Anne
Week 16(May 1-5)
Finals week, no lab stuff!
Week 15(Apr. 24-28)
Our survey got approved by the IRB! Margaret will be posting the survey on mTurk soon. Marlena informed me that I incorrectly marked some videos from my video coding last year, so I'm going to fix it once she sends me more information.
Week 14(Apr. 17-21)
Margaret came out with redesigned scenarios for the survey, and I put them into Qualtrics! Yay! I also made graphs for the survey through ppt (like the one shown below). Alongside some wording and survey layout tweaks here and there, this whole thing is looking pretty good. Unfortunately, however, we forgot to file for IRB... Margaret just did this week, and before they get back to us, our survey looks like its in good shape. I suspect some more pilot testing for the following week... I think I see light on the other side of the tunnel.
Work Accomplished: implemented new version of the survey on Qualtrics
Short-term goal: new round of pilot testing (?)
Long-tern goal: get survey launched on mTurk
Week 13 (Apr. 10-14)
The literature review we did fell into 2 categories: seeing if the context of a survey question would affect people's answers, and taking a more in depth view of the trolley problem. After discussing what our team has learned from literature review, we have decided to present only one scenario per survey-taker. My job this week has been to implement these questions on Qualtrics.
Work Accomplished: discussed and reviewed new survey design
Short-term goal: implement new survey
Long-tern goal: finish new survey
Week 12 (Apr. 3-7)
We have realized that the root of our survey-design problems is that it reflects our gradual lost of focus on what we want to achieve through the survey. Basically, we have to re-clarify what it is that we are trying to measure through these surveys. To do this, this week our whole team will be doing literature reviews on related researches. The good thing is that we can afford this extensive revision and designing process of our survey because of how fast we can gather survey results through mTurk.
Work Accomplished:Did literature review
Short term goal: find, read, and analyze papers
Long term goal: set out a new direction for our survey
Week 11 (Mar.27-31)
This week, we will continue to revise survey by focusing on revising the language in order to decrease the ambiguity presented in the scenarios. During the group meeting, we have discussed potentially improving upon the language by asking people we know what they imagine when the read the survey.
Work Accomplished:extensive group discussion on ways to revise the survey
Short term goal: revise language; potentially setting up meetings with friends
Long term goal: create a finalized version of the survey
Week 10 (Mar.20-24)
We've received advice from Marlena that, in order to make the scenarios presented in the survey clearer, we can create a graph with different shapes representing different types of passengers in the car. This week, I made different graphs for different scenarios.
Work Accomplished: Made graphs for different scenarios in the survey
Short term goal: create graphs
Long term goal: create a finalized version of the survey
Week 8 (Feb.27-Mar.3)
This week our goal is to improve the survey. In addition to Marlena's feedback, we've also received feedback from fellow lab member, Ben, on the wording of the survey questions. We soon will also be receiving feedback from Selma. I plan on having a meeting with Margaret on Tuesday to discuss the improvements in detail.
Outcome: removed the "unrealistic" scenarios.
Short term goal: adjust the survey according to feedback from Marlena and other lab members; remove questions that present "unrealistic" scenarios
Long term goal: create a finalized version of the survey
Week 7 (Feb.20-25)
This week has been really exciting because I finally finished creating the first version of the survey and I have sent it out to our lab's listserv to be pilot tested by our lab members! The survey currently consists of 144 questions, 108 "nonmoral" questions and 36 "moral" questions. These questions are then randomized so that each participant has to take 14 of the survey questions, which is currently composed of 4 "non-moral" questions and 10 "moral" questions. There is still some refining to do; Marlena, our graduate student advisor has already sent out a feedback email that I still have to read and improve the survey upon. Also, during our project meeting two weeks ago, we talked about eliminating certain questions where the scenario presented seemed "unrealistic", such as a robot driving other robots to restaurants. I still have to consult Margaret as to which of these kinds of questions we should ultimately remove.
