Eleanor Humphrey's Blog
Check this page for weekly summaries of my individual work. My colleagues' pages are linked right below, as well as a link back to the team blog.
This week the team met for the first time as a whole, and discussed our plans going forward. I worked on completing the first several sections of ToBI tutorial we are using to familiarizing ourselves with the software. I also spent some time familiarizing myself with using the Praat software. We also each, individually, practiced rapid transcription and are planning on comparing and discussing our results at our next meeting.
During the first part of the week I worked on completing several more sections of the ToBi tutorial. I also spent some time learning how to use Audacity to chop up sound files in order to make them easier to label. After getting far enough in the tutorial to have learned enough of the labeling conventions I began labeling the chopped files in Praat using the conventions learned in the Tobi tutorial.
This week the entire group met with Dr. Alejna Brugos, who led some exercises, and later a discussion, on ToBI labeling. Having the entire group do the exercises together was very helpful as it allowed us to get feedback from the rest of the group. During the rest of the week I finished up the remaining sections in the ToBI tutorial and continued to practicing labeling.
This week I began labeling the sound files collected by last year's team for the first time. Natalie, Michelle and I decided to work on some of this labeling together and compaire results. This was extremely helpful because, while we mostly agreed on labels, it gave us the opportunity to discuss the ones we didn't agree on and get options on sections we were having trouble with.
This week the team met as a whole and went over the labels we had chosen. It was extremely helpful to be able to do this with the entire group because we not only had the professors there to answer any questions, but we also had another group of students with whom to compare our results. During the rest of the week I continued to label sound files. I was unfortunately not able to meet with Michelle and Natalie this week due to conflicting schedules, but look forward to comparing results with them and the other students at tomorrow's meeting.
This week we met as a group and discussed some of the labels we had disagreed on. There was more disagreement this week than in previous weeks. Meeting was extremely helpful because it allowed us to discuss the struggles we have with a more difficult file. I also chose to go back over some sections of the tutorial to hopefully become a stronger labeler.
This week I completed the labeling for the P20 files. In previous weeks, I had chosen to break up my labeling sessions into smaller sessions, labeling for no longer than an hour at a time. This week, I decided to work for a longer period of time and faced some unexpected difficulties. I found that, the longer I labeled the more difficult the labeling became. I think next week I'll go back to labeling for shorter periods of time.
This week we moved onto P19. Natalie, Michelle and I met to label together. There was some disagreement, but less than other weeks. I'm looking forward to finding out if the Wellesley students had similar results at our team meeting tomorrow.
This week we met and continued going over the labels we disagreed on. We also decided that it might speed things up if we tried looking over the labels we disagreed on before our meetings. So after collecting all our labels, Michelle sent the entire group the labels from two files that we disagreed on. I worked on looking over my labels and comparing them with those of the other group members.
This week Michelle and I met began working together to figure out how to best display and format the consensus labels the group has been working on up to this point. Michelle and I both had a lot of fun doing this. I was really excited to have the opportunity to utilize my knowledge of R on organizing this data, with later analysis in mind. We came up with several different possible options and are looking forward to sharing them with the group when we next meet.
This week I worked on manually placing our consensus labels into excel spreadsheets, using the format Michelle and I previously discussed. It was difficult at first figuring out which labels correspond with each other and it took be some time to figure out the labeling convention for the file names. I was eventually able to figure the system out and the work was relatively straightforward from there.
This week a lot more time spent working in person, because we're still in the planning and testing phase of our analysis, Michelle and I have been meeting much more frequently. We did some preliminary planning after our group meeting on Thursday as well as twice more during the week. We also met as a whole team to work on our midyear report. I've really enjoying having the opportunity to work more with a partner and I think it's been very helpful for both of us.
The team is on winter break. Happy holidays!
The team is on winter break. Happy holidays!
The team is on winter break. Happy New Year!
The team is on winter break. Happy holidays!
The team is enjoying their last few days of winter break. See you next week!
