Bad news - did not get the gambling scholarship. Very disappointing - that was a lot of work! However.
In other news, the SurveyMonkey stuff is pretty much ready to go. I have fixed the instructions and am in the process of fixing the blurry photographs (which happened when a too-small photograph was used and then made bigger). We will be ready to run very soon. Dustin has tested it, though, and it took him 1.5hours, so not sure what to do about that. We may have to split up the task between subjects.
The review continues apace. Spending today locked up in Ben's lab working on it from 9 -4 - that ought to get us somewhere. Tomorrow we will meet and get a new draft of the outline ready to go.
Did a lot of thinking on the value learning stuff, and will discuss with Mark tomorrow.
To do:
- sort out external advisor
- finish up hot or not surveys
- send Mark data from Stonehenge
- SD review - 2 days of hard work - all papers summarised
And finally, need to sort out my advisory committee ASAP, so spending some time researching that this week.
Thursday, June 25, 2009
Monday, June 22, 2009
Mark & Anne Meeting June 17 (plus updates)
1. Hot or Not: at a good point with this. Angele and I have photoshopped (well, Graphic Converter-ed) all 400 stimuli. Angele has set up placeholders in SurveyMonkey so we can add them in. This will take forever, but am hoping to have at least one ready to go by Wednesday or so. The rest we will do as quick as we can, but this way at least we could start running.
2. Value learning: had a talk about how best to do that. Need to decide on what type of stimuli to use (have pretty much decided to have people learn classes of stimuli). Mark suggested either mondrians or Leptons. Need to discuss this further, although I have written out a few experiments we could try and can get moving on this once the details are decided.
- The value learning bit is basically written in Python and could be ready to go with a bit of fiddling once we have stimuli
- the inhibitory devaluation will likely use the same paradigm we are already using for the Stonehenge stuff, so also more or less ready once the stimuli are done.
3. SD Review - met with Ben last Wed and again today. He's now got some articles summarized, just not in the outline section yet. I am hopeful that will be done by Friday. My stuff is nearly done - one more topic to go.
- we have planned to spend this Thursday and Friday exclusively working on this.
Also went to the SWAP conference in Hamilton this week. Lots of grad student and other presentations.
2. Value learning: had a talk about how best to do that. Need to decide on what type of stimuli to use (have pretty much decided to have people learn classes of stimuli). Mark suggested either mondrians or Leptons. Need to discuss this further, although I have written out a few experiments we could try and can get moving on this once the details are decided.
- The value learning bit is basically written in Python and could be ready to go with a bit of fiddling once we have stimuli
- the inhibitory devaluation will likely use the same paradigm we are already using for the Stonehenge stuff, so also more or less ready once the stimuli are done.
3. SD Review - met with Ben last Wed and again today. He's now got some articles summarized, just not in the outline section yet. I am hopeful that will be done by Friday. My stuff is nearly done - one more topic to go.
- we have planned to spend this Thursday and Friday exclusively working on this.
Also went to the SWAP conference in Hamilton this week. Lots of grad student and other presentations.
Wednesday, June 10, 2009
Mark and Anne meeting June 10
Whoops, accidentally skipped the June 3 update, but nothing much happened. That was a productive week SD Review-wise, and I collected some Stonehenge women data. I analysed it to see if there was anything useful (in case I could present at this SWAP conference), but no luck. Two of the women had really low Go/No Go accuracy, which was very disappointing and left us with not much to go on. Here is the email I sent Mark regarding that data:
> Well, the women are all over the place. However, I just read a paper
> for the SD review saying that women use different brain regions for
> inhibition, so they would have to have a high N to get decent
> results (it was an imaging study). So maybe they are just more
> variable in general. Also, there are only three of them currently.
>
> At any rate, I added in the controls (from the data we used for the
> poster) and have attached the new graph. On the no-go Trust ratings
> the women are going entirely in the opposite direction of all other
> groups. Apart from that, looking at all the SD people vs. the
> controls, they are not that far off. I am going to go back in a sec
> and see if I can find my original control data including women, and
> add that in to see if it helps clarify.
>
> I'm not really sure what to make of this, but I get the impression
> that we could really use a few more participants before being able
> to really tell what we are seeing.
So basically, nothing doing at the moment.
HOWEVER, I am set up to go out to Stonehenge tomorrow afternoon. I will give my little spiel about my research, and see if I can't convince some women to either be tested right then, or to set up a time for next week. Fingers crossed.