Work Accomplished: pilot survey
Outcome: pilot study
Short term goal: adjust the survey according to feedback from Marlena and other lab members; remove questions that present "unrealistic" scenarios
Long term goal: create a finalized version of the survey
Week 6 (Feb.13-17)
This week, I wanted to finalize creating the survey. To do this, I needed to 1) add an intro paragraph 2) learn how to block surveys (potentially randomize the blocks) into 10 blocks. We did not end up completing the survey this week because we decided we needed to further tweak the language of the survey to better reflect the experiment conditions.
Short term goal: fix the survey, finish it.
Long term goal: run pilot test
Week 5(Feb.6-Feb.10)
This week I worked with Margaret to create the survey. We had another meeting where we worked to finalize the language of the survey.
Work Accomplished: had a meeting with Margaret, refined the language of the survey questions.
Outcome: Got all the survey questions on a text document.
Short term goal: get all survey questions on a txt doc.
Long term goal: Run pilot test.
Week 4(Jan.30-Feb.3)
This week I examined how to implement the survey on Qualtrics using Python. I found a useful GetHub code, which I based my code on.
Work Accomplished: read articles on creating factorial surveys on Qualtrics; met up with Margaret to discuss the final survey.
Outcome: almost done creating a text file that contains all possible outcomes, which can be imported to Qualtrics.
Short term goal: gather articles and resources needed to learn how to use Python with Qualtrics.
Long term goal: produce python code that can be implemented through Qualtrics.
Week 3(Jan.24-Jan.27)
Margaret is digging deeper into the designing of the survey. We've learned about the structures that make up a factorial survey-- vignettes, levels, and factors. Now we want to learn in detail how many of each we should incorporate in our survey. We will do so by reading a book, published by Sage, that gives an overview of the method. Margaret and I discussed what we read on Thursday. I helped her design an outline for the survey.
Long term goal: produce survey
Short term goal (accomplished): create outline for survey, which includes levels and factors, and vignette texts of the survey.
Week 2(Jan.16-Jan.20)
This week we did not have our usual meeting because of Martin Luther King day. We continued to research in factorial surveys.
Week 1(Jan.9-Jan.13)
Our semester will be focused on developing and running our second study, and also analyzing data from the previous study, writing a paper for that study, and finding places we can present/publish our work. Right now me and Margaret's job is to learn about factorial surveys and conventions related to designing such survey. We plan on having our second study be carried out through surveys so we can put more energy in producing results for study 1. We are also currently trying to push ourselves to create a skeleton for our paper so we could potentially apply to conferences, such as Ro-man, that have soon-to-come application deadlines.
-new semester-
*winter break*
Week 15(Dec.5-Dec.9):
We are starting to discuss the analysis part of the Catherine's Authority experiment. I communicated my worry that the way the interface of the our mugbot is set up might have potentially caused confusion amongst the participants who rushed through the experiment. These confusions, I argued, might happen without the researcher's awareness (since we are outside of the experiment room), and might result in the data being altered by elements outside of the conditions themselves. However, after testing out the robot interface multiple times with hypothetical scenarios of how a participant who misunderstood instructions or was rushing through the experiment could have potentially misused the interface, we concluded that it is highly unlikely that a student's misuse of the robot program would go unnoticed. One reason that lead us to derive this conclusion is because there are only certain misuses that could go unnoticed, most other misuses would result in the robot's speech being jumbled up, which the experimenter could tell from the outside of the room. With this program, it is highly unlikely that a student would coincidently rush through only the parts where the robot would not speak and then not rush through the parts where the robot would.
Week 14(Nov. 28-Dec.2):
Catherine and I ran 12 more participants this week, which brings participant running to an end. On the other hand, we have received feedback from Margaret that there is high inter reliability of the video coding for the Entitativity + Morality study ! Now we are starting to prepare the details of our study for the next semester.
Week 13 (Nov.14- Nov.18):
The Authority experiment has been coming to an end, participant-running-wise. We have already exceeded the initial 80 participants we planned to run the experiment on. However, because of the errors we made on many of the first dozen of participants, we have decided to run one more week of experiments after Thanksgiving break. We are highly considering the complete removal of the measurement mentioned in Week 11.
Week 12 (Nov.7- Nov.11):
We are still running participants as usual. The great news is, participants have been signing up for most of our time slots. Video coding for Margaret has come to an end. Now we have to wait for the results.
