Ubigreen/CurrentListOfARs > Ubigreen/HomePage ? > Ubigreen/MeetingNotes

March 26, 2008

Attendees

Beverly, Sunny, Pedja, Tawanna

Overview of today's meeting:

Details on how to continue running participants (2 in each city after the study is over?). Initial concerns raised regarding IRB and payment, should I ask them if they are willing to volunteer, how will this process work?
- Continuing subjects may be risky as it adds cost to the project. For instance, there are a limited number of MSPs, having subjects keep them longer increases the chances of the MSPs and phones to be damaged, and how will updates be managed/maintained. Is this time better spent making a better study (or improving the software)? Also, who will continue paying for the unlimited data plans? Is this in the IRB? Will subjects be volunteers from this point? What will we get out of having them continue?
- It may be better to think about this first. At this point, let's ask who's interested in participating in the next study and think about continuing that study once it has been thought out more properly. Go ahead and close out all subjects - we need to close out the study so there is time for data analysis and paper writing.
Paper
- Jon and Tawanna are leaving for CHI next Thursday, April 3rd. James is leaving on Friday. We should aim to have the paper completed before Jon and Tawanna board the plane.
- We need to work out a plan for submission. Who will submit the final paper? What types of analysis will go in the paper? What makes sense? How much analysis will go in the paper? Again, in order to do analysis, we need to close out participants.

Plan of Action and ARs

Jon and Tawanna: Work out a plan for final submission. Based on survey and initial interview results, identify the main contribution of the paper.
Pedja is currently making revisions to the paper. He will pass the paper to Jon within the next couple of days once he has completed making revisions.
Jon will look at the current state of the paper and inform Beverly and Sunny when they should expect to receive the next revision of the paper (please let them know ahead of time so they can plan ahead. They have other deadlines).
Jon and Tawanna: Send out link to post-survey so participants can begin completing surveys. Ask to have these completed before their scheduled interview or by Sunday (whatever comes first). Therefore, we can begin data analysis.
Jon and Tawanna: Try to close out all participants by Sunday and Monday if possible.
Pedja will be available to consult regarding data analysis since he is currently working on this with another study. He will be able to help determine what types of data to include in the paper.
The paper should be submitted on April 3rd before Jon and Tawanna board the plane.

Feb 27, 2008

Attendees: Sunny, Jon, Jen, Pedja, Tawanna, James
Brief Notes*
- MSP Status
  - Still waiting to see if new MSPs will be available. It will take them about 1 week to build once it has been determined that they can be built.
  - There are ~9 msps in the new cases at UW but a couple may be broken (Jon and Pedja have one also)
  - Tawanna had clips and safety straps ordered for the MSPs in Pittsburgh; however, since those devices only have 1 battery, the hope is that we can use the newer versions (which have clips and the straps). Suggestion to drill the hole for the safety strap on top of the clip instead of in the device.
  - Jon helped Tawanna get the MSPs set up to her T-Mobile device last week (phone is up and running with the MSPScope app)
  - Tawanna is currently getting ~11.5 - 12 hours of MSP battery life(still testing)
- Software and Design status
  - Design - James sent images to neighbor's daughter yesterday. Estimated that at the absolute latest, this should be done by Monday.
- This shouldn't hold us up because Jon's design is almost done - Yang needs to address an ActivityDesigner defect this evening. Once this is done, we should be able to install the software and start testing.
- See Pedja's email for details - current location tracking is working; however, there is a conflict with the survey and detecting two locations. Jon is working this out at the moment. The survey shouldn't be launched if there is a conflict. We should be okay for testing with this defect in but it will need to be addressed before the study.
- Overview of issues
  - Hardware hold up - not enough MSPs
  - Need to get the vehicle travel code working. About 24 hours behind.
  - IRB approval at CMU
  - IRB not yet submitted at UW
General Notes
- Participants shouldn't leave the country
- CMU team is slowly finding backup while Tawanna is away. A couple of students have volunteered to help already.
- Jen and Tawanna to get their data plans set up (Tawanna is not a T-Mobile or AT&T/Cingular customer - will need to work this out)
- Tawanna will set up the phones here for testing once sw is ready

