Gentle reminder for MIT courses: A way to think about your time on the class is the total credit hours. This is a 12-hour class, which means you should on average spend 12 hours/week on it. For this class, 2 hours/week are "in" class, and the other 10 are work you should invest (on average per week) outside of class. In order to keep from getting super-crunched at the end of the term, we urge you to put in 10 hours on your project this week, trying to get it as far as possible this week. Also, most of you need to submit something to COUHES: if you're doing a study collecting new data, then you're not exempt until COUHES responds to you and agrees that you are.
This week, we list the homework in two parts and give you two weeks to do the non-project reading.
In the first part, we want you to show us progress on your project, especially finalizing your COUHES and study design, and a presentation for our class. The second part is a preview of what we'll do in two weeks, which you can start early if you have time this week after moving your project forward. Next week we'll add some questions/debate topics around the readings.
Due Tuesday Oct 13 10am: Submit copies of your COUHES materials (or COUHES' proof of exemption) to Roz. If you are conducting a study that is going through COUHES, make sure you also download and prepare a Subject Consent form and your proposed recruiting materials (e.g. emails or posters) and plan. If you're unsure if you're exempt or not (sometimes it's not clear on their website), then reach out to COUHES directly (phone/email) and explain and ask. If you are 100.0% sure you are exempt, and Roz has agreed with you on that, then describe what other work you did on your project to move it forward. If it's blocked due to waiting on data, or other reasons, feel free to start the other part of the homework due in 2 weeks.
Due Wednesday Oct 14, 10am: Prepare slides that you will present in class. These should address the likely-final version of your class project plan. Each *project* will have 15 minutes total. (Even if it has only one person on the project team.) Here are some guidelines:
1. For your first slide, give a one-sentence statement of your project's main question/goal.
2. Give us details on the human-affective components. If you are designing a study experience for human study participants, please teach us that part in detail: Walk us through what your participants are going to see/do/be evaluated on. We should be able to imagine the complete experience as if we were a participant in the study. Show a diagram describing what steps they will experience (and if you have different control groups, what each group experiences.) Label the estimated times for each step. What do you plan to pay people for participating, or what will they get out of the experience? Optimizing experimental designs is hard work and let's help each other. If you are NOT running a study, but using existing human data, then walk us through the background details of how your data were collected, what the participants' experiences might have been like, so that we all understand what conditions might have influenced the types of affective expressions in the dataset. Clarify how much data you have for each affective or study condition.
3. Finally, summarize what *specific* kinds of answers/insights you think your planned study/analyses are likely to give.
4. Before 10am, put your slides in a new Panel you make on Miro (same link as before, and put your "Names(s): Title" at the top of the panel and use big font. Note: We will NOT do live commenting like last time, as we think it was a bit distracting to the presentations. However, people will be invited to revisit your slides to add any thoughts afterwards. You will present in class only from your slides in zoom.
READING and WRITING (additional work on these will be due : 10am Tues Oct 20 - 2 weeks away)
(1) Many of you mentioned in the previous homework due Sep 29 that you thought there are cases where technology should monitor only the affective information that people perceive. Several studies may add a surprising twist to this:
a) Emotion contagion can happen when you "catch" the affect of another, even when they don't express their emotions in a clearly visible/audible way -- these findings suggests that we can possibly perceive more than we think. Please read these:
Hatfield, Elaine, John T. Cacioppo, and Richard L. Rapson. "Emotional contagion." Current directions in psychological science 2, no. 3 (1993): 96-100.
Kramer, Adam DI, Jamie E. Guillory, and Jeffrey T. Hancock. "Experimental evidence of massive-scale emotional contagion through social networks." Proceedings of the National Academy of Sciences 111, no. 24 (2014): 8788-8790.
b) We humans may also see much more from each other's faces than facial expressions -- perhaps we see enough blood flow and facial color information to tip the balance. Please skim this for the main idea:
Benitez-Quiroz, Carlos F., Ramprakash Srinivasan, and Aleix M. Martinez. "Facial color is an efficient mechanism to visually transmit emotion." Proceedings of the National Academy of Sciences 115.14 (2018): 3581-3586.
We will give out questions/debate topics on these in a week, but for now just become familiar with the articles.
Yes, the Facebook study above is "The big controversial one" that led to a worldwide outcry against their practice.
(2) We're going to give you an introduction to some of the capabilities related to interpreting emotion or "sentiment" from text and use of emojis. Here's a Media Lab publication that moved the work forward significantly.
Felbo, B., Mislove, A., Søgaard, A., Rahwan, I., & Lehmann, S. (2017). Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm. arXiv preprint arXiv:1708.00524.
(3) Extra credit/optional: If you're interested to spend more time revising your prior "pseudo-legislative report" from last week (the one you and your partner did) and getting more input from Gretchen, let me know and we can arrange something there.