[ECOLOG-L] Model Fitting and Data Quality
I have been following the AIC thread with some interest. While I'm a newcomer to the subject and don't know much about the ins and outs of model selection, it seems like data accuracy and precision should drive how much we penalize extra parameters. Kepler rejected circular planetary orbits and went with elliptical ones only because he believed Tycho Brahe's data was of such high quality that even a very small discrepancy between observation and prediction was worth taking seriously. Data that was not known to be as precise as Brahe's would not have convinced him to fit elliptical rather than circular orbits to the observations. I'd very much like to hear people's thoughts on this. Jane Shevtsov -- - Jane Shevtsov Ecology Ph.D. candidate, University of Georgia co-founder, www.worldbeyondborders.org Check out my blog, http://perceivingwholes.blogspot.comPerceiving Wholes The whole person must have both the humility to nurture the Earth and the pride to go to Mars. --Wyn Wachhorst, The Dream of Spaceflight
[ECOLOG-L] Ecology study abroad opportunity in India
The Davidson College Semester-in-India program is expanding to include theme-based study abroad opportunity focusing on ecology and environmental issues in India, mostly southern India. The program will run during Fall 2011 and is designed for, but not exclusive to, biology and environmental science/studies students. If you know students looking for a study abroad program in this area, please forward this to them, forward to your campus study abroad coordinator, contact Chris Paradise (chparad...@davidson.edu) and/or check out the following website to learn more (http://www.bio.davidson.edu/people/chparadise/india/india2011.html).
[ECOLOG-L] Rép : European ecology list?
Dear Simone, I don't know of any European ecology list. The French Ecological Society host a list (called ECODIFF) that advertise PhD and postdoc in europe and the rest of the world. You can look the adds and subscribe to this list on this page : http://www.sfecologie.org/ecodiff/ The site is in French, but don't be afraid, most of the post are in english ! Bye for now, Nicolas Am 2010-12-02 um 15:19 schrieb David Inouye: Hy all, I'd like to know if is available an Ecology list focused on Europe, I'm asking this because I'm a Graduate Student from Italy, searching for a PhD, some weeks ago I've wrote a letter searching some information about it, but I had not received any reply from Europe. Best Regards Simone Demelas From: Simone Luciano Demelas simone.deme...@gmail.com Nicolas Mouquet Institut des Sciences de l’Evolution - CNRS UMR 5554 - Université de Montpellier II - CC 065 34095 MONTPELLIER Cedex 05 nmouq...@univ-montp2.fr Tel +33 4 67 14 93 57 Fax +33 4 67 14 40 61 Skype : nmouquet personal web site : http://nicolasmouquet.free.fr/ group web site : http://www.eec.univ-montp2.fr/
Re: [ECOLOG-L] Model Fitting and Data Quality
An interesting aspect of this story is that Kepler's decision to accept the accuracy of Tycho's data was based on his subjective evaluation of the quality of the data. The idea that we can drive all subjectivity from science is an illusion. Bill Silvert - Original Message - From: Jane Shevtsov jane@gmail.com To: ECOLOG-L@LISTSERV.UMD.EDU Sent: sábado, 4 de Dezembro de 2010 4:40 Subject: [ECOLOG-L] Model Fitting and Data Quality I have been following the AIC thread with some interest. While I'm a newcomer to the subject and don't know much about the ins and outs of model selection, it seems like data accuracy and precision should drive how much we penalize extra parameters. Kepler rejected circular planetary orbits and went with elliptical ones only because he believed Tycho Brahe's data was of such high quality that even a very small discrepancy between observation and prediction was worth taking seriously. Data that was not known to be as precise as Brahe's would not have convinced him to fit elliptical rather than circular orbits to the observations. I'd very much like to hear people's thoughts on this. Jane Shevtsov
Re: [ECOLOG-L] quantifying scent
Scent is one of the standard examples of an important variable which is extremely difficult to quantify but which is well adapted to fuzzy classification. Although there have been many attempts to measure scent, I have not seen anything that succeeded. Smells are very complex and each molecule is unique. The mind can interpret these smells, but are there any computer programs that can tell pheromones apart? Bill Silvert - Original Message - From: R K podocop...@yahoo.com To: ECOLOG-L@LISTSERV.UMD.EDU Sent: quinta-feira, 2 de Dezembro de 2010 17:32 Subject: [ECOLOG-L] quantifying scent Would someone be able to recommend a survey paper on olfactory sensitivity in different mammals? I'm interested in the differences in sensitivity across taxa--whether certain species are more focused on certain elements of the olfactory environment. I realize this is something of a naive question, but I know very little about the scent landscape and how mammals live in it, so any guidance whatsoever would be most appreciated. Thanks much to all.
