Week 1: Help-seeking

Led by Jeannie Kim

CSE 191: Inclusive CS Education

Mindfulness

(1 min)

Paper: Factors Influencing Students’ Help-Seeking Behavior while

Programming with Human and Computer Tutors

(Thomas W. Price, Zhongxiu Liu, Veronica Cateté, and Tiffany Barnes. 2017.)

​

https://dl.acm.org/doi/10.1145/3105726.3106179

​

Focus

“What do we know about how and why students seek programming help from computer tutors – or from human tutors?”

“…how and why students actually seek help, not just how they should.”

​

RQ1 Why do novices seek and avoid help when programming?

RQ2 How is the process of help-seeking different when novices are working with human help and computer-based help?

​

How often did/do you seek help in CSE? When? In what ways? (tutoring, office hours…)

​

What are/could be barriers to seeking help?

• things you’ve personally experienced or things you see your students experience

take down notes on next slide

Discussion: Seeking help in CSE

How often did/do you seek help in CSE? When? In what ways?

What are/could be barriers to seeking help?

• things you’ve personally experienced or things you see your students experience

​

How often:

​

Barriers:

• not knowing how tutor system works
• knowing when tutoring hours are
• preparation/pressure pre-going to tutoring
• mental barrier: i should be able to do this by myself
• equating asking for help with not being able to complete something by yourself
• discouraging/disparaging responses from asking for help
• logistically: too many other students asking for help
• easier to ask friends

Quick paper overview: Objectives

factors that influence novice programmers’ help-seeking behavior

how this differs with human and computer tutors

qualitative analysis of 15 students’ interviews, in which they reflect on solving two programming problems: one with a human tutor and one with intelligent, computer-based help.

discuss design implications and hypotheses that arise from these results.

​

​

Quick paper overview: Details

15 undergraduate students at a large U.S. research university; recruited students who had completed or enrolled in an introductory programming course (AP CS, Java, Matlab or Python), but not a more advanced course.

4 females and 11 males, with most students identifying as White (11) or Asian (2), one as Hispanic/Latino and one as Other. Participants’ majors were primarily CS (5) or other engineering �fields (8), and 9 had taken or were enrolled in the introductory Java course for CS majors

Snap! - Block based programming environment

iSnap - computer tutor; an extension that provides hints, possible edits

Quick paper overview: Analysis

Qualitative analysis using Grounded Theory

Open coding on interviews → codes → broader categories

The Three Code Categories:

Things that Impact Help-seeking

Inputs: static factors that exist outside of the problem solving session

• i.e. student’s previous experiences, the student’s existing knowledge and beliefs, and attributes of the programming assignment. �
• tutor (or ITS designer) has no direct control over these factors. unlikely to change in the course of problem solving.
• while they may not influence help-seeking behavior directly, Inputs set a baseline for the Student Mindset.

1

The Three Code Categories:

Things that Impact Help-seeking

​

Student Mindset: how a student interprets their relationship to the programming problem, the tutor and the help the tutor offers.

• i.e. perceptions of the tutor’s trustworthiness, accessibility of help, how stuck the student is
• unlike Inputs, the mindset is dynamic and can change throughout problem solving, especially but not exclusively when the student seeks help
• student’s help-seeking behavior is a response to this mindset

2

The Three Code Categories:

Things that Impact Help-seeking

​

Attributes of Help: qualities of a tutor’s help which determine its effectiveness and perceived value

• i.e. specificity, interpretability, utility
• unlike codes in the Student Mindset category, these attributes can only be experienced by the student after seeking help from the tutor. �
• they affect the outcome of receiving help, possibly leading to changes in the Student Mindset.

3

Time to read the paper!

Sections:

• 3.2 Procedure
• 4.1 Inputs - Andrew and Jeannie
• 4.2 Student Mindset - Jose and Allison
• 4.3 Attributes of Help - Rachel and Mia

(if needed; brief summaries of these at the end)

• 5.1 Hypotheses
• 5.2 Limitations
• 6 Conclusion

​

3.2 Procedure: Notes

4.1 Inputs: Notes

Previous experiences

• human tutors: negative/positive experiences
• negative experience -> avoid human help
• ex. having a program in mind, tutor says “just do it this way”
• not being used to asking for help
• relating computer tutor to other ides, previous programming experience

Expectations of the tutor

• students had high expectations of tutor’s knowledge and experience
• low expectations of computer: led to either being satisfied with results or not using it at all

Independence

• beliefs about working alone: “figuring it out myself”
• asking for help lowers self-efficacy
• programming in particular may lend itself to independent work because “you can just kinda puzzle your way through it and �figure it out on your own”

​

4.2 Student Mindset: Notes

4.3 Attributes of Help: Notes

"Perceptiveness" - how much the tutor/computer help tool knows about *why* you're stuck

​

"Interpretability" - how easy it is to use the help that's provided by the tutor, trust

​

Maybe difference between human and tool help is interactivity

Sometimes useful not to get full help

5.1 Hypotheses: Notes

5.2 Limitations: Notes

6 Conclusion: Notes

Hypotheses

H1: The�e same factors shape students’ help-seeking process with both human and computer tutors.

H2: Help-seeking is not simply a triggered response to di�fficulty; it can be a problem solving strategy.

Limitations

Small dataset

Researcher bias

Game may have been too easy

Snap! interface

No gender discussion: too few females (4) vs. males (11)

• this is a major factor in help-seeking

Conclusions

human tutor seemed more trustworthy, perceptive and interpretable. computer tutor seemed more accessible and less threatening to a student�’s sense of independence.

future work: address aspects of the help-seeking process that we identi�fied but did not discuss here:

• the ways students seek and avoid help
• the types of help the tutor provides
• the interface through which help is mediated
• outcomes of receiving help
• how the students’ perceptions are updated a�er seeking or avoiding help.

​

Discussion

Reactions to the paper?

Anything you related to?

What factors were missed?

What can instructional staff take away from this discussion? How can we improve?