[Heartlogic-dev] Re: Re: ahead of schedule

[William L. Jarrold]

On Sat, 29 Nov 2003 address@hidden wrote:

> > Some more responses RE the amazing website you have put together...
> >
> > Okay, I'm FINALLY loging in...
> >
> > Loved the little gearheart that shows up in the URL line.
> >
> > Picky: I'd vote to remove "I want to donate my heart and brain to
> > science."
>
> Yah, fine.
>
> Actually I consider the whole thing tenative.  Is there
> such a thing as a "disclaimer agreement"?  The wording
> might be wrong.  Maybe it should be an indemnification
> agreement?  Or something else?  I need to re-read
> the text which you used for your study.

You are probably referring to informed consent.  The concept here is an
ethical one.  Unlike, e.g. Nazi's, we do not force other's to participate
in our experiment.  Rather we offer them an informed choice.  We tell them
what might be the side effects of the study, we tell them, to the extent
appropriate, what the study is about.  Now our experiment has a very
small risk of actually causing anyone the slightest amount of distress -
we don't really deceive our subjects or take their blood or anything like
that.  Nonetheless, as a psychologist I am compelled to follow ethical
practice.  I don't know what the guidelines are with respect to
free-for-all type web surveys.  I'm aware that the APA (American
Psychological Association) has some newly released guidelines regarding
internet surveys.  We should track those down and see what they say.  If
we stick close to what I got approval for on my dissertation, I really
doubt it is a problem.  I have added...

* Clarify informed consent RE Open Heart Logic

...to my to do list.

>
> > I think approximate time should be less.  My subjects had 30 min
> > to complete everything and read the instructions.
>
> Trivial.
>
> > I kinda think we want to do a strict replication, therefore I should
> > send you the instructions I had for the subjects.
>
> I have a tex file for Group#1 (or #4?).  I will
> copy the exact wording.
>
> > Also, I'd much prefer to not have the split.  For replication purposes
> > it is bad.  Also if we want to do a traditionally valid statistical
> > analysis of the data, then life will be much easier if we do not
> > have the split option.
>
> OK, it is trivial to take out.  I guess you wanted
> the split thing for the next survey or something.

Yes.  In future studies we might have them rate the (a) emotion alone,
(b) the explanation alone and (c) the emotion and explanation together.
Right now, all we do is (c).  I don't think we'd ever want to give them
the choice of either i: (a) and (b) OR ii: (c).  Intuitively it feels
kind of scary and complicated.

>
> > Also, I'd prefer to have the surveys look like the paper surveys.
> > Therefore, the likert ratings shoudl be horizontal, no?
>
> OK
>
> > Also the explain your rating part, "If it is not obvious then please
> > briefly explain your rating:"...That is not how I did it in the surveys.
> > Shoot, I sent you a suvery didn't I?
>
> Yah, I will go back to your exact wording.
>
> > On the replication part, I don't think there should be any hot links.
> > We want survey takers to be able to do one thing and only one thing,
> > take the survey.  Not go off on a tangent.  Just like the thing I did
> > for my dissertation.
>
> Are you sure?  I mean, it is easy enough to keep two
> web pages open and switch between them.  Or what if
> I complete half the survey then complete the second
> half tomorrow?  I guess we can ask people to complete
> it in one sitting.  What do you think?

For the replication we should prefer that they do it in one sitting, just
like my study.....Maybe the program could be written such that if they
quit early, we'd still have their data...In the early days we should see
how many people flake out half way through.  If we loose 10% or more, that
seems like a serious issue and we should think about relaxing the
requirement that they do it in one sitting....Remember, if we fail to
replicate the diss it would suck.  And we would want to have as few
factors that might be contaminating why we aren't getting a replication.

>
> > Also, I mildly prefer the word "dissertation" in place of "thesis".
> >
> > Hmm, I'm going to need to send you the surveys from all 12 groups.  I
> > think there were 12 groups in study 3.  Do you want them now?
>
> Why?  Currently everything is assigned randomly, but I
> can generate permutations if random isn't good enough.
> Do we need permutations?  If we use permutations then
> we can involve a minimum number of subjects which
> is probably good.
>
> > When I finished the survey it said "Approx. Time 0 min".  It would be
> > better to say "Approximate time remaining."  However, it should only
> > do this for the series of items they complete AFTER they do my
> > dissertation replication.
>
> Hrm, I have to check.  I think my version says "Completed"
>
> > We'll need about 10 ss per group. 12 groups means 120 subjects.  Once
> > we analyze that data and show a replication, we can not require that
> > folks do it.  But it might be a good training thing.
>
> Yah.
>
> > So, I should really REALLY send you the instructions that I gave
> > the kids who did my study, right?  Were those instructions in the diss?
> > Ug, I really need to unearth all my dissertation papers.
>
> No, I think I have everything.  I just need to plug
> it in.

Oh, cool.  If you happen to notice, on what page did you find the
instructions?  I should really make sure that the instructions
are as close to the same as in the diss.

>
> > So, interesting additions you've made but for the replication my
> > preference would be to not have them.
>
>
>