Saturday, May 5, 2012 0

Almost finished. I have 15 students left which could mean anything from 30 to 9o separate experiments to read through (my present school I am working on has 6 students and 36 separate assignments). My last school had 8 students and sixteen experiments as each student designed two complete experiments.

I have found the whole experience invaluable. It has shown me what other schools do and really brought home to me what to look for in each criteria that I mark. It will most definitely impact my classroom next year.

So what have I learnt or seen from the work I have moderated?

• When designing a lab it must be one that can be credibly done within the classroom. Breeding horses for artificial selection is not appropriate and will not gain a CCC
• That the sample size must be BIG. Work on the premise of 5*5 for continuous data. Its understandable if it has to be 5*3 but it cannot be less than 5 different temperatures or light intensities etc if its looking for trends. If its looking for a significant difference between two variables then it must be 2*10(minimum)
• Error bars are not compulsory yet invaluable in showing the variation of the data and allowing for informed discussion in the conclusion of an experiment on the validity of the data. So don’t mark your students down in DCPa3 for not including them.
• Its important that at least 5 data points are used minimum to determine the standard deviation from which error bars are often generated.
• That overlapping error bars do not mean anything in themselves.
• That a minimum of two sets of 10 data points are required for the t-test.
• That the t-test when carried out, should include information as to whether its a one or two tailed test and whether its a paired or equal variance test (how else can I repeat it?).
• That the ANOVA statistical test is well worth teaching to your students.
• That scattergraphs should not have data points connected and that a line of best fit as a trend line can be sufficient for processing data (provided its commented on in the conclusion).
• That when an experiment involves participants, they must sign that they agree to take part (Ethics).
• That is acceptable to do experiments with animals as long as the parameters within which it is exposed are similar to that in which it would normally be found.
• That qualitative raw data does NOT have to be included if its not appropriate to the investigation.
• That a literature value must be included if the thing being determined is of a known value. e.g. optimum temperature of catalase.
• That teachers should be reading the report from the IA moderator and making sure their IB co-ordinator gets it too them. I am seeing the same issues arise as in the reports from past years which I get to read.