Feb 20, 2008

Attendees: Sunny, Jon, Jen, and Tawanna, James, Beverly
What is the battery lifetime of the MSPs in Pittsburgh?
- MSP in Pittsburgh only has 1 battery so ~11 hours and there are more exposed sensors
Sunny has a different version of the MSPs; they have a second battery.
- We need to add sample survey questions to the IRB. Tawanna can add additional survey questions to submit with the IRB and Jon will send the questions that were asked in Pedja's study
UW Exemption Status: Jon spent time talking to James about filing for exemption; however, he doesn't think they will be able to file for exemption on the UW IRB application
IRB Status
- Come up with survey questions to give a flavor of what may be asked of participants
  - We can use Footprints and what has already been asked
  - Tawanna will add new questions
  - Jen will submit IRB tonight
  - As an FYI: IRB approval is needed before recruiting can begin
  - Jon will send Tawanna a list of questions used in Pedja's study (DONE)
  - Appendix B questions that exist - Questions that we already asked would be pre and post
Status of Hardware: How many MSPs and how many 2125s?
- CMU received 8 last week but these were offered to anyone interested in testing the devices in the department. Tawanna will try to reserve all 8 (but we will need backup devices in case).
- We would need extra devices "just in case" but Sunny has different devices.
- James L and James F will be getting in more devices.
- Jon and James will see if they can get more devices to run studies in UW.
- Sunny needs to do the inventory. Some are loaned out, some are doing Beta testing.
How many subjects will we run in each city? Right now we're only looking at Pittsburgh and 16 is too many for Tawanna to run. How about 12?
- It's another confound to have 8 subjects each in both cities (again, we have IRB problems to overcome). Not sure if UW can file for exemption but they should be able to do minimal risk.
- If we only run the study in Pittsburgh, should we look at 12? UbiFit was run with 12 subjects (w/2 failures)
What if the IRB is not approved by March 3? If we don't have approval by CMU by then, we should submit an IRB to UW as well.
Polar Bear Design Status:
- We need to talk about the design.
- Polar bear has more space at the top and tree has more space at the bottom. We may be able to add user status at the top of the polar bear design assuming design work is done such that the design does not look out of place.
- We need to update the aspect ratio
- We will need about 19-20 different polar bear images
- Current tree and icon design - currently, there are different states for the tree - leaves are already on the trees (there is no script to do this) - James decided not to do the scripting.
Tree Design Status:
- Jon pretty much has an end to end demo but he still has some cleaning up to do with the scenes/states

Opens/ARs

Jen: Add Craigslist and word of mouth, bulletin boards, etc... to IRB so we can recruit from anywhere
Jon will advertise that a designer is needed - Jen to send Jon a list of specific requirements for James' designer.
Tawanna will be in Seattle from March 11 - 16
The goal this week is to have good enough to work on people's devices next
Tawanna: send Jen additional survey questions this by this evening

Feb 13, 2008

We've decided that incorporating the Stepgreen website into this study may be asking too much (given our tight deadline).
Study design
- 2-3 weeks
- 3 groups
  - polar bear: has phone + MSP + polar bear UI
  - tree: has phone + MSP + tree UI
  - control: has phone + MSP (logging only, no interface)
  - note: each group is to wear the MSP
There is some discussion about whether having a control group is actually a good idea as it begins to reflect a "quantitative" study rather than a more qualitative, exploratory study (which is the goal at this point, right?)

Jen argues that we should see some behavioral differences between control group vs. study group even if it's just novelty effect.

Vehicle transportation inference
- The MSP at this point, will NOT be trained for vehicle transportation inference (e.g., bus, car, scooter)
- The MSP is trained for walking, running and bicycling, however. I'm not sure people "run" as a replacement for their commute, some would bike.
- It's unclear whether the MSP's inference of biking is good enough to use to affect our interfaces
- Current plan: use jon's GSM-based motion inference code to ask the user a survey after they finishing moving
- James mentioned that we might avoid triggering a survey when the MSP says that the subject was walking / bicycling.

