The OCC website is a valuable resource which I believe to be overlooked by the busy teacher. However I am here to share a recent experience and to say that if you have not looked or used it recently, its good and you should.
Its Grade 11 internal assessment season at my school as I finish marking the students first full design lab on Enzymes and busy myself prepping them for their second on Plant Science before the end of the year. Following a discussion with a colleague on what constitutes minimal data for Design aspect 3 I took my query to the OCC.
To start a discussion you click on “new topic” which on the top of the screen under the top banner. I received almost immediate support.
Here is a summary of the points raised / discussed.
The methodology that is written by a student should demonstrate that they plan to collect either:
an minimum 5 sets for the continous independent variable with 5 repeats (e.g. 5 different temperatures each repeated x5) so that the data can be analysed through standard deviation
a minimum for 5 sets of data if doing a comparison between two variables (A and B) with a plan to do a t-test etc.
In reality within the constraints of a busy classroom it can be difficult to carry through with the 5×5 model and in such circumstances a 5×3 is acceptable but should be explained within any modifications included within the paper.
Important to note that the SD is now not appropriate (5 is the magical number for SD, and the error bars will now reflect the range of values. Students should always indicate what the error bars represent.
It is a good lesson to discuss with the students that the larger the sample size the more representative the sample is, but the less realistic the design is in terms of time demands. Big picture concept we want students to recognize is that you need sufficient data to justify a conclusion.
The discussion also touched on minimal data processing. It was suggested that an average is acceptable for a partial but only if its value is recognized and used in the conclusion otherwise it constitutes N. I still believe that a good research question will set out the path for suitable data processing; be it a correlation coefficient or a t-test for a “significant difference” and a complete has to measured against the question / focus of the lab. However, that is only my take on the picture. Yours may differ from mine.
It was not long before I posted again as I was asked whether SL students were okay doing labs on HL material for internal assessment purposes (The Plant lab). Though I believed it to be fine I again took my query to the biology forum and received a knowledgeable response by the end of the morning.
I encourage you to go take a look at the discussions, perhaps it will inspire you to give advice or possibly you will extend a discussion.