Catherine

Week 33 (May 1 - May 5):

We got IRB approval, so we can soon put the survey up on mTurk! Hopefully we will get feedback fairly quickly. Because this was finals week, we had no lab meeting, though we did work on finishing up the final CREU report. I will also be graduating college tomorrow, so that is taking up a lot of headspace right now! Once graduation is over and I've moved out of my apartment (Monday), I am going to investigate going to this year's Grace Hopper conference. My long-term goal is to be helpful to my lab-mates should they need my assistance over the summer!

Week 32 (April 24 - 28):

We received feedback on the IRB, and Margaret and Marlena have been working to make sure all the necessary edits are made, and that all the amendments have been submitted. We had one last lab meeting, and it seems like the survey will be up and ready to go fairly soon. The issue of participant payment is still present, but I believe the current estimate is 50 cents per person.

Since the semester is coming to a close and all of my final projects are due this week, I am spending most of my time working on those. Because we are so close to beginning the study, I'm afraid I don't have any specific goals to work on for the project, other than to be available and helpful if my fellow CREU folks need me over the summer.

Week 31 (April 17 - 21):

We've submitted to the IRB, and sent out the pilot survey! Most of this week was again spent on the exact language of the survey and the scenarios we would use. There were a few we got stuck on (namely, how to say that a caravan of human drivers collectively ran over someone), but we have decided on what I think is a good solution. Anne has also worked on creating graphics of each scenario for the survey, which will hopefully help participants visualize the possibly odd-sounding situations.

Week 30 (April 10 - 14):

This week we talked more about how exactly we want to word the survey, and about what exactly we want to measure. We decided to still use the trolley problem and look at group vs. individual agents and human vs. robot/autonomous car actors, but we are not looking at different contexts for this particular survey. We might do another survey looking at the different contexts (driving, medical, etc.) at a different time. The goal for the next week is to submit a proposal for the survey to the IRB and send out a revised pilot survey for members of the lab to look at and comment on.

Week 29 (April 3 - 7):

This week at our lab meeting, we decided we needed to choose which variables to manipulate for the trolley problem factorial study. These variables are:

Human vs. robot actors
Action vs. inaction (in other words, whether we even want to use the trolley problem at all)
Groups as actors vs. solo actors
Context (a driving situation, or a medical situation)

I did some literature review on whether or not context makes a difference in how people respond to the trolley problem. The short answer is yes, so I think it would be an interesting factor to look at, though I know it would add another layer of complexity to the study (and to future data analysis). My goal for the next week is to continue doing lit review, and my more long-term goal is to help revise the language of the survey and show it to my friends so we get a feel for what people are actually picturing when they read it. Of course, the long, long-term goal is to finish the survey and get it on mTurk.

Week 28 (March 27 - 31):

Now that the paper's done, I am spending more time acquainting myself with Margaret's factorial survey. We all met on Monday to discuss changes that different people suggested, though at the end of the meeting I think we had more questions than answers. We're having difficulty finding the right phrasing for all of the scenarios, especially the scenarios involving group decision-making. It seems we need to re-evaluate what exactly we want to study with this project--do we want to focus more on groups? On the moral dilemma of the trolley problem (inaction vs. action)? On different contexts in which the trolley problem might occur?

This week I learned that the original plan for the factorial survey may not be the one we want to go with in the long run, but it's better to find this out earlier rather than later. Our goal for the next week (and likely the coming weeks) is to decide on which factors we want to focus, and to come up with the right phrasing for each of those scenarios.

Week 27 (March 20 - 24):

This week was spent finishing up the paper and making it shorter. I met up with my grad mentor on Saturday to seriously edit the sections for clarity and brevity. The final paper was finished and submitted Sunday night, so it's now out of my hands until I hear from the reviewers at the conference. After several weeks of saying, "The paper's almost done!" it's finally done. I hope that for the next paper I work on, I organize my time better so I do not creep up so close to the deadline. I had very little experience working on a paper before, and this semester was the first time I was not assigned a particular section to work on by someone else--my assigned section was the entire paper. I am extremely grateful to my mentors for helping me so much on writing and editing, especially so close to the deadline.

