 |
 |
Frank My lab can show you some tests you could run. In out of town, but I'd be happy to look at the data or have a student run some quick analyses next week. We'd just run analyses to help us figure out the best tests for you to run Juls Sent from my iPhone On Aug 31, 2010, at 1:41 PM, "Frank D. Granshaaw" <[log in to unmask]> wrote: > Hi all... > > I am writing everyone with two requests. > > 1) This year there will be yet another session on geoscience in > community college at GSA in Denver. I've been asked to speak on the > experience of adjunct faculty in community college earth science. > To get ready for this I've together an on-line survey for part-time > geoscience faculty. I have sent this out to all of our part-time > faculty, but am wanting to cast the net a bit wider. So if you are > part of a community college with part-time geoscience faculty or > know part-time faculty who might be interested in participating, I > would appreciate your forwarding them this email. The URL for the > survey is < http://www.surveymonkey.com/s.aspx?sm=6SjUkFphD1WBO8ZpLkeyRA%3d%3d > >. Please ask them to complete the survey by 09/24/2010. > > 2) I am also looking to get some advice on statistical tests for > some research that I am doing. I recently completed a set of > interviews with 22 professional geologists (experts), 17 grade 6-12 > earth science teachers, and 45 community college students who have > less than two classes in geology (novices), where I asked each > person to look over a 360° digital panorama of a scene on Mt. Hood O > regon. Each person was asked to select four features from the scene > that interested them for any reason (though they were asked to imag > ine themselves on a geology field trip to this place) and ask a ques > tion about each feature (either something that they wanted to know o > r something they would ask some students). Following the interview > s I listed and tallied the unique questions for each group. The que > stions were then coded according key operational and cognitive chara > cteristics. The final result of this coding is the percentage of th > e total unique questions for each group having the various operation > al and cognitive characteristics (e.g. 44.4% of the questions asked > by novices involved processes shaping the feature, 2.4% of their que > stions asked about internal structure, and 26.3% of their questions > were convergent). My aim in doing this is to compare these characte > ristics between groups. While I've been able to make some prelimina > ry statements on the basis of the sample statistics, the big issue > right now is testing the significance of these comparisons. I have > tried a z-test involving proportions that has provided me with looks > like reasonable answers, but the question of an n for each group is > a bit troublesome since n as a group size is less than 30 for two o > f the groups and given the structure of the test that I selected, n > for each group would appear to be the number of distinct questions a > sked by members of each group. The latter seems somewhat counter-in > tuitive. Both t-tests and ANOVA have been suggested as alternatives > , but those seem to assume some knowledge of variance, which is not > within the nature of the data have. In any case, forgive the long-w > indedness of this second request. This being said any insight any m > ight be able to offer would be much appreciated. > > Take care > Frank G. > > > Frank Granshaw > > Earth Science Instructor > Portland Community College > Sylvania Campus > Portland, OR > 503-977-8236