Log in

No account? Create an account


gib emos about vignettes

Apr. 9th, 2007 | 08:59 pm
posted by: inhumandecency in ihdresearch

Questions about negemo evoked by situation predict helping others for leg and ups, not for apt or kidney. For leg, helping an acquaintance more also predicts more uneasiness, probably because it's so physical.

splits by friend type typically v. low n, hard to tell anything.

form 7 is messed up! remove extra friend question for next time.


scbroad PRELIMINARY!!!!!

Apr. 7th, 2007 | 06:36 pm
posted by: inhumandecency in ihdresearch

this is a big secret.

two-way interxn with acad and gpa. other seem possible w more n.
tried interxn between cond and goals. learning goals enhance predictions. but, when the interxn goes in, main fx disappears. it's entirely because learning goals buffer against the fx of social comparison.
will continue running until subject pool closes.


(no subject)

Apr. 3rd, 2007 | 07:41 pm
posted by: inhumandecency in ihdresearch

fragmented day with no clear, long stretches of work. helped some RAs, got the last pieces of ms submission ready, talked to diss committee, wrote to A, and talked to B for a while about tests to run on lkt3 and ppl to see for postdocs. will get ready for the merger of gib1 and gib2, and that's probably it for the day.


gib results, 070320

Mar. 20th, 2007 | 12:15 pm
posted by: inhumandecency in ihdresearch

I finally got the new data processed... only 44 people from all that running! The condition effects are totally null, though. no emo diffs after the questionnaire is done, no diffs in helping self or other.

type of friend predicts helping. ios and weness are obviously corred with friend type, but even residual ios still predicts.

helping self and helping other correlate, right on the edge of significance. So they're not negatively correlated after all.

people help themselves more than others (remember not to use z-scores for this comparison!)

PE is correlated with amount helped!

let's find out whether PE correlates with amount helped regardless of IOS.

help self & help other correlate significantly.
looks like unweighted vars are bad.
more self-help seems to reduce posemo!
emopos corrs NEGATIVELY w help1st (0=other, 1=self). iow, if you help the other person first, you end up feeling better (only if it's an acquaintance).

helping other only makes you happier if it's a friend!
also, IOS and oneness only affect amt. of help if it's a friend.


cfp work

Jan. 20th, 2007 | 02:55 pm
posted by: inhumandecency in ihdresearch

1) the two cond variables are the same! references in surveydata are bogus; it's assigned by control.html

2) emocond must be added manually from old versions of the file

3) realized that headache is referenced whether she says she avoided the headache or not. am changing the item from "no headache" to getting a small headache but not a bad one.

4) both new ras say the resps are too enthusiastic. going to tone them down.

5) coding resps for whether they noticed connections. will check for emo fx if I can get emo conds into the file...

6) going ot make the other changes I said I would in the aps app.


(no subject)

Nov. 7th, 2006 | 05:12 pm
posted by: inhumandecency in ihdresearch

worked on photo data today. smile / nonsmile rank correlations are .5-.6... lower than I'd expect. confused... the hottest ppl don't look all that hot. must check and make sure these are correct. Check data entry... guuuurgh.

Also worked on the programming for lkt3 and b-e. going slowly, but I'll get there.

have added a bunch of stuff to the HoPP chapter. Still need to reinsert HoPP-specific sections and jazz up the tone. A few other changes remain in the outline document from my meeting with barb, too.


(no subject)

Oct. 20th, 2006 | 04:42 pm
posted by: inhumandecency in ihdresearch

made updates to the irb for the symmetry study. ready to go as soon as we get the questionnaire and the photos ready. will meet anna next week and show her how to get a website up.



Oct. 20th, 2006 | 07:27 pm
posted by: inhumandecency in ihdresearch

deleted fake / bad participants from the dataset. Left in the outliers, but ran test with them excluded.

Yesterday I wrote code for converting all the scores to effort-weighted numbers from the pilot.

Today I converted to weighted variables ("w") and also created z-scored variables within each item, to address the problem that you can't go as high, e.g., on the apartment item as on the kidney item. These create z-scored summary items.

re-ran yesterday's findings with the outliers removed (4 ppl who scored below 4 on self-help). findings get weaker. see below:

The best vars to use are the non-z-scored weighted variables. This has negemo increasing self-investment and helping self first marginally increasing other-investment. posemo no longer affects self-investment. picpos marginally correlates with self-help, but only the unweighted vars (z or not). For those, neg also stays significant (for non-z; p=.08 for z).

with outliers removed, the help1st*friend interaction falls out. help1st has a main effect for both types of friend.


(no subject)

Oct. 19th, 2006 | 12:20 am
posted by: inhumandecency in ihdresearch

I imported the data from the latest run of gibblit. Now we have people in pos, neg, and neut.

higher pos emotions in pos than neut -- this differs from the diag version. neut and pos are the same on NE.

Re-ran all the syntax from gib.sps. it's fine.

no outliers on helpo, although close-selfFirst has a negative skew.

manip chex:
ios corr with friend & helpo
helop corr w friend & ios

across all:
helpo slightly higher when you help self first; helps when you helpo first. p<.15
helpo much higher for close than acq. helpo-close same as helps.
helpo unaffected by current emotions, but helps + with PE and - w NE (p<.08)

helpo based on friend * help1st:
if you help self first, you help friend more, but only if they're an acq. a close friend you help even more, and equally.
both fx, and interxn, are sig.

cond does not affect helpo when added to that model.

checked and made sure I was using the right attributions for each form. I was!