Week 26 (March 13 - 17):

Spring Break! Since the deadline for the paper has been extended and because I'm still waiting to get feedback on my previous draft, I am taking a bit of a break this week. I have done some edits on my paper, and I will send those to my mentors in the next few days. I have almost completed my final revisions on the paper, though it is still too long, so the biggest focus now is getting it down from seven pages to six. My goal for next week is to be finished with the paper and ready to submit!

Week 25 (March 6 - March 10):

The deadline for the paper has been extended once again, this time to March 26. I am almost done, and am just taking care of finishing touches at this point. I am having some people look at my paper and am mostly just waiting for edits at this point. My goal for the next couple weeks is still to finish the paper. Since all I'm doing are edits for clarity and length at this point, that's the only goal I can really have.

This week has been spent learning about, well, editing a paper, mainly by making it shorter. I tend to be a wordy writer, so eliminating words I don't need has been a bit of a personal struggle. I suppose it is a bit easier to make less from more, rather than the other way around.

Week 24 (February 27 - March 3):

Like the last few weeks, I am still working on editing the paper. I am focusing more on just editing for clarity and length now, since all the "meat" of the paper (results, discussion) is done. Some extra background wouldn't hurt, so I am combing through my lit review to add the relevant information. My goal for the next week is to finish the lit review and the edits, and submit the paper on March 6th!

Week 23 (February 20 - 24):

I am still working on editing the paper. All the important sections are in, but it still needs some polishing and re-writes for clarity. I've taken some of my mentors' suggestions (I'm still working on the other ones) and I'm trying to make the paper more cohesive as a whole. I met virtually with my grad mentor this week and will likely do so a few more times before my paper is turned in.

I am learning a lot about how to write these types of papers, namely, that one must remember that future readers will know nothing about the experiment, so clear and precise language is imperative. Learning to say something in fewer words (rather than more) is a difficult lesson, but one I needed to learn.

My goal for the next week is to finish the paper! The due date is rapidly approaching, but I believe I can make it so long as I stay focused.

Week 22 (February 13 -17):

Good news this week! The deadline to submit my paper has been moved to March 6, instead of February 19! This gives me more time to polish it and make sure all the parts are in the right places.

I completed a draft of the entire paper (minus the abstract) and finished my data analysis. It seems I somehow forget to analyze the one variable I was most interested in--whether or not participants turned the robot off! Since my other results weren't particularly exciting, I had assumed this one would turn an insignificant result. But it did! It seems participants in the Ingroup/Low-authority condition were the least-likely to turn the robot off after all! This is exciting, and it gives me more to write about in the Discussion section of my paper. I will be polishing the draft and taking my mentors' suggestions soon.

My goal for the next week is to work on implementing my mentors' suggestions and polishing the paper. I will have it done by March 6 (and hopefully before that!).

Week 21 (February 6 -10):

I have finished my first versions of the Intro, Background, and Methods sections of the paper, and am mostly done with data analysis. I also spoke with my graduate mentor about how to report my experimental results, and about what literature I could be looking at to add to my paper. My goal for the next week is to finish the remaining sections of my paper (Results, Discussion, Conclusion, and of course Acknowledgments and References). Once these sections are done, I can write my abstract. I also need to send out my completed draft of the paper to my grad mentor and faculty mentor, to see what suggestions they have.

Week 20 (January 30 - February 3):

I am still finishing up my data analysis, but I have also begun to work on writing the Introduction, Background, and Method sections of the paper. I will need to do some more literature review for the Background, but I have a few good papers to reference in the meantime. My goal is to finish at least these three sections of the paper and finish my data analysis this week, though I do have a lot of homework for my classes right now, so I may not meet my own deadline. I do feel have made a decent start, though.

