMI)




    Client                   Client Interface (RMI)        Keep Alive
                                                                                               DB
       UPnP Functions
                                                                        Chat Room
                                  Server Response

        Teacher Extensions
                                                                                            Text /
                   Client                Problem                                            Excel
                   Monitor                Editor
Auto Grading System: DICE




                            5
       Screenshots (DICE)




                              Figure 2: Grading result



Figure 1: Client of student
              Screenshots (DICE)




                                        Figure 4: Client of instructor




Figure 3: Monitor a particular client
                 Running DICE
   Run in Hsing Kuo High School (Taiwan,
    R.O.C.) over 3 years
     C + Java + DS + Algorithm
     Over 4,000 students via 4 teachers

     Over 400 test cases in 80 teaching unit;

     over 50,000 learner’s source codes be collected
                  Training Method

   Instruction-oriented training
       deductive, programmed, law learner control,
        compete material and feature control (S.D., Devis,
        1989)
       Our implementation: DICE with TDD
   Exploration-oriented training
       inductive, trial and error, high learner control;
        incomplete learning materials and relevant task
        focus (S.D., Devis, 1989)
       Our implementation: DICE with Non-TDD
                                                            9
    Test Driven Development (TDD)
   TDD is a code
    development strategy, one
    always writes a test case
    before adding new code
    (Beck, 2003)
   Extreme programming
   TDD in education since
    2004




                                    10
The TDD model in DICE




                        Γ: Skills




                   Ω: Conceptual
         Training Method: DICE TDD
                  Concepts
              XΩø           XΩ>           XΩm             XΩ╞

        YΓø   Exploration Concept-like    Concept-        Concept
                          (2)             Modular(4)      Instruction
                                                          (7)
        YΓ>   Skill-like    Like          Concept-        Concept-
              (1)           (5)           Modular,        Instruction,
                                          Skill-like(9)   Skill-like (11)
Skill




        YΓm   Skill-        Concept-      Modular         Concept-
              Modular       Like,         (12)            Instruction,
              (3)           Skill-                        Skill-
                            Modular(8)                    Modular(14)

        YΓ╞   Skill-        Concept-      Concept-        Instruction
              Instruction   Like,         Modular,        (15)
              (6)           Skill-        Skill-
                            Instruction   Instruction(13)
                            (10)                                         12
Training Material: an example of DICE
           with Non-TDD
   Prime Cut (the 406th problem of ACM UVA
    online judge system)
        a program that will cut some number of prime numbers from the
         list of prime numbers between 1 and N. It reads in a number N;
         determine the list of prime numbers between 1 and N; and print
         the C*2 prime numbers from the center of the list if there are an
         even number of prime numbers or (C*2)-1 prime numbers from the
         center of the list if there are an odd number of prime numbers in
         the list
     Sample In                             Sample Out
    21 2           21 2: 5 7 11
    18 2           18 2: 3 5 7 11
    18 18          18 18: 1 2 3 5 7 11 13 17
    100 7          100 7: 13 17 19 23 29 31 37 41 43 47 53 59 61 67
                   1 2 3 5 7 11 13 17 19 K=9, C*2-1=3

                                                                         13
Training Material: an example of DICE
              with TDD

             countK()
                    primeM()
                                isPrime()




           getL()
                     isEven()      getR()




                                            14
Examples by ACM Contest Q476, Q477
        Q476         Q477




15
       TDD relevancy
                         Parse Tree

 Q476.C




     DICE
     TDD
       (
     Typed
     Mind
      Map
     Model                                           Test Suite
       )




16           Functions                Concept Maps
     Mind Maps of teaching unit ACM




17
                  Test Suite C1=Q476
        Skill Test Units (Functions that be
         defined by instructor for teaching
         programming skill)
            S={main, allocate2D, readToArray2D,
             readARowToArray, checkUntil,
             checkPoint, checkARow, deAllocate2D}


        Concept Test Units (Other functions)
            O = {scanfRectangle, IsRectangle, isInX,
18
             isInY}
          Learning Style Theories

   Educational psychologists have long been
    of the opinion that different people learn in
    different way
   But, the relevant learner characteristics
    need to be identified (Bostrom, R.P, 1990)
   Learning styles
       an individual’s preferred approach in learning
        process (Riding and Rayner, 1998)

                                                     19
         Kolb’s Learning Style Inventory
   KLSI defined by (Kolb et al. 1979, 2005), includes four
    learning styles:
       Converger
            best at finding practical uses for ideas and theories
            Important for effectiveness in specialist and technology
             careers.
       Diverger
            Who performs better in situations that call for generation of
             ideas and have board interests, tend to specialize in the arts
       Assimilator
            who solves problems by inductive reasoning and ability to
             create theoretical models
            important for effectiveness in information and science careers
       Accommodator
            Their tendency may be to act on “gut” feelings rather than on
             logical analysis
            It is important for effectiveness in action-oriented career such
             as marketing or sale
                                                                              20
              Research Model
                  Learning Style

                    Converger

                   Assimilator


                    Diverger

                  Accommodator

Training Method
                                 H2    Learning
     TDD                              Performance
                          H1
   NonTDD

                                                    21
Result of Data Analysis




                          22
        Discussion of Result (1/2)
   Training gradually is good for programming
    learning
   Considering Confuciu's Theory of "Teaching
    Students in Accordance with Their Aptitude“
       DICE with TDD is worth to carry out for Non-
        Assimilator
       For the Assimilator,
            it is better for them to use Non-TDD training method to
             prevent from suppressing their creativity
       TDD training method has best performance for
        Diverger, secondly on Converger            23
           Discussion of Result(2/2)
   From a practical point of view,
       TDD promotes a climate of discussion between
        learners and their mates.
       In Non-TDD group, the main support of the
        students is the teacher, but we only have an
        instructor in each class
   From a pedagogical perspective,
       TDD scoring mechanism does have a positive
        reinforcement on learners as they can acquire
        scores after every sub function is solved.
                                                        24
                  Future Work
                    Individual Difference
Training Method
                     Learning Style