[ECOLOG-L] Costa Rica REU opportunity for summer 2011
Texas AM University has a new REU site funded by the National Science Foundation for 10 undergraduate students to conduct cutting-edge research in the cloud forests of Costa Rica. Specifically, students will have an opportunity to conduct research in the areas of: . Climate controls and change in a cloud forest . Plant-atmosphere feedbacks . Soil-atmosphere interaction . Partitioning of atmosphere and soil moisture Successful applicants will receive all expenses paid trip to Costa Rica in addition to a $500 a week stipend. Further information about the program can be found at http://costaricareu.tamu.edu/ http://costaricareu.tamu.edu/ and a description of the Soltis Center for Research and Education in Costa Rica can be found at http://soltiscentercostarica.tamu.edu/ http://soltiscentercostarica.tamu.edu/. Please advertise this study abroad research opportunity to your undergraduate students. The application deadline is January 31, 2011. Georgianne Moore, PhD Assistant Professor of Ecohydrology Ecosystem Science Management Texas AM University gwmo...@tamu.edu / 979.845.3765
[ECOLOG-L] Postodoctoral Position: Environmental Physiologist (Congo)
Environmental Physiologist: Specializing in flux towers, 12 month post doctoral fixed term contract. Cirad is recruiting a post-doctorate researcher with a view to evaluating the influence of the change in land use on the partition between the “green water”, transpired by plants, and “blue water”, which is a resource for soil, rivers, lakes and aquifers. Assigned to the Ecosystems and Plantations research unit, and the new EcoSoils “Functional Ecology and Biochemistry of Soil and Agro-ecosystems” mixed research unit (UMR), (s)he will be responsible for coordinating the network of flux towers (savannah, plantation, natural forest), for identifying the determinants of the partition between transpired water and drained water on the three eco-systems, and modelling the dynamics of water storage and flows essentially. Description of the Position More specifically, the candidate will have to (i) take part in the creation of two flux towers (savannah, which will be planted, and natural forest; 6 first months of the project) and monitor the entire system with the help of the CRDPI's technical personnel and an international volunteer who will also be recruited to carry out this task; (ii) follow training courses organised as part of the climafrica project and more generally any graduate school that may be able to complete his/her profile; (iii) contribute strongly to the unit’s summarising publications in terms of water and carbon flows, particularly for the savannah ecosystem, which has been monitored for three years (alongside the unit’s researchers and post-doctorate researchers); (iv) contribute to studies on water flows in the three ecosystems, notably using the isotope analyser and models operating at ecosystem level; (v) more generally, contribute to the CRDPI's scientific life (seminars, environmental physiology training for technical personnel, management of a student from Brazzaville university). Profile Required Doctorate in environmental physiology, with significant expertise in flux towers. Good knowledge of SVAT models. Aptitude for multidisciplinary work and team work. Capacity to work in Southern countries (assignments or expatriation). International research experience would be a plus, particularly in tropical environments. Fluent English (written and spoken) essential. Location: Congo – Pointe Noire For more information: Jean-Michel Harmand UPR Correspondent for the Operation and Coordination of Plantation Ecosystems s/c UMR EcoSols 2 Place Viala Bât 12 34060 Montpellier cedex 01 France Tel.: +33 4 99 61 21 68 Email: jean-michel.harm...@cirad.fr
Re: [ECOLOG-L] Model Fitting and Data Quality
Interesting comment, Jane. It would be interesting to know what was the basis of kepler's confidence in Brahe's data. Did Brahe make repeated measures of some unchanging object (like the peak of a distant mountain, perhaps) to give a feeling for the errors in his measuring system? Even though modern statistics hadn't been developed, I presume that they did have arithmetic averaging. Were Brahe's data published as raw data points or averaged points, and did he make this known? Here's something else to consider: would Kepler have had the same level of confidence in the data if it showed something that seemed REALLY weird? Going from circular to elliptical orbits may not be too much of a stretch, but what if the data showed sharp-cornered orbits or funny helices? If he would just as readily have accepted these, then he was truly letting the data lead him, but if his mind rebelled against the idea of sharp-cornered orbits, and he therefore reject the data as untrustworthy, he would have been exhibiting subjectivity and letting his preconceived notions affect his interpretation of the data. I have often heard of scientists double-checking surprising results to see if they didn't make some mistake. That's fine, and probably many mistakes are caught that way, but still, I consider it unscientific unless one just as rigorously rechecks ALL data in the entire project. Otherwise one is biased toward accepting the world according to one's expectations. In fact, even that standard of re-checking data leads to bias because it is only triggered when something unexpected happens. Erroneous data that happen to conform to expectations wind up supporting those expectations when that support is not deserved. How much junk becomes dogma because of this level of subjectivity? Martin Meiss 2010/12/4 William Silvert cien...