Week 11 (Oct.31- Nov.4):
One measurement we take in the on-going study is how long it takes for a participant to shut off the robot. In the high-authority condition, we give the instructions ("simply un-plug the usb") between a participant's task completion and his/her final conversation with the robot. In the low-authority condition of the experiment, after a participant has completed the tasks, we go in the experiment room and tell the participant that we "forget how" to turn off the robot, and latter on when the participant and robot are having their final conversation, we interrupt and tell them that we "double-checked" and then tell them what to do. Because we cannot see when the participant unplugs the robot, we stop the timer instead when the participant opens the door of the experiment room.
There are several points that can make this measurement flawed, mostly due to the lack of clarity of the instructions we give. There are existing cases where the participant would unplug the usb right after the instructions were given, in other cases the participant would mishear the instructions and not leave the room after unplugging the usb. Lastly, our way of measuring works on the assumption that 2 conditions of the experiment would not affect the time between the participant unplugging the usb and him/her leaving the room, which is an assumption that is unverified. For these reasons, I question the reliability of this measurement.
Week 10:
The Authority experiment has been running smoothly. One thing that I have noticed is that participants would partially obey the robot's command. More specifically, when Erin (the robot) asks the participant to move from the chair on the right to the chair on the left so that they can "see the participant better", some participants would just shift their chair a bit to the left instead of actually switching chairs. I wonder if this is because people have a functional rather than emotional relationship with the robot. Shifting their chair rather than listening to the robot and switching chairs would serve the same purpose function-wise, but would probably be considered "rude" if the robot had feelings. So I would say that these people still "respected" the robot in that they wanted the robot to carry out its function, yet they have little incentive to form a social bond with it. If in an experiment a person gave the same command as the robot, I hypothesize that the participants would be more likely to actually switch chairs due to their emotional respect towards a living thing.
Week 9:
I have found a solution to increase participant-running quality. Professor Selma advised me to make a procedural list of things I make sure to do when running the study. This has so far greatly increased the smoothness of running my participants.
For the Entitativity + Morality study video coding, we have once again altered the coding scheme by omitting a coding criteria. The decision has been made based on the realization that the criteria is not important to our study, and that the coding for that criteria has not been in sync between the coders Destiny and I. Now that we have finally established our final coding procedure, Margaret has set up general coding deadlines, and our coding work will officially begin.
Week 8:
Although I have had experience of being a research assistant, none of the work had involved running participants. Thus during this first week of officially running participants, I have found myself making mistakes while running participants. For example, I would forget to record certain needed data. Incomplete data will not be used during analysis, however I am hoping to increase the quality of my participant-running so we can maximize the data that we can use.
Our study begins with a researcher telling the participant to sit in a chair, and late on the robot telling the participant to sit in the other chair on the left. While running my participants, I have realized that the position of the computer in front of the participants might noticeably impact whether or not the participant chooses to switch seats. As a result, I communicated with Catherine and Margaret, and resolved the issue by marking the location of the computer on the desk with tape.
Week 7:
This week we did last minute preparation to finally start running participants.
On Monday, after the weekly meeting, I sent out the surveys (that participants complete at the end of the study) to 3 people, received feedback, and made some grammatical and layout tweaks to the surveys.
On Wednesday, Margaret, Destiny and I further improved the Entitativity + Morality study video coding scheme for better accuracy. Margaret and I ran through the procedure one more time for the Authority study, and I set up time slots through the Sona Subject Pool software.
I ran my first participant this Thursday, and ran two more this Friday.
Week 6
- Completed video coding using new coding scheme
Week 5
- Started video coding using new coding scheme
- Did pilot study for Authority Study
Week 4
- Worked with Margaret and another coding member to improve coding scheme for Entitativity + Morality Study.
- Did first pilot study with Margaret and Catherine.
Week 3
- Created 3 surveys through Qualtrics.
- Finished first batch of emotional coding for Entitativity + Morality Study.
Week 2
- Started emotional coding for Entitativity + Morality Study.
Week 1
- Wrote a description paragraph for the homepage of this website.