NonTDD (TDD 0)      Field Dependency




                           …
    TDD 1

                          Logical
    TDD 2

                                    H2       Learning
      …




                                            Performance
                            H1
   TDD 15

                                                          25
Thank You for Your Listening!



                                26
EXPERIMENT AND DATA
ANALYSIS




                      27
             Hypotheses (1/2)

   Hypotheses1: Participants in DICE with
    TDD will score significantly higher on
    learning performance measures than
    participants in DICE with non Non-TDD




                                             28
                  Hypotheses (2/2)
   Hypotheses2: Participants in training method
    with learning style will have different
    influence on learning performance
       Hypotheses2.1: DICE with TDD has a stronger
        positive effect on learning performance for
        Divergers and Accommodator than Assimilator
        and Converger.
       Hypotheses2.2: DICE with Non-TDD has a
        stronger positive effect on learning performance
        for Assimilator than non-Assimilator

                                                      29
                         Sampling
   Random assignment of Sampling

Item       Category          Training Method   Summary

                           Non-TDD      TDD

Learning   Assimilator       62          57        119
Style
           Accommondator     25          30         55

           Converger         58          65        123

           Diverger          18          15         33




                                                         30
                 Training Materials
   Textbook: “C How to Program, 4/e” published by Deitel
    in 2004.
   Training material: from ACM UVA online judge system

       Session1 Introduction of programming language
                background
       Session2 the basic input and out expression
       Session3 Decision making , arithmetic operators,
                relational operator and logical operators
       Session4 Repetition essentials
       Session5 Array
       Session6 Function
                                                            31
                           Procedures
              Content                   Progressing
Introduction of the course        Conducting KLSI to
                                  discriminate learners

Session1~Session4                 Conducting the same
                                  training method


Session 5,6

Problem solving training          Conduct the different
                                  training method to
Test                              students by random
                                  sampling

                                                          32
              Descriptive Statistics(1/3)
Item                Category            Frequency   Percent

Gender              Male                237         71.8%

                    Female              93          28.2%

Training Method     DICE with Non-TDD   163         49.4%

                    DICE with TDD       167         50.6%

Learning Style      Assimilator         119         36.1%

                    Accommondator       55          16.7%

                    Converger           123         37.3%

                    Diverger            33          10%

Classification of   First               34          10.3%
Department
                    Second              155         47%

                    Thitd/Fouth         141         42.7%
                                                              33
           Descriptive Statistics(2/3)
Training     Learning Performance               Learning
Method                                          Attitude

             Mean       Standard    Min   Max   Mean
                        Deviation

Non-TDD      14.25      19.79       0     80    3.00

TDD          21.51      27.14       0     100   3.01

Summary      17.92      24.03       0     100   3.01




                                                           34
              Descriptive Statistics(3/3)
Learning       Mean            Standard            Min          Max
Style                          Deviation

               Non-    TDD     Non-        TDD     Non-   TDD   Non-   TDD
               TDD             TDD                 TDD          TDD

Accommonda     9.47    14.13     18.8      21.96     0     0     68    80
tor

Assimilator    19.08   20.08    22.17      23.73     0     0     68    100


Converger      13.35   24.44    18.56      30.01     0     0     80    100


Diverger       7.18    28.93    11.84      33.78     0     0     36    100


                                                                         35
       Data Analysis Method(1/4)

   Unbalanced two way ANOVA: multiple
    regression which can demonstrate the
    quantitative information
   Regression Model


       Learning Style   Dummy     variable
       Converger        (0,0,0)
       Accommondator    (1,0,0)   klst1
       Assimilator      (0,1,0)   klst2
       Diverger         (0,0,1)   klst3
                                             36
        Data Analysis Method(2/4)
The estimated model is



Variable           Regression Estimator   Estimated        P-value
                   Coefficient            standard error

Intercept          β0         13.1037     2.5213           <0.0001
KLST1              β1         -7.1389     3.7033           0.0548
(Accommondator)
KLST2              β2         5.9931      3.9172           0.1270
(Assimilator)
TM                 β3         11.0978     3.2544           0.0007
(DICE with TDD)
TM x KLST2         β4         -10.1069    5.4179           0.0630
(DICE with
TDD*Assimilator)                                                     37
         Data Analysis Method(3/4)

   Accommondator has poor performance in
    programming
       Accommondator has fewer grades 7.1389 than
        Converger, P-value with 0.0274
   Learners with Diverger have the same range
    with Converger in the two training methods




                                                     38
Data Analysis Method(4/4)




                            39
                   Summary of Result
Item                            Hypotheses                            Result

H1     DICE with TDD will score significantly higher on learning Supported
       performance measures than participants in DICE with
       NonTDD
H2     Participants in training method with learning style will have Supported
       different influence on learning performance.
H2-1 DICE with NonTDD has a stronger positive effect on learning Partial
       performance    for   Divergers   and   Accommodator    than supported
       Assimilator and Converger.
H2-2 DICE with nonTDD has a stronger positive effect on learning No
       performance for Assimilator than non-Assimilator.           supported



                                                                               40
stronger positive effect on learning Partial
       performance    for   Divergers   and   Accommodator    than supported
       Assimilator and Converger.
H2-2 DICE with nonTDD has a stronger positive effect on learning No
       performance for Assimilator than non-Assimilator.           supported



                                                                               40

								
To top