Feb 6, 2008 UBIGreen Notes from 02/06/08

Attendees:

Pedja, Beverly, Jon, James, Sunny (new to the team from Intel),Tawanna

Agenda

IRB Status
IRB Workaround
Questions regarding Study
Individual Status
Opens
ARs

IRB Status

Jon changed focus last week to work on a prototype to send Jen and Tawanna so there haven't been much updates to the IRB. This week, the focus should be on the IRB and since the process has been said to take 7 weeks minimum for a minimal week study, we need to expedite this process.
Quick look at timeline:
- UbiComp deadline: April 4th
- We need 1 week - 10 days to do analysis (and this is pushing it)
- The study is 2 - 3 weeks
- We need to be launched and loaded by March 1st

Discussion:

Scott has done a multiphase study but we're not sure if ithey've all been minimal risk. How long did "Voting with your feet take"? Jon: about 2 months? not sure...it was a long time.
- If it goes off great, if it doesn't, then it's not much we can do about it.
- What about a pilot?
- Can we only do through CMU? No, we have to go through UW with affiliation with the lab in Seattle. We should just move forward on this and flush out details. Can we go with the simplest qualitative HSB? Can we go exempt? Can we modify the IRB so that it can go through faster?
- Can we have modded application with Scott?
- Nothing has been put in for the ESM study (the ESM study is an Intel project)
- Can we do it at CMU and get an exemption from using their data "trick"?
- We can call it an Intel/CMU study and get an exemption for UW to do the study.
  - Intel does the data collection, not UW. Get CMU IRB.
  - Scott and Sonny kind of did this.
  - How can we run it out of Intel? Pedja and Jon are not Intel employees.
  - How does UW do subjects on their end? Can they do interviewing?
  - If UW sends an IRB revision on Friday? Could we get it back in time? Maybe... perhaps we could do a 2 week instead of a 3 week study?
    - Tomorrow and Friday free for Beverly, Tawanna free on Friday and Monday to work on IRB

Outcome:

We should try the CMU approach.
Sonny, Beverly, Jen, and Tawanna file at CMU, Data will be public (anonymized) - will file at UW as an exemption using public data.

Questions:

When do we do that? When do they file the UW one?
Does the CMU one need to be approved before we do the exemption? Not sure... (moved to open issues)

Questions regarding study

Issue: AT&T 8900 - can't get this to work on the T-Mobile. Only used AT&T subjects on UbiFit. Don't work with T-Mobile.
What are the rules for pilot testing and recruiting subjects (front end things we'll want to do) - Is there an issue with pilot testing?
Do we have budget?
We may have to do audio transcription? recording?
Do we try to recruit ourselves? We can try mailing lists. There may be issues with

Individual Status

Tawanna:
- Tawanna worked with Yang last week to complete prototype and sent out screenshots of polar bear application. Yang reviewed prototype and everything looks okay according to Yang (Tawanna prototype). Tawanna still needs to add functionality to StepGreen to post to ActivityServer and allow posts to StepGreen.

Question:

How much do we match these interfaces? Do we measure the negative response as a variable? We use a within subjects design? We'll use a between subject design. Too long, too many weeks to design. Tree or polar bear? People are only exposed to one, we want to test the positive versus the negative. We can have people point out what they like and dislike. They can't discuss them if they haven't been exposed to both.
Jon:
- Jon has been testing with the Cingular 2125 devices (no touchscreens)

He has also been testing with the AT&T tilt (GSP phone) - windows mobile fix phone. He has the phones talking to the MSP - the data is coming in. He needs to figure out how to upload it to ActivityServer. Jon is running into logistical issues and is having problems accessing the Intel server. He is looking to download more samples. Jon has also looked into the MSP sourceforge code (he wants to see how ubifit was using this code). Suggestion is to talk to Ryan and ask. It appears that the versions have been lost on the transfer from CVS to Sourceforge (it's only one commit). Jon has also been looking into Skyhook Wireless but he's not sure if its using GSM and he can't get it to work with C#. Overall, Jon is pretty close to being done (was hoping to have it done today but may be a little late). We will need to port Tawanna's to work the same way once this is done.