@silvert.org An interesting aspect of this story is that Kepler's decision to accept the accuracy of Tycho's data was based on his subjective evaluation of the quality of the data. The idea that we can drive all subjectivity from science is an illusion. Bill Silvert - Original Message - From: Jane Shevtsov jane@gmail.com To: ECOLOG-L@LISTSERV.UMD.EDU Sent: sábado, 4 de Dezembro de 2010 4:40 Subject: [ECOLOG-L] Model Fitting and Data Quality I have been following the AIC thread with some interest. While I'm a newcomer to the subject and don't know much about the ins and outs of model selection, it seems like data accuracy and precision should drive how much we penalize extra parameters. Kepler rejected circular planetary orbits and went with elliptical ones only because he believed Tycho Brahe's data was of such high quality that even a very small discrepancy between observation and prediction was worth taking seriously. Data that was not known to be as precise as Brahe's would not have convinced him to fit elliptical rather than circular orbits to the observations. I'd very much like to hear people's thoughts on this. Jane Shevtsov
[ECOLOG-L] Landscape Ecology PhD/postdoc position available
Landscape Ecology position available for 2011 A PhD or Posdoc is needed for spatial modeling research on an Ecosystem services project (USDA-NIFA). The Department of Entomology, Purdue University seeks a researcher to work on the spatial modeling of biocontrol insects. The successful candidate will be responsible for conducting field work, managing seasonal field workers, and collecting, storing, and identifying insect predators. The insect samples will be used to identify patterns of soybean aphid predator distribution within the agro-ecosystem context to better understand how grassland protected areas and restorations contribute to the protection of soybean yields. Experience with spatially explicit modeling, a valid driver's license, and ability to do field work are essential. Experience with entomological surveys and GIS are advantageous. The assistantship will start in late May or early June 2011. Please contact Dr. Jeff Holland (jdhol...@purdue.edu) or Dr. Helen Rowe (hir...@asu.edu) for more information. Purdue is an equal opportunity/affirmative action employer and encourages applications from women and members of minority groups.
[ECOLOG-L] AIC and parsimony
Dear ECOLOG: I think there was an important aspect of Dr. Bigelow's quote from Burnham and Anderson (2002) page 131 that was not highlighted. THe passage quoted was: Models having delta-I within 0-2 units of the best model should be examined to see whether they differ from the best model by 1 parameter...in this case, the larger model is not really supported or competitive, but rather is close only because it adds 1 parameter... As I understand it, this is a means for distinguishing between between two models when one of the models has a higher dAIC and more parameters. However, if a model with the lowest AIC (dAIC = 0) has more parameters than a second model with dAIC2, the second model cannot be considered to be better than the first on the basis of parsimony. This is because the addition of more parameters to the first model is improving fit and lowering the AIC despite the penalty more parameters incurs. Is this true? Below is an example I sketched out: Say you have four models with wither 2 or three parameters. Model deltaAIC #1 response = B1(x) + B2(y) + B3(z)0 #2 response = B1(x) + B2(y) 1.2 #3 response = B1(x) + B2(y) + B3(u) 1.8 #4 response = B1(u) + B3(v) + B3(w)4 First, I believe this is the situation that kicked off this thread on AIC: Two models, the one with the lowest AIC has more parameters, the second with dAIC 2 #1 response = B1(x) + B2(y) + B3(z)0 #2 response = B1(x) + B2(y) 1.2 So, between #1 and #2, one should not conclude the #2 is better simply because it has fewer parameters. The addition of the third parameters to model #1 improves the fit of the model and has a lower AIC DESPITE the penalty for adding another parameter. Second, two models, the one with higher AIC also has more parameters #2 response = B1(x) + B2(y) 1.2 #3 response = B1(x) + B2(y) + B3(u) 1.8 Here, #2 and #3 both have dAIC 2, but #3 has more parameters. Therefore, model #2 can be favored b/c it is more parsimonious Third, #1 response = B1(x) + B2(y) + B3(z)0 #3 response = B1(x) + B2(y) + B3(u) 1.8 Here, #1 and #3 have the same number of parameters and are within 2 dAIC units. These models, I believe should be considered equivalent. Finally My question: what do you do if you have all three models under consideration? #1 response = B1(x) + B2(y) + B3(z)0 #2 response = B1(x) + B2(y) 1.2 #3 response = B1(x) + B2(y) + B3(u) 1.8 Nathan