Week 19 (January 23 - 27):

The data analysis is coming along a little better now. After running some within-subjects tests, I am beginning to get significant results! For example, it seems in that in conditions where subjects had an in-group robot and a low-authority researcher, they were less likely to rate the researcher as "knowledgeable" on a scale. This is all just looking at my survey data, which isn't as interesting as the behavioral data, but it is a start. My goal is to finish running within-subjects tests and to finish analyzing data period, hopefully be the end of the next week. I plan to meet with my graduate mentor again one or two times this week (always virtually, as she is currently doing work in Japan) to have her look at my data and give her input.

Week 18 (January 16 - 20):

I am still working on data analysis. Most of the tests I've run so far (all between-subjects) have not produced any statistically significant results, though I will meet virtually with my graduate student mentor soon and ask for her input. My goal remains the same, to finish data analysis and, once that's done, write up a draft for my paper very quickly. I believe I will be able to write quickly once my analysis is done, but until then, more analysis is all I can do. Luckily, the tests I did run indicate that there was some difference between the four conditions, though it may not necessarily be a difference that supports my hypotheses.

Week 17 (January 9 - 13):

After coming back to school from Winter Break, I began working on analyzing the data from the experiment. It's been a while since I've done any sort of data analysis using SPSS, so it's been a bit tricky learning (and re-learning) how to use it. My goal for the next few weeks is to finish analyzing all the data (both the survey data and participants' behavioral data) and to begin work on my submission for the IEEE RO-MAN 2017 conference, due February 19. If I am unable to make the deadline, then I will have to submit my paper to a different conference, though for now I am choosing to be optimistic and work toward the early deadline.

Week 16 (December 5 - 9):

I ran the last participants, bringing the total to 101! The next steps are to organize and analyze the data. I have done some of the organizing, putting some of the data into a big Excel spreadsheet, but I haven't been able to finish it because of my final exams and projects. I will work on it some more over the coming winter break, and next semester I will begin writing my paper.

Week 15 (November 28 - December 2)

Ran more participants! We're almost done! Our grand total is 99, and because that number bugs me, I'm going to run another participant next week (though I set up two time slots, just in case one of the participants doesn't show up). I'm very excited to finally look at all the data I collected from the experiment, though any data we get from the videos will likely have to wait until next semester, since I have not yet decided on the exact coding scheme we will use. One of my fellow undergrads on the project found a scheme that worked well for one of her experiments, but I think we are looking for slightly different things. I'll have to think more about how I want to code the 80-some videos we recorded...

Week 14 (November 21 - 25):

Thanksgiving Break! Since I am not on campus, I actually did not run participants this week! Final project deadlines for my classes are approaching, so I did not do much work on the project, but I will be done running participants very very soon.

Week 13 (November 14 - 18):

We are still running participants, but we're almost done! It still looks to me like the low-authority condition is having an effect on how often participants agree/disagree with the robot, however I predict this effect will be rather small, unfortunately. There are a few instances of participants really agreeing or really disagreeing with the robot, but for the most part people seem to disagree most of the time and agree...sometimes. I had hoped to get a little more variety across conditions for this measure, but of course, I won't know for sure until I analyze the data.

Week 12 (November 7 -11):

You'll never guess what I did this week! We ran a grand total of 17 participants, bringing our total up to 67! Wow! It looks like I'll have plenty of data to work with next semester. It would be great if we could get 90 or even 100 participants by the end of the semester, but I don't want to get cocky (and I don't want the other researchers to feel overworked, especially since they're both working on other projects).

I began this study asking, how far would one have to go to get people to obey a robot's orders over those of a human? Glancing at the data in our spreadsheet, that answer already seems to be, "farther than what we've tried." My sort-of hope for the study was that, in the Low-Authority/In group condition, people would unplug the robot less often than in the other conditions, but we've only had a very small handful of people do this at all. I doubt a real analysis of the data would show these few instances to be significant. Still, there's about 20 participants left to run, and of course I am collecting lots of other data, too. And I haven't even looked at the surveys, though I hope I don't have to use them as a crutch!

Week 11 (October 31 - November 4):

Still running participants! We didn't run as many this week since Margaret was away at a conference in Kansas City, but we're still ahead of schedule.

