Tom Lowe

Tom Lowe commented on Perform a one-sample t-test 3 years, 11 months ago

When trying to test run the question I get an error, and in the variables section the variable "scenario" is not defined, with the following error message:

Error evaluating variable scenario: No definition of 'sum' of correct type found.

Tom Lowe on Perform a one-sample t-test 3 years, 11 months ago

Gave some feedback: Doesn't work

Tom Lowe on LSD and Tukey yardsticks, and two-way ANOVA - SES2003 3 years, 11 months ago

Gave some feedback: Has some problems

Tom Lowe commented on LSD and Tukey yardsticks, and two-way ANOVA - SES2003 3 years, 11 months ago

Could use some adaptive marking in parts e) and f), as the LSD and Tukey yardsticks depend on the value found for the sqrt(RMS) in part a), and which differences are found to be significant depends on the means for the groups found in part d).

Also there is no advice for parts a), b), c) and d), and the advice for part e) doesn't explain where most of the terms in the calculation of the yardsticks come from (ie the critical values or the sample size of one group).

Tom Lowe commented on LSD and Tukey yardsticks, and one-way ANOVA - SES2003 3 years, 11 months ago

Could use some adaptive marking in part e), as the LSD and Tukey yardsticks depend on the value found for the sqrt(RMS) in part a), and which differences are found to be significant depends on the means for the groups found in part d).

Also there is no advice for parts a), b) and c), and the advice for part e) doesn't explain where most of the terms in the calculation of the yardsticks come from (ie the critical values or the sample size of one group).

Tom Lowe on LSD and Tukey yardsticks, and one-way ANOVA - SES2003 3 years, 11 months ago

Gave some feedback: Has some problems

Tom Lowe commented on Perform a one-way ANOVA 3 years, 11 months ago

Question could do with adaptive marking, as later parts of the question rely on the answers to earlier parts. Also, there is no advice for parts d), e) and f), and the advice for part b) only gives working for the TSS, for the BTSS and RSS it just gives the answer. 

There is also no advice for how the standard error in part c) is worked out, and the wording of the question is, to me, confusing. The expression in the statement at the beginning of the question implies to me that the question is asking for the the standard deviation of the entire dataset divided by the square root of the total sample size, which is the standard error of the total mean. However, the question is actually asking for the standard error of the mean of a treatment group, which is the square root of the Mean Square Error (or Residual Mean Square) divided by the square root of the sample size of that treatment group (in this case all treatments have the same sample size). I am not sure if this would be more clear to someone currently taking the module but to me it seems ambiguous. 

Tom Lowe on Perform a one-way ANOVA 3 years, 11 months ago

Gave some feedback: Has some problems

Tom Lowe on Pattern matching - student's answer is a fraction 3 years, 11 months ago

Gave some feedback: Has some problems

Tom Lowe commented on Pattern matching - student's answer is a fraction 3 years, 11 months ago

Pattern matching doesn't work when the answer simplifies to the form x/(x+b), ie when a = 0. Answer is marked as incorrect unless the answer is specifically of the form (x+0)/(x+b).