Standards for Assessments in Stat 340

--- tags: stat340, specs --- # Standards for Assessments in Stat 340 This document contains all the details for what constitutes “successful” or “acceptable” work on each type of assignment in the course. Use the table of contents in the upper-left to navigate to specific assignment types. ## Learning Targets There are two ways to demonstrate skill on a Learning Target: By working a problem on a Learning Target quiz or by working a problem in an oral quiz. Both written and oral quizzes are graded on the basis of *completeness* and *correctness*. A mark of `Successful` is recorded if the work meets the standards; a mark of `Retry` is recorded otherwise. | The work receives a mark of... | If it meets these criteria | | ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Successful | <ul> <li> Each part of the problem has a good-faith attempt at a response, and</li> <li> The response contains no more than two minor errors and no major errors, and </li> <li> Unless specifically stated otherwise, the response must be justified by sufficient relevant work or written explanation. </li></ul> | Retry | Not all the standards for “Successful” are met. | **Note:** The above are general criteria for successful demonstrations of skill on a quiz problem. Some quiz problems may have different or more targeted criteria that will be clearly stated on the quiz. Minor vs. major mistakes: - A **minor mistake** doesn't indicate a lack of understanding. It is a mistake that doesn't significantly affect your completion of the particular learning objective(s) being practiced or tested. These mistakes often result from a "typo," or from being rushed. A small algebraic error is a good example. - A **major mistake** reveals a lack of understanding of either the material we're covering, the logic and vocabulary needed to clearly explain your answers, or the prerequisite material of the course. It is a mistake that means you have not yet met the learning objective(s) being practiced or tested. Incorrect or poorly explained logical reasoning is a major mistake. Using the wrong formula or technique is another example. An algebraic or notational error that leads to a practically nonsensical answer is a major mistake. Not attempting a question is a major mistake. ## Case study and Project components Case study and project components are also marked as "Success" or "Retry." | The work receives a mark of... | If it meets these criteria | | ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | Successful | <ul> <li> Understanding of the concepts is evident through correct work and clear, audience-appropriate explanations, and</li> <li> The response contains no more than two minor errors and no major errors, and </li> <li> Unless specifically stated otherwise, the response must be justified by sufficient relevant work or written explanation. </li></ul> | Retry | Not all the standards for “Successful” are met. | #### Audience The audience for your case studies and project is your classmates---people who have the background knowledge that you have, but who have no familiarity with the specific model you are discussing. I am not the audience, and neither are you, so don’t assume that the person reading the solution has a Ph.D. or has been studying the model for a while. #### Elaboration Below, I elaborate slightly on what is required for a "Success" on each component of a case stuy/project. ##### 1. Given a research question, construct a Bayesian model that captures the information on the parameter prior to conducting the study. * The model is clearly and completely specified. Mathematical expressions for the priors, likelihood, and posterior distributions are given. (The posterior can be expressed up to the normalizing constant.) * The model is appropriate for the research question. * The model is justified to the audience. * Correct notation is used. ##### 2. Clearly communicate the prior information your are incoporating into the model. * Justification for the choice of prior is included in the report with sufficient detail for the audience. ##### 3. Given your data and model, implement an appropriate MCMC algorithm to sample from the posterior distribution using JAGS via R. - Necessary and sufficient details about the MCMC sampler are given in the report. - You submitted an .Rmd file containing your final MCMC implementation. - The analysis is correctly implemented via MCMC and I can run your code without alteration. - It is clear how MCMC convergence was assessed and that the procedure converged. ##### 4. Clearly summarize the relevant posterior information. - Relevant graphical/tabular summaries are included. - The posterior summaries are linked to the research question. - Summaries are statistically correct. ##### 5. Clearly communicate the implications of the analysis. - Results are clearly described. - All interpretations are accurate and logical. - All interpretations are in context. - The research question(s) are clearly answered. - The conclusions drawn follow logically from the results. ##### 6. Adhere to standards for technical writing. - The report has very few grammatical mistakes (≤ 3), and any mistakes present do not detract from the content. - The report has very few spelling mistakes (≤ 3), and any mistakes present do not detract from the content. - R code is not referenced in the text, and no raw R code/output is shown in the body of the report. - All figures/tables are properly polished and sized. - All figures/tables have informative captions. - All figures/tables are cited in the text. - Use proper mathematical and statistical notation and terminology are used. For example if you want to write “x squared”, type $x^2$ and not x^2. The use of plain text formatting of notation will result in the work being marked "Retry". #### Not assessable A component may be marked as "Not Assessable" if it does not represent a good-faith effort to meet the standards. This mark will be given in any of these circumstances: - The component is omitted, partially complete, or contains insignificant attempts at a solution. - The submission is significantly disorganized, illegible, messy, or unprofessionally written up. ## Homework While homework assignments do not contribute to your course grade, they will still receive feedback. In order to receive feedback on a homework problem, it must meet the following standards: - Be submitted by the deadline. - Assignment is scanned or otherwise compiled and uploaded to Gradescope. Each page should be properly tagged with the question number/part. - A good-faith effort was given. - Irrelevant R code and output should be omitted (e.g., messages printed when you load packages.) - Relevant R code and output should be included. You should annotate your code so that it is clear what each step is doing. Delete any irrelevant R output and clean or summarize anything excessively long.