This week was the first full week back from winter break. It’s nice to be back to get back into the swing of things. Natalie, Michelle and I met with Professor Veilleux to talk about the preliminary plan for the rest of our project. We discussed the time line for designing and conducting an experiment later in the semester as well as more technical emblements such as IRB paperwork. Later in the week Natalie and I met to discuss our findings from our comparisons of our annotated sound files. The results did not present the clear pattern we would have liked. We were able to find some pattern when comparing the data based on the amount of emphasis the speaker was given. We also had some difficulty’s surrounding the unmasking of the files. Both Michelle and I discovered some improperly unmasked files. We are working with Professor Veilleux to figure out the issue and solve it as soon as possible.
This was an exciting and different week as we embarked on two new aspects of our research. First, we came up with a preliminary design for our experiment. Our hypothesis is that participants will be able to identify which sound file matches which comic based on amount of evidence given in the comic. We are planning on giving the participants the correct comic, a comic that has a much different degree of evidence and a comic that has a different, but closer degree of evidence (photo on main blog. The other new thing we did this week was begin learning a new method of labeling called SToBI. We went over this new labeling convention as a group, practiced labeling a few files and compared them to previously labeled files.
We also continued to discuss the new SToBI labeling conventions and what we disliked about the new system. We discussed with professor Veilleux how we did not find the new system to be any easier than the old convention and how, in some ways, it’s more complicated. We also worked together to create a spreadsheet of possible combinations of comics for our study. We decided that we want to show participants three comic’s, one matching the voice recording we plan on playing, one having a similar degree of evidence, but not exactly the same, and one having a completely different degree of evidence. The spreadsheet was to examine the possible combinations.
This week we continued to delve further into our experiment planning. We talking with professor Veilleux about the possibility of using Amazon’s Mechanical Turk to run one or more additional experiments. As a group, we decided it would be best to run the in person experiment we have been planning and then use those results to do further testing online. We also discussed possibly of not continuing to annotate files.
The team continued work on experimental design with the goal of beginning experimentation shortly after spring break.
This week I competed my CITI course which all three of us are requited to complete before we send in our proposal for approval from the IRB which we are planning on doing before spring break. I also worked on finishing up our IRB proposal writing in our plans to include subjects completing the same experiment we have outline for Simmons Students but conducted virtually using Amazon’s Mechanical Turk.
It's spring break! See you next week.
Right off of spring break we had a huge storm which resulted in two snow days. Because of this I was not able to make it to our meeting with Michelle and Natalie. I know they discussed which two comics, aside from the correct one should be shown to the participants. My idea for our experiment set up seems to be similar to the one Michelle and Natalie discussed. I’m planning to find a time to discuss the exact details within the next few days, before we meet with professor Veilleux on Tuesday. Additionally, I have been reading about Mechanical Turk and Turk Tools to better familiarize myself with the exact details of how we will implant our experiment.
This week we met with Professor Veilleux for the first time in three weeks due to spring break and a subsequent snow storm. We updated her on the feedback on our IRB proposal and ironed out the final details of out experiment. We previously decided on giving each participant three comic options to match with the sound files, one that matches, one that’s close and one that’s farther away. We decided that we are going to present our subjects with all possible combinations of comics for each level of speaker hearer evidence. This means we will not be able to repeat any of the combinations more than once, however we do not foresee this being a problem due to the high number of people who will be participating in our experiment. Later in the week Michelle and I met and began work on our presentation board for Simmons College Undergraduate Symposium.
This week I edited the consent form with the feedback I received from our IRB reviewer, resubmitted it and received the final approval for our experiment so we’re all set to begin the experiment as soon as we get Mechanical Turk set up! Additionally I worked on designing the writing the poster for the Undergraduate Symposium.
This week worked on I worked on chopping up the sound files in audacity for the speaker we agreed to use for our experiment so they’re ready to go when Mechanical Turk is finished being set up. In addition to this I continued to work on designing the poster for the Undergraduate Symposium.
Back to the top