9288 Views (as of 05/2023)
  1. Beverly Woolf
  2. https://www.cs.umass.edu/faculty/directory/woolf_beverly
  3. Research Professor
  4. Presenter’s NSFRESOURCECENTERS
  5. University of Massachusetts Amherst
  1. Danielle Allessio
  2. http://people.umass.edu/~allessio/index.html
  3. Graduate Student College of Education
  4. Presenter’s NSFRESOURCECENTERS
  5. University of Massachusetts Amherst
  1. Ivon Arroyo
  2. http://www.wpi.edu/academics/facultydir/ia2.html
  3. Assistant Professor
  4. Presenter’s NSFRESOURCECENTERS
  5. Worcester Polytechnic Institute

Impact of Adaptive Interventions on Student Affect, Performance, and Learning

NSF Awards: 1324825

2015 (see original presentation & discussion)

Grades 6-8

This research address how an online instructional tutor can respond to student emotion in online contexts. What to do when students are frustrated or anxious, bored or disappointed? We are evaluating the hypothesis that tailored emotional interventions can positively impact students’ emotions, attitude, and performance during learning. Our team has created MathSpring, an intelligent tutor, and has built models to detect student emotion in real-time, as students use the tutor, with and without sensors of physiological activity. We are measuring the impact of pedagogical treatments, such as reports of student progress and providing a choice to students about future curriculum. We intend to identify concrete principles about how a system might respond to student emotion as it occurs and to contribute to a theory of affect in interactive learning environments. In our past work, we integrated gendered and ethnic learning companions (male and female, White, Hispanic and African American), which offer advice and encouragement by talking to students. Learning companions can gesture and train attributions for “success/failure” – for example, the idea that intelligence is malleable, perseverance and practice are needed to learn, making mistakes is an essential part of learning and that temporary failure is not due to a lack of innate ability. In sum, we target students’ emotion in-situ in online learning, are gathering fine-grained data on emotions during learning, and will assess the impact of specific affective interventions on a moment-to-moment basis.

This video has had approximately 691 visits by 588 visitors from 148 unique locations. It has been played 360 times as of 05/2023.
Click to See Activity Worldwide
Map reflects activity with this presentation from the 2015 Teaching & Learning Video Showcase website, as well as the STEM For All Multiplex website.
Based on periodically updated Google Analytics data. This is intended to show usage trends but may not capture all activity from every visitor.
show more
Discussion from the 2015 Teaching & Learning Video Showcase (14 posts)
  • Icon for: Carolina Milesi

    Carolina Milesi

    Facilitator
    Senior Research Scientist
    May 11, 2015 | 03:46 p.m.

    Very cool to present the video from the perspective of the student! Do students select their online tutor or is it assign to them? How have you been able to assess how different students react to online tutors of different gender and race/ethnicity?

  • Icon for: Danielle Allessio

    Danielle Allessio

    Co-Presenter
    Graduate Student College of Education
    May 12, 2015 | 04:45 p.m.

    Hi Carolina, thank you for your question. At the moment students are assigned a learning companion. We have collected data indicating that males prefer the male companion. We are making plans to design additional multi-cultural companions with student input and plan to add functionality that allows students to design aspects of their tutor.

  • Icon for: Allison Hoff

    Allison Hoff

    GIS Technician for GRACE Project, GIS Graduate Student
    May 11, 2015 | 08:46 p.m.

    What is the program tracking to evaluate the students possible emotions? Is it the amount of time since they have been active, how they are doing on similar problems, and how they are doing on problems during that session? This is a great project – I want to try it!

  • Icon for: Danielle Allessio

    Danielle Allessio

    Co-Presenter
    Graduate Student College of Education
    May 12, 2015 | 05:16 p.m.

    Hi Allison, thank you for your question. The program is tracking all of the things you mentioned, how many problems they got wrong/right, how many hints used, amt of time on problems and it also asks the students directly how are you feeling? It asks how frustrated, interested, bored or excited are you on a scale. The system provides the student an intervention based on what they report on how they are feeling. The student maybe directed to the my student progress page, to a new topic or to review a topic. Please feel free to log in as a guest and try out Mathspring.

  • Icon for: Ivon Arroyo

    Ivon Arroyo

    Co-Presenter
    Assistant Professor
    May 14, 2015 | 03:31 p.m.

    Hi Allison, We have been working for several years on creating accurate predictive models of student emotion based on a combination of: a) recent history of student activity (time, hints, correctness); b) baselines of student affective predispositions for math, problem solving and motivational factors such as growth mindset. All of these have shown to be important predictors of student emotion when a student solves a new problem.

  • Icon for: Tamara Moore

    Tamara Moore

    Facilitator
    Associate Professor
    May 12, 2015 | 01:47 a.m.

    This is a very interesting platform.