In other news, my half of the SD review is going well. Have figured out a great method for quickly going through papers and summarizing, then re-summarizing to a one-or-two-line sentence that could go straight into a paragraph of the finished paper. Really pleased as this is quick and totally traceable (always know where to find the paper in question but also have a detailed summary for memory-refreshing purposes).
Unfortunately, Ben hasn't started his yet (bar doing some reading). I am a little disappointed, but hopeful that we can get moving on it. In the meantime, I've shown him my method and we've divided up the remaining sections so there is always something to work on.
Mark decided to put the London alcoholics collaboration on the back burner for now, until we have some good, consistent data so we have something to try and replicate.
Had a meeting for the Hot or Not study. Angele is working hard on collecting stimuli. Still having a bit of trouble with the blond men. I sent her what I found, but it turns out some is not useful if they are not standing up. However, rumour has it she is leaving for ten days in June, so we are meeting Friday to re-evaluate where we are and make a plan of action. I would be happy to sort it out while she's gone (just want to be careful as it is her project!)
And finally, I want to try out some of this Value Learning stuff, similar to Jane's but a bit more direct. To that end, I wrote up an experiment and sent it to Christian. He has made it into an awesome Python program, which we worked on today in Python class. My next steps are to gather more info about expected value for Mark, and to figure out what our output data should look like for Christian.
To do:
- read up on EV
- find key references and send that and a summary to Mark
- decide how data output should look and send to Christian/Alex/Asma (copy Mark)
- meet re: Hot or Not, do whatever needs to be done re: E-Prime etc
- Stonehenge - give talk, maybe collect data
- SD Review: ongoing
- find book on python and read it
> Well, the women are all over the place. However, I just read a paper
> for the SD review saying that women use different brain regions for
> inhibition, so they would have to have a high N to get decent
> results (it was an imaging study). So maybe they are just more
> variable in general. Also, there are only three of them currently.
>
> At any rate, I added in the controls (from the data we used for the
> poster) and have attached the new graph. On the no-go Trust ratings
> the women are going entirely in the opposite direction of all other
> groups. Apart from that, looking at all the SD people vs. the
> controls, they are not that far off. I am going to go back in a sec
> and see if I can find my original control data including women, and
> add that in to see if it helps clarify.
>
> I'm not really sure what to make of this, but I get the impression
> that we could really use a few more participants before being able
> to really tell what we are seeing.
So basically, nothing doing at the moment.
HOWEVER, I am set up to go out to Stonehenge tomorrow afternoon. I will give my little spiel about my research, and see if I can't convince some women to either be tested right then, or to set up a time for next week. Fingers crossed.
In other news, my half of the SD review is going well. Have figured out a great method for quickly going through papers and summarizing, then re-summarizing to a one-or-two-line sentence that could go straight into a paragraph of the finished paper. Really pleased as this is quick and totally traceable (always know where to find the paper in question but also have a detailed summary for memory-refreshing purposes).
Unfortunately, Ben hasn't started his yet (bar doing some reading). I am a little disappointed, but hopeful that we can get moving on it. In the meantime, I've shown him my method and we've divided up the remaining sections so there is always something to work on.
Mark decided to put the London alcoholics collaboration on the back burner for now, until we have some good, consistent data so we have something to try and replicate.
Had a meeting for the Hot or Not study. Angele is working hard on collecting stimuli. Still having a bit of trouble with the blond men. I sent her what I found, but it turns out some is not useful if they are not standing up. However, rumour has it she is leaving for ten days in June, so we are meeting Friday to re-evaluate where we are and make a plan of action. I would be happy to sort it out while she's gone (just want to be careful as it is her project!)
And finally, I want to try out some of this Value Learning stuff, similar to Jane's but a bit more direct. To that end, I wrote up an experiment and sent it to Christian. He has made it into an awesome Python program, which we worked on today in Python class. My next steps are to gather more info about expected value for Mark, and to figure out what our output data should look like for Christian.
To do:
- read up on EV
- find key references and send that and a summary to Mark
- decide how data output should look and send to Christian/Alex/Asma (copy Mark)
- meet re: Hot or Not, do whatever needs to be done re: E-Prime etc
- Stonehenge - give talk, maybe collect data
- SD Review: ongoing
- find book on python and read it
Subscribe to:
Posts (Atom)