help seeking model

13
1 Help seeking model By: Geraldine Clarebout (Univ of Leuven, Belgium) Nik Nailah Binti Abdullah (National Institute of Informatics, Japan) Track: DataMining

Upload: amber-drake

Post on 01-Jan-2016

27 views

Category:

Documents


0 download

DESCRIPTION

Help seeking model. By: Geraldine Clarebout (Univ of Leuven, Belgium) Nik Nailah Binti Abdullah (National Institute of Informatics, Japan) Track : DataMining. Help Seeking Model. What is it? Describes student’s ideal help seeking behavior. It determines - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Help seeking model

1

Help seeking model

By: Geraldine Clarebout (Univ of Leuven, Belgium)

Nik Nailah Binti Abdullah (National Institute of Informatics, Japan)

Track: DataMining

Page 2: Help seeking model

2

Help Seeking Model

• What is it?– Describes student’s ideal help seeking

behavior. It determines• what type of action the students should ideally

perform• The number of steps of actions so far on the

step• The number of time students spends answering

Page 3: Help seeking model

3

Help Seeking Model

We are here!

Page 4: Help seeking model

4

Research question

• Do different patterns in “help seeking” correlate with student’s attitudes about the cognitive tutor?– We think so, what do you think?

Page 5: Help seeking model

5

Method

• Data from Ryan S. Baker PhD.

• Environment– Geometry scatter

plot.

Page 6: Help seeking model

6

The Analysis• Learning to work with Excel, Pivot Table.• Building rules.

– Try step abuse1) Too fast (VARTF)If “not first attempt” = 1, AND time <.10, AND if answer = 0; then too fastIf “not first attempt” = 0; And if time <.7, And if answer = 0; then too fast2)Guess quicklyIf probability >.80, and if time <.7, and if answer =1, (then guessing)) =>

VARguessIf errors >2 then help is need, if help requested = 0, ; and if answer=1) then

guessing =>VARguess2(answer can be right or wrong when guessing; but we only see it as a problem

for guessing when the answer is correct, because then a student can think he actually knows the problem; but he doesn’t necessarily)

Page 7: Help seeking model

7

Results

• Correlations

VARTF VARguess VARguess2

VARTF 1 .007 .736**

VARguess 1 .225*

VARguess2 1

Page 8: Help seeking model

8

Results• Descriptives

Min Max Mean SD

VARFT 1 22,13 8,55 4,62

VARguess

0 6,32 2,52 1,39

VARguess2

0 15,18 6,45 2,88

Page 9: Help seeking model

9

Results

• For each of the variables (i.e.,VARTF)– We divided student into two groups: high or

low scores on VARTF,VARguess,VARguess2 median split.

– We defined 9 groups based on this 3 variables.

• E.g., high on the 3 variables or high on 1 variable and low on 1 variable.

Page 10: Help seeking model

10

Results

Total

02468

101214161820

TFl-PGl-PG2l

TFl-PGl-PG2h

TFl-PGh-PG2l

TFl-PGh-PG2h

TFh-PGl-PG2l

TFh-PGl-PG2h

TFh-PGl-PG2l

TFh-PGh-PG2l

TFh-PGh-PG2h

Fig 2. Histogram showing the patterns of the VAR with the count of

hi/low.

Page 11: Help seeking model

11

Results

• Correlation with the attitude data?– Remains mystery….

Page 12: Help seeking model

12

Conclusion and perspectives

• 18% try step abuse.• 17.5% for high on guessing and low on

the other ones.• 17.5% high for too fast and high on

guess2 and low on guess.• 14.4% low on all of the variables.

Page 13: Help seeking model

13

Conclusion and perspectives

• The % does not explain why students has the tendency to try step abuse.

• What is the relationship of try step abuse with the other help seeking factors?

• What is the relationship with student’s attitude?