    Are the tutors primarily providing word problems that focus on applying mathematical concepts? Do you have any open-ended, higher-order thinking problems for students to engage in?

  • Icon for: Ivon Arroyo

    Ivon Arroyo

    Co-Presenter
    Assistant Professor
    May 14, 2015 | 03:33 p.m.

    Hi Tamara,
    We are adding new activities that have other forms of responding as we speak. These involve math problems of the style of PARCC, as well as mini-games and activities that involve more sophisticated narratives and forms of responding.
    —Ivon

  • Icon for: Danielle Allessio

    Danielle Allessio

    Co-Presenter
    Graduate Student College of Education
    May 12, 2015 | 05:22 p.m.

    Hi Tamara, thank you for your question. At the moment the system provides multiple choice questions but we are about to implement short answer questions too. This is an exciting new direction that will provide opportunities for engagement with open-ended, higher-order problems.

  • Icon for: Dacid Lustick

    Dacid Lustick

    Facilitator
    Associate Professor
    May 13, 2015 | 09:09 a.m.

    While watching the video, so many question kept popping up in my head: What is the target audience for Mathspring? How many students have you worked with on this project? How many responses does the tutor have? How long before the interactions become repetitious? How much time on average do students spend with the program? I think my favorite part is the plant as a model for learning and a mechanism for giving feedback to the student. However, I was also wondering in a half joking manner, if the tree ever fails to thrive? What if the student fails to progress? Does the tree wither? Finally, how are teachers involved in using the program with their students?

  • Icon for: Beverly Woolf

    Beverly Woolf

    Lead Presenter
    Research Professor
    May 13, 2015 | 09:33 a.m.

    Dear David,

    Thanks for your comments. The target audience for this tutor includes students in grades 4-12. Currently the tutor covers most of the math topics listed in the Common Core Standards in mathematics for Grades 4-8 and we are developing new problems for grades 8-12.

    Tens of thousands of students in four states have used the tutor.

    The tutor has several response types and numerous responses within each type. It produces hints ( 1-6 hint for each problem); companion comments (2-3 for each effort and emotion level — e.g. comments for a student who has not exerted much effort and comments for a student who appears frustrated); and external activities (linked Internet problems, e.g. from Kahn Academy, that are very similar to the one the student is working on).

    The hints and external activities do not become repetitious. However, the Companion comments would become repetitious after a few hours. Students work on the tutor generally for a few hours. They could also work with the tutor for a few hours/week over a semester.

    Currently the tree does not fail and wither away. It does stay in the same position if the student failed to make progress.

    Students do fail to make progress and their trees do not grow.

    Teachers are engaged in the use of the tutor. They select the CC standards to be presented and help students as they work on the tutor.

    Regards
    Beverly

  • Icon for: Dacid Lustick

    Dacid Lustick

    Facilitator
    Associate Professor
    May 13, 2015 | 02:33 p.m.

    Thanks Beverly. I personnally really connect with the growing tree. In the next version, I would like to see a much more robust plant to grow as a model for my own learning. Maybe the tree could be different for different ideas in mathematics? Or, perhaps there are a dozen different plants any of which a student might get when they work with the program. Just my 2 cents.

  • Icon for: Ivon Arroyo

    Ivon Arroyo

    Co-Presenter
    Assistant Professor
    May 14, 2015 | 03:38 p.m.

    Hi David,
    Actually the tree grows topics out of it! If you use it for a few days with a real account, such as usr: FOWLERSTUDENT and pwd: FOWLERSTUDENT, you will see the different topics growing and eventually blooming. This is a test user, doesn’t really have a lot of activity but you can see how the tree is changing gradually after logins during different days.

    We were considering the idea of withering… We wanted to try it for the student progress page. This way you have to “keep your plants growing happily” until they bloom and give … peppers!

  • William Hersh

    Guest
    May 14, 2015 | 08:40 a.m.

    The video looks terrific. What sort of response are you getting from students with disabilities: Learning, sensory, cognitive, etc.? Are you planning on trying the tutor with any assistive technologies to determine if there are compatibility concerns?

  • Icon for: Beverly Woolf

    Beverly Woolf

    Lead Presenter
    Research Professor
    May 14, 2015 | 09:53 a.m.

    Dear Bill,

    Nice to talk with you. Yes, we have tested MathSpring with students who have an Individual Education Plan in the classroom, primarily those with learning disabilities. We found that these students show increased confidence and reduced frustration when the companion is used. These student also learn more and are more interested in the tutor than the control group of typical students.

    We definitely want to use the tutor with student with sensory and cognitive disabilities.

    We are actively looking for teachers of this populations of students to work with us to evaluate the system in their classroom. Are you able to work with us?

    Regards
    Beverly

  • Further posting is closed as the event has ended.