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

2
Help Seeking Model

• What is it?
• Describes students 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

3
Help Seeking Model
We are here!
4
Research question

• Do different patterns in help seeking correlate
  with students attitudes about the cognitive
  tutor?
• We think so, what do you think?

5
Method

• Data from Ryan S. Baker PhD.
• Environment
• Geometry scatter plot.

6
The Analysis

• Learning to work with Excel, Pivot Table.
• Building rules.
• Try step abuse
• 1) Too fast (VARTF)
• If not first attempt 1, AND time lt.10, AND if
  answer 0 then too fast
• If not first attempt 0 And if time lt.7, And
  if answer 0 then too fast
• 2)Guess quickly
• If probability gt.80, and if time lt.7, and if
  answer 1, (then guessing)) gt VARguess
• If errors gt2 then help is need, if help requested
  0, and if answer1) then guessing gtVARguess2
• (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
  doesnt necessarily)

7
Results

• Correlations

VARTF VARguess VARguess2
VARTF 1 .007 .736
VARguess 1 .225
VARguess2 1
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
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.

10
Results
Fig 2. Histogram showing the patterns of the VAR
with the count of hi/low.
11
Results

• Correlation with the attitude data?
• Remains mystery.

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.

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 students attitude?