Tom Lowe

Tom Lowe commented on Apply Kruskal-Wallis test - SES 3 years, 4 months ago

The number of ties table in the advice section is not displaying correctly, I believe this is something to do with + signs being used in the table that are being interpreted as delimiters when that is not what was intended. Also the first block of text in the advice needs rewording for readability.

I have reworded some of the decision variable and choices in the final part for consistency and to avoid the use of "accepting" the null hypothesis.

Tom Lowe commented on Two factor ANOVA - SES 3 years, 4 months ago

There is a lot of advice here that isn't related to the questions that are asked to be completed. Also, the dec function that is defined in extensions and scripts does not seem to be working properly, as in the final table of the advice it merely seems to return the number input into it rather than a decision on whether to reject the null hypothesis based on that number.

Tom Lowe on Two factor ANOVA - SES 3 years, 4 months ago

Gave some feedback: Has some problems

Tom Lowe commented on 5 number summary for sports science 3 years, 4 months ago

Corrected some typos in the advice.

There are other methods of calculating upper and lower quartiles which may give slightly different answers and would thus be marked as incorrect, so perhaps the method used in the advice could be briefly explained in a "show steps" for part a)?

Tom Lowe commented on Measures of central tendency from a list 3 years, 4 months ago

Checked, looks fine.

Tom Lowe commented on Measures of central tendency from a list 2 3 years, 4 months ago

Advice contains an answer to a part b), which doesn't exist in the question.

Tom Lowe commented on Perform a one-sample t-test 3 years, 4 months ago

Question now works fine. I've made some corrections to the advice, by ensuring that the modulus signs are in the calculation of the absolute value of T, and changing part e) to say that we reject the null hypothesis if our test is significant at the 5% level, as is consistent with the marking (it previously said to reject if the test was significant at the 10% level).

I think the question would benefit from adaptive marking, but can't see any other issues.

Tom Lowe commented on Perform a paired t-test 3 years, 4 months ago

Question could use some adaptive marking as each part requires an answer from the previous part. Also it seems slightly strange that 12/18 of the marks on offer are for calculating the differences between two values - could these be awarded half marks?

Tom Lowe commented on Perform an independent two sample t-test 3 years, 4 months ago

I've corrected some typos, moved the steps that tell you how to calculate the test statistic to the part that asks you to calculate it, made the steps show more clearly that the absolute value must be taken, and changed the test statistic formula in the advice so that it matches the one in the steps. Question now seems fine

Tom Lowe on Perform an independent two sample t-test 3 years, 4 months ago

Saved a checkpoint:

Have corrected some typos, and am about to move the step showing how to calculate the test statistic from part a to part b (where the test statistic question is asked) but want to save first in case I accidentally break it.