As for phones, they are going to try the hack on the 2125 phones (we have 15 of the AT&Ts total right now). The preference is to use the old phones as there is a manual, and we have 30 of them. It's worth waiting to see if we can make everything work on the phone.

Opens

Does the CMU IRB need to be approved before we do the exemption? Not sure. Who will find this out?

Jan. 30

On Jan 30, 2008, at 11:01 AM, "Harrison, Beverly" <beverly.harrison@intel.com> wrote:

need to work out the following:

List of functional requirements from each UI, plus screen shots so we can compare and identify gaps to match up functionality across each UI
- Jen and Tawanna to list off functions and screens for polar bear; Jon and James to list off for tree UI
- List of survey questions from Jen�s web study � do we want to include any of these in interviews in this study?
- IRB � will say general population in IRB app but need to decide if we want to include or avoid certain demographics
- Jon to send Jen graphics requirements/limitations for design of images
- Jen to ask Sue about whether or not certain demographics were predictive of more motivated green behavior
- What data will be collected by manual ESM or web journaling in addition to MSP stuff?
- Need to lock down on which phones we would be using (and whether we need to order more).
- IRB draft � Ubicomp deadline is April 4th (no UIST deadline is visible or listed anywhere � I sent email to Steve Cousins to find out) � UW process is often 6 weeks to turnaround/approve so we need to file next week and even then it�s going to be very tight! Can we try to get a draft circulated by end of this week??

Feel free to add � I wasn�t taking very thorough notes!

Jan. 24

Tawanna Dillahunt wrote:

I thought I'd send out my status from last week to determine what work can be kept assuming I use MyExperience instead of ActivityDesigner (based on today's discussion) and also to have some additional questions answered

Last week's status
- Sent out proposed task list and design (are there any questions?)
- Began adding tasks, themes, and images to ActivityDesigner - worked on prototype but had a couple of questions regarding the prototype design (see below)
- Found a server to host ActivityServer - loaded ActivityServer to system (not sure how to test that everything works as expected though)
- Located 7-8 mobile devices (these are T-Mobile Nokia 6600's) (THESE WON'T WORK)
Overview
- It was proposed in today's meeting that instead of using ActivityDesigner, I should be using the MyExperience "hack" instead. This "hack" essentially would pull in the StepGreen website and animated polar bear and set the screen size to fit on the mobile device. If this works as stated, there would be no need to create a prototype using ActivityDesigner.
Open Questions:
- What is the definition of our study? (we talked about this today somewhat but further details are needed)
- If I use the MyExperience "hack", is there support for Flash?
- If I use MyExperience, do I still need to have a dedicated ActivityServer running in Pittsburgh?
  - James wrote: I don't think you can eliminate ActivityDesigner. You need both it AND MyExperience. MyExperience is possibly less required, but it will make lots of things easier and also make it easier to change the prototype (e.g., using MSP vs. not using it).
  - Our prototype is assuming SENSED activities from the MSP for walking, running, biking and then we will use either cell phone-based speed estimation based on cell tower changes or GPS-based phones to tell when you are in a vehicle, and then use MyExperience surveys to ask the person what type of vehicle they were in.
- What are the milestones for this project?
  - James wrote: Yes, we need a schedule w/ deliverables. I was hoping we'd have an end-to-end prototype to show you by the end of this week (without the vehicle part working). It might be early next week.
- Will 7-8 T- mobile devices be enough for the study?
  - James wrote: The Nokia's are not going to work as they don't work with MyExperience or the MSP. You need Windows Mobile 5 or 6 phones.
- Will T-Mobile Nokia 6600 suffice for our study?
  - Tawanna wrote: No, Windows Mobile 5 or 6 phones
Regarding last week's tasks
- Can't users simply access StepGreen via the StepGreen website? If so, what's the purpose in adding actions to ActivityDesigner? Should I only add transportation based tasks to ActivityDesigner (so these can be sensed)?
- Is there a way to verify that ActivityServer is set up correctly? I may or may not need to do this based on the second question from above...

Meeting Notes