[ECOLOG-L] Faculty position in GIS at Rowan University
Rowan University Position Announcement Department of Geography and the Environment Assistant Professor-GIScience/Environmental Planner Description: The Rowan University Department of Geography and the Environment invites applications for an assistant professor tenure track position to begin September 1, 2011. We seek a Ph.D. specializing in GIScience and Environmental Planning or closely related field. The successful candidate will teach introductory and advanced courses in GIS and Planning focusing on issues of environmental management and sustainability. The candidate may also teach some introductory environmental studies courses. Development of advanced courses in the candidate's area of expertise is also welcome. We also expect an active research agenda, which is encouraged by the university through travel support, internal grants and load adjustment. The successful candidate will also demonstrate the potential to secure external funding. Applications must include: letter of interest, curriculum vita, graduate transcripts, a one-page teaching statement, course evaluations, a research statement, and names and contact information for three references. Applications should be sent to: Dr. John Hasse, Department of Geography and the Environment, Rowan University, 201 Mullica Hill Road, Glassboro, NJ 08028. Applications may also be sent electronically to ha...@rowan.edu. Applications must be received by January 15, 2011. Rowan University values diversity and is committed to equal opportunity in employment. The position is contingent upon budget appropriations. For more information about this position please contact: Dr. John Hasse (856) 256-4812, ha...@rowan.edu. Qualifications: Applicants must have the Ph.D. in hand by September 1, 2011. Demonstration of excellence in teaching is a fundamental requirement as this is emphasized in our program. Salary: Competitive General Info: Rowan University is a comprehensive (bachelors and masters level) institution that values high quality teaching, scholarship and service. Our classes are small (20-30 students), and emphasize project based and interdisciplinary approaches to learning. We are conveniently located less than an hour's drive from the Atlantic Ocean, 20 miles from Philadelphia, and midway between New York City and Washington, DC (approx. 100 miles). The University's immediate surroundings provide a variety of housing and recreational opportunities in urban, suburban and rural settings. For more information about Rowan University and the (proposed) Department of Geography and the Environment, please visit: http://www.rowan.edu/ http://www.rowan.edu/colleges/las//departments/geography/ http://www.rowan.edu/environmentalstudies
Re: [ECOLOG-L] AIC and Occam's Razor
Hi, James 1. Because AIC has already been used: that's how the potential models had been chosen. Also, I was pointing out that it's useful to know in absolute terms how much variation in the data the model is explaining: dAIC only give a comparative measure. 2. You're right that prediction assumes conditions are similar enough - I'm not sure what I wrote caused confusion on this point. Can you amplify? 3. I agree that finding the range of stability is important, but this is going far beyond what I was writing about: finding the range in which predictions are reasonable is a whole research agenda. 4. True. But in practice, simpler models are usually easier to understand. 5. Again true. I'm nor aware that I wrote anything to contradict this. Bob Bob O'Hara Tel: +49 69 798 40226 (in Germany) Mobile: +49 1515 888 5440 WWW: http://www.bik-f.de/root/index.php?page_id=219 Blog: http://blogs.nature.com/boboh/ Journal of Negative Results - EEB: www.jnr-eeb.org James Novak 12/03/10 8:42 PM There are a couple of confusing points in your response Bob. 1.) Why would you use R^2 rather than dAIC and wi to see how large the differences between models are? 2.) Doesn't all prediction assume that conditions are similar enough that the prediction conditions are valid? 3.) Related to above then is it not more important to try and predict the range of stability rather than just throwing up our hands and saying things are not stable? 4.) Parsimony does not imply nor guarantee an interpretable model. 5.) Actually what we want is the most parsimonious model that adequately explains our data, not just the most parsimonious model. AIC is merely a metric that has some desirable properties and in a model selection procedure performs better than R^2 or BIC. Interpreting and explaining the model(s) comes down to biology as it should. I do not think anybody is treating AIC as an explanatory black box, but rather as a tool to help us select from a range of models. (((º (((º (((º (((º (((º (((º Jim Novak Biological Sciences Department 1162 life Sciences Annex Eastern Illinois University Charleston, IL 61920 (217) 581-6385 (217) 581-7141 (FAX) jmno...