I’ve been trying the following approach for around a year now, prompted by the vexed question of the dreaded ‘practical questions.’ I was never satisfied with some earlier strategies, such as teaching accuracy and precision using dartboard graphics (this seems to be an extra, unnecessary layer of thinking), and the more definitional approaches adopted by some exam boards struck me as not easily accessible, particularly to younger students who would benefit from getting these ideas in their heads during KS3.
Validity
I noticed that many questions about validity, at least from the biology perspective, expected an answers related to variables. There’s an easy link: V for Valid, V for Variable. Of course, variables are an area that can also cause problems, but when I say down and thought it through there were actually a limited number of variables that are ever in play. I came up with the following list:
C – Concentration
L – Light intensity/wavelength
A – area/length
p- pH
S – species
T – Time/rate
T – temp
V – volume/mass
That’s it, just these eight. Sure, there may be other potential variables, and I only added species in later when I started teaching more of the ecology topics and realised it could be a useful variable to include. Essentially, most experiments in biology has its variables, independent, dependent and controlled, within this list. My GCSE and A level students now know that on a question about validity, to think first in terms of what the variables are in the experiment and then work out which ones haven’t been controlled. That is invariably (ho-ho) the way to answer the question.
Here are the AQA trilogy RPAs, and how the variable list fits in. Variables are not necessarily relevant to all of them. IV = Independent variable, DV = Dependent variable, CV = Controlled variable
IV | DV | CV | |
Microscopy | |||
Antibiotic growth | Conc | Area | pH, Time, Temp, Volume |
Osmosis | Conc | Mass or length | Time, Temp, Volume |
Food tests | Volume/mass | ||
Enzymes | pH | Time/rate, Volume | Conc, Temp, Volume |
Photosynthesis | Light intensity or wavelength, length (distance), Species | Time/rate or Volume | Conc, Area (of leaves), pH, Species, Time, Temp, Volume |
Reaction time | *hands, Conc e.g. caffiene | Time or length | *Same person, *ruler position |
Germination | Light, *gravity | Length | Conc, Area (of dish) pH, Species, Time, Temp, Mass (number of seeds) |
Quadrats | Light intensity, pH, species (e.g. grazer presence) | Species (frequency) | Area, pH, Time (seasonal) |
Decay | Temperature | pH | Conc, Time, Volume |
Apart from the ruler drop experiment, most of the experiments can be viewed through a limited number of variables. Yes, there are a few details to be clarify, but from the point of view of giving a starting point for validity questions, having an easily recallable structure is a good starting point for students. They can recall the mnemonic CLAPSTTV fairly easily, and then they have a better shot at identifying the variables that have not been adequately controlled.
Accuracy and repeatability
I find the best place to start here is with the word accurate. I describe an accurate result as the true, or real result. Even in Year 7 students can work out that experiments to measure things are prone to getting wrong measurements, so the idea that an accurate experiment gives the ‘true’ result is a useful starting point. No need for dartboards and archery.
After that, time to go to everyone’s pal, Excel. If an experiment would usually result in a proportional relationship, I would gather a set of class results, usually putting up a blank spreadsheet and getting the students to put their results in. There are normally going to be some anomalous results in there. For example, I recently got this from a yeast respiration experiment:
Temperature oC | Bubble rate per minute |
38 | 7 |
45 | 22 |
47 | 33 |
47 | 30 |
47 | 3 |
36 | 2 |
40 | 14 |
39 | 9 |
44 | 10 |
Firstly, we organised the results into ascending order, then made a scatter graph on Excel (this is important!). The graph was placed next to the table and adjusted so it was clear to see. Next, students could identify where a rough line of best fit would be. I then used Excel to fit a line, and it became even more obvious where the anomalous results appeared. Knowing that anomalous results should be identified and removed, we deleted them from the table. Each time you do this, the graph updates, shifting your line of best fit to a better (more ACCURATE) position. Students can see in real time the effect of this, and the reason why anomalous results are removed.
In other words, this is how I approach the accuracy/reliability question:
They are about METHOD. If you were writing out an experiment, one of your numbered points would hopefully to be to repeat so you can identify anomalous results and calculate a mean. How do you make an experiment more accurate? Look to the method, there will usually be a step missing about the repeats.
Fine, but students usually write ‘repeats’ as a guess anyway on these type of questions, so have we really improved the situation? I think so, since it has modelled WHY removing anomalous results improves accuracy, and that repeats (you can model this too on Excel by having repeated results and again, identify and remove anomalies from the table to get a real time update on the graph).
There is one trap here though, and it usually appears on the Osmosis and Enzyme RPAs. Here, the accurate result is not a relationship but a specific value, optimum pH or isotonic point. In this case, stress repeats should be done oat smaller intervals around the accurate value. This would of course be another method instruction, since intervals measured should be mentioned within a method.
So now students should be looking for answers involving repeat to identify anomalous results BETWEEN a range of points. This approach will answer most of the questions that they come across asking about accuracy. Sure, there will be exceptions and caveat, but when answering this type of question I feel having a framework is useful for many students.