CRITERIA
     
     1.  The goal should be to recognize the student, not the advisor or the 
     drafting department.  It is not usually possible to distinguish how many of 
     the ideas are those of the student alone, so it is best to try not to judge 
     the science.  It is also difficult to compare different types of 
     science--is a field program "better" than a theoretical model?
     
     2.  Therefore, evaluation should be based on things such as timing, clarity 
     of expression, effective use of illustrations, organization and logic.  All 
     of these are primarily or completely under the control of the student.  You 
     might want to use different criteria for talks and posters, or attempt one 
     set of standards for both.  Things to include might be:
     
     a.  PRESENTATION
     
     FOR A TALK:  Should be audible from rear of room, with reasonably clear 
     enunciation and absence of "um, er, you know" (some concession could be 
     made for nervousness, and for those who are not native-English-speaking, 
     but not too much).  Time should be used effectively; the introduction 
     should not take half the time with results crammed into the last minute.  
     Points should be deducted for running more than a minute over maximum time, 
     and extra credit given if talk ends early with time for questions.  The 
     student should have practiced the presentation often enough to have the 
     timing right, so there is no excuse for running over.  Slides/viewgraphs 
     should be legible from the back of room, well-labelled, and not 
     overcrowded.  The main point of the figure should be obvious without 
     explanation.  There should not be too many slides; points should be 
     deducted for     large data tables or multiple graphs on one figure.  If 
     there are questions, the student should handle them with some poise, should 
     understand the point of the question and be able to answer it.
     
     FOR A POSTER:  Students should, again, be able to speak clearly; no "um, 
     er, you know."  They should tell you enough to explain any item, without 
     going into excessive detail (unless you ask for it).  They should explain 
     the poster logically, starting with background and going on to results and 
     conclusions.  The figures need not be high-quality materials (color 
     zipatone and fancy lettering), but should be neat and legible.  Poster 
     should be logically arranged, not just a lot of figures tacked up in no 
     order.  Title should be easily legible from 10 feet away, and there should 
     be an abstract or some short summary for people who just want to read.  
     Either too much text or not enough are minuses.  There should be some sort 
     of summary diagram or list of conclusions.  The figures should be designed 
     to be informative in a poster context, not just copies of something for 
     publication.  Extra credit might be given for visually catchy set-up and 
     use of color.  Student should be able to handle poster alone; points should 
     be deducted if he/she turns to the advisor for help.  If the advisor 
     attempts to take over, continue to address your questions to the student.
     
     b. CONTENT 
     
     FOR A TALK:  Arrangement should be logical; it should explain the problem
     to be addressed, describe methods (briefly), present results, and draw 
     explicit conclusions.  Points should be deducted for diverging into 
     unnecessary details. The purpose of the study should be clear, not just a 
     description of data.  At least one conclusion should be reached, and 
     substantiated by the data.  Although hard to do, try to assess whether 
     student understands the significance of the work, or is just parroting 
     his/her advisor.  Study may not be earth-shaking, but should be elegant and 
     contribute something new to the field: useful new data, a new model, a test 
     of an old model.  There should be evidence of familiarity with the 
     literature and work of others.  
     
     FOR A POSTER:  Same criteria as above.  Data should be enough to support 
     conclusions but not too much; a few results that show the trend are better 
     than trying to show every single piece of data.  Either verbally or 
     visually, there should be a statement of the problem and of the 
     conclusions.  With a poster it is easier to determine whether
     student understands the work.
     
     3.  You might want to have an overall ranking for each session, as well as 
     the individual score sheets.  The judge will probably end up trying to give 
     the most points for what is thought to be the "best" talk.  To get around 
     this (or in case of a tie), you could allow for "extra credit" or
     "special considerations" ("Even though Bill has a slightly higher score 
     than Jane, I thought Jane's talk was better, though I can't quite verbalize 
     it.")