Glancing at the data, it looks to me like the Low-Authority condition is having some effect on how often participants agree or disagree with the robot's suggestions, but not necessarily on whether/when they unplug it. I wonder if being an in-group member somehow gives the robot more authority or makes it more persuasive, but there is still a "line" that participants won't cross, in this case, not turning the robot off. I can see that potentially being good in real-world situations, as I think it would mean people are willing to defer to robots' judgement as long as said robots don't go "too far." I say this all with some hesitation, however, as I am just glancing at the data in the spreadsheet, and am doubtless biased toward my hypotheses being correct. I don't want to do too much theorizing before the study's even over!

Week 10 (October 24 -28):

Yet another week of running participants. We ran about 15 people this week! This means we are slightly ahead of schedule. While I could begin analyzing data early, I think I will be patient and try and get some extra participants the week after our Thanksgiving break. There were a few hiccups during the first couple weeks of running participants, meaning that a few people may have weird data that may need to be thrown out (if the robot didn't speak, if a researcher wasn't super comfortable with the script, etc.). This is good--if the data show any significant differences, it it is more likely to be caused by our manipulation, and not some sort of accidental variable. I am excited to eventually look at the data, though.

Week 9 (October 17 - 21):

This week was spent running participants, as will every week after this one for the rest of the semester. Everything is going well. As usual, our group met up to discuss how the studies were going and to voice any questions or concerns, but I had little to say since the authority study is going smoothly. We were able to recruit eleven participants this week!

Judging from the few participants that I have already run, it seems that people are not more reluctant to turn the robot off in the low-authority condition, as I had hypothesized. I wonder if this is because the difference between the conditions is not having an effect, or if this is more due to participants being confused in the high-authority condition, since the researcher never comes back to reiterate the importance of turning the robot off. It's difficult to say, however, since I have run fewer than twenty participants, and this is only looking at one data point. We'll have to wait and see.

Week 8 (October 10 -14):

The authority study is officially underway! I've spent the majority of my time on the study running participants. We will need at least eighty by the end of the semester, with twenty participants for each condition (ideally, we would have even more, but I do not want the other students working on the project to spend all of their research hours running participants). We ran a decent amount this week, though I would like to raise our average to about ten. This is only our second week, however, so it takes some time for all the wrinkles to be sorted out.

At one point during the study, the robot we were using stopped speaking when it was supposed to. After troubleshooting, it seems there is some sort of hardware issue. I plan to look at it more later, but in the meantime I was able to replace the malfunctioning bot with one that works. Other than that tiny hiccup the study is going smoothly.

Week 7 (October 3 - 7):

This week we started officially running the experiment! We began setting up time slots for the study and recruiting participants. All of the bugs in the robot's program have been fixed, and we're ready to roll!

As of this writing, we've had two participants come in whose data we will be able to use. We treated the first few trials as "pilot studies," so we could work out the kinks in our methods and write better, more-detailed instructions for the researchers to follow. It took a while, but now the real study is finally beginning! Everything seems to be going well, though we will have a meeting at the beginning of next week for all the researchers to meet up and discuss any problems or concerns they may have. Because this study requires a bit of acting on the part of the researchers, it is important that we keep our methods consistent as the semester continues. It's still too soon to tell if there's any issues with how we're doing things, but for now, I think we're off to a good (if a bit late) start!

Week 6 (September 26 - 30):

Met with CREU collaborators
Practiced running the Authority Study in its two conditions
Fixed the bugs in the robot's program
Typed up some instructions for researchers to follow when running the experiment (myself included)

Week 5 (September 19 - 23):

Met with CREU collaborators
Finished programming the robot!
Practiced a "pilot study" with the newly-programmed robot and identified bugs that needed to be fixed

Week 4 (September 12 - 16):

Met with CREU collaborators
Continued programming robot to be used for the Authority Study

Week 3 (September 5 - 9):

Met with CREU collaborators
Began programming robot to be used for the Authority Study

Week 1 & 2 (August 22 - September 2):

Met with fellow CREU collaborators and discussed the studies that we have planned
Worked on personal blurb for our website (this one!)