Timeline for How to choose a good grading curve for yes/no exams?
Current License: CC BY-SA 4.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 18, 2018 at 19:36 | history | edited | sds | CC BY-SA 4.0 |
emphasize repeated shuffle
|
| Jun 18, 2018 at 19:23 | comment | added | Acccumulation | Arguably, someone who gets everything wrong has displayed one bit fewer of information. | |
| Jun 18, 2018 at 17:14 | history | edited | sds | CC BY-SA 4.0 |
dependence on Bernoulli rng
|
| Jun 18, 2018 at 17:08 | comment | added | sds | @EspeciallyLime: yes, see edits. | |
| Jun 18, 2018 at 15:18 | history | edited | sds | CC BY-SA 4.0 |
add Protocol
|
| Jun 18, 2018 at 15:00 | history | edited | sds | CC BY-SA 4.0 |
add Caveats
|
| Jun 18, 2018 at 14:07 | comment | added | Especially Lime | The Matthews coefficient has some (presumably) undesired behaviour in not treating questions equally. For example, if the test has five statements, three of which are actually true, then a student who gets only one wrong will score 0.61 if the mistake was thinking a false statement was true, but will score 0.67 if the mistake was thinking a true statement was false. | |
| Jun 17, 2018 at 18:18 | comment | added | sds | @JohnK: I would treat blank as submitting a random answer. | |
| Jun 16, 2018 at 19:09 | comment | added | John K | How would leaving an answer blank be handled with these systems? | |
| Jun 16, 2018 at 18:32 | comment | added | Erel Segal-Halevi | Looks good. I wonder if I can convince the university to accept such a scale. | |
| Jun 16, 2018 at 1:45 | comment | added | Superbest | As a stats nerd, I love this answer, but I suspect some students might not understand it and conclude you are grading unfairly. | |
| Jun 15, 2018 at 20:33 | history | edited | sds | CC BY-SA 4.0 |
expand
|
| Jun 15, 2018 at 19:38 | history | answered | sds | CC BY-SA 4.0 |