@eiu.edu http://www.ux1.eiu.edu/~jmnovak/ (((º (((º º))) (((º (((º (((º On Dec 3, 2010, at 9:08 AM, Bob ohara wrote: A couple of people have mis-interpreted what I wrote, so I think I should clarify. I was not suggesting that R^2 should be used for formal model selection. But rather, given your model selection suggests that there are several models that are similarly adequate, it's worth asking what you lose by selecting the model that is formally non-optimal. If you don't lose a lot (e.g. R^2 goes down by 0.1%), then from a model fit point of view, it doesn't really matter which model you chose. OTOH, if it changes by 10% it does make a difference, and you would have to decide if it was worth it for the parsimony (most likely not). There are a couple of issues underlying what I wrote. Firstly, AIC is inly optimal in a narrow predictive sense: in the sense of predicting the same data (this has a stability assumption buried in it, i.e. you're assuming that the conditions in which you collected the data will be the same for where you want to predict for). Now, I think it's very rare that we, as ecologists, want to make predictions in this narrow sense: we're not political pollsters. I think we're usually more interested in understanding what our data is telling us. Hence, having a parsimonious model is more important. There is really no point in fitting a model and then finding out we've no idea what it's telling us. The second point is that I come from a statistical background which doesn't blindly run the numbers. We're trying to understand our data, so the OP's question makes sense. The problem is to understand more about the models, and what the consequences are of using a model which isn't optimal. This will involve some subjectivity, but we're human beings so we're going to interpret the results with some subjectivity anyway. The important thing is to understand why the model we chose is the best. Just using AIC encourages a black box mentality, and doesn't remove subjectivity: unless one is doing prediction in a rather narrow sense (i.e. under exactly the same conditions as those used to collect the data), what objective reason is there for thinking that AIC is optimal? Bob Bob O'Hara Tel: +49 69 798 40226 (in Germany) Mobile: +49 1515 888 5440 WWW: http://www.bik-f.de/root/index.php?page_id=219 Blog: http://blogs.nature.com/boboh/ Journal of Negative Results - EEB: www.jnr-eeb.org Hal Caswell 12/02/10 11:38 PM There are a couple of strange things about the description of the scenario. First, the idea of thinking about a model with almost the same AIC (or, better, AICc) but fewer terms, in pursuit of parsimony is doing parsimony twice. The AIC already accounts for the relative number of parameters.
Re: [ECOLOG-L] Model Fitting and Data Quality
In fact, a while back I read a paper on modelling which stated, there is no best model, only useful ones. If you are relying on a computer or computer program, or a statistician, or some other resource which has no background in what you are studying, then you are setting yourself up for problems. First, you must understand the system in which you work before you can properly interpret statistical outcomes. Malcolm McCallum On Sat, Dec 4, 2010 at 8:35 AM, William Silvert cien...@silvert.org wrote: An interesting aspect of this story is that Kepler's decision to accept the accuracy of Tycho's data was based on his subjective evaluation of the quality of the data. The idea that we can drive all subjectivity from science is an illusion. Bill Silvert - Original Message - From: Jane Shevtsov jane@gmail.com To: ECOLOG-L@LISTSERV.UMD.EDU Sent: sábado, 4 de Dezembro de 2010 4:40 Subject: [ECOLOG-L] Model Fitting and Data Quality I have been following the AIC thread with some interest. While I'm a newcomer to the subject and don't know much about the ins and outs of model selection, it seems like data accuracy and precision should drive how much we penalize extra parameters. Kepler rejected circular planetary orbits and went with elliptical ones only because he believed Tycho Brahe's data was of such high quality that even a very small discrepancy between observation and prediction was worth taking seriously. Data that was not known to be as precise as Brahe's would not have convinced him to fit elliptical rather than circular orbits to the observations. I'd very much like to hear people's thoughts on this. Jane Shevtsov -- Malcolm L. McCallum Managing Editor, Herpetological Conservation and Biology Peer pressure is designed to contain anyone with a sense of drive - Allan Nation 1880's: There's lots of good fish in the sea W.S. Gilbert 1990's: Many fish stocks depleted due to overfishing, habitat loss, and pollution. 2000: Marine reserves, ecosystem restoration, and pollution reduction MAY help restore populations. 2022: Soylent Green is People! Confidentiality Notice: This e-mail message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.
[ECOLOG-L] information management in a graduate seminar
Some suggestions about relatively recent options for running a reading-intensive course. Inouye, David W. 2010. Evolution of Information Management in a Graduate Seminar. Bulletin of the Ecological Society of America 91:361362. [doi:10.1890/0012-9623-91.3.361] http://www.esajournals.org/doi/full/10.1890/0012-9623-91.3.361
