ISA: How Science Works

There's virtually always a 4-mark question asking you to evaluate a conclusion made based on the results of an investigation (although I'm sure there wasn't one on the paper we did the other week?)Do as many of these as you can. The marking points will be specific to the investigation and you must read all the information but it's similar things that tend to come up. The sort of question could be: 12 rats were split into 4 groups and tested with different concentrations of the drug imositol. The results are as follows:Concentration of imositol               Mean lifespan of rat in months (+- standard deviation)10                                                 25.7 (+-2)30                                               26.4 (+-3.8)50                                               27.9 (+-2.1)70                                               28.1 (+-1.8)A pharmaceutical company claims: "IMOSITOL MAKES YOU LIVE LONGER!" Evaluate this conclusion.This is a slightly exaggerated example, but the sorts of things you'll have to say are the same.=> You usually get one mark for saying that the data agrees with the conclusion. Here you would say "the experiment shows that the greater the concentration of imositol, the longer the lifespan of the rat".The rest of the marks will come from slagging off the conclusion. The sorts of things you should mention include:=> Investigation was with rats and not with humans/effects could be different in humans=> Only 12 rats - not representative sample/not reliable=> Only one investigation - not representative/conclusive=> Genetic variation in rats affects lifespan=> No control group=> Overlaps in standard deviation - so differences are not significant=> Bias as the pharmaceutical company want to sell the drug=> Other factors could have been involved/correlation does not show causal relationship.The devils writing the mark scheme want you to write what they're thinking and sometimes you can make a perfectly valid point that won't get credited because it's not on the mark scheme, even though it may have been on the mark scheme a previous year, and was still applicable to this investigation.For this reason, I would always try and write as many valid points as possible - you won't get penalised for saying things that aren't on the mark scheme as long as they aren't plain wrong, so try and get in a couple of insurance points.

Reliability, accuracy and precisionAll these horrible How-Science-Works words you need to throw into your answers. If you look through past papers, you will see recurring question styles and the least you can do is decide how you would answer them for your own investigation. In the case of the glucose concentration investigation:What limitations of your experiment could have made your results less reliable/accurate? There was no control. We didn't do an experiment with no glucose to see if it was really the glucose reacting or just the potassium manganate and sulphuric acid. We didn't do any repeats. The end point was qualitative, not quantitative. Stirring with rod may not have been consistent/difficult to maintain consistent vigour. There could have been apparatus error - air bubbles in syringe, volumes read incorrectly. Don't just say "human error" without qualifying it! What did you do to ensure that your experiment was reliable? Used fixed volumes of glucose solution, potassium manganate, sulphuric acid/any other confounding variable. Rinsed the pipette out each time - to remove residual glucose solution that would have reacted with potassium manganate/affected time taken to decolourise. (P.S. if you're getting tired of writing out potassium manganate, you can abbreviate it to KMnO4 (the 4 should be a subscript but it won't work on here) Compared end points to glucose solution/solution which had already turned colourless to ensure consistency of end point. How would repeating your experiment have improved the reliability of your results? Anomalies could be identified. Effect of anomalies would be reduced. A reliable mean could be calculated. Do not say "discount" or "ignore" anomalies - the fussy AQA examiners have decided they don't like that anymore, and you'll get marked down for it. A student did a similar experiment to yours, but doubled the volumes of all the reagents. How would this have affected the accuracy of his results? Greater precision/less percentage error in reading from syringe Therefore more accurate/consistent readings of volumes More accurate results To see why doubling the volumes of the reagents would give a greater precision/increased accuracy, imagine you are using a pipette to obtain 5cm^3 of sulphuric acid, but there is a degree of error in your reading: you could misread the volume by as much as 0.5cm^3 in either direction. Now that would give a percentage error of (0.5/5)x100 = 10%. Imagine instead that you are pipetting 10cm^3 of sulphuric acid; once again your reading could be out by up to 0.5cm^3, giving a maximum percentage error of (0.5/10)x100=5%.

Interpreting dataDescribing graphsIf there is a 2-mark question asking you to describe the shape of a graph, the mark breakdown is usually as follows: 1 mark for stating the general shape of the graph (e.g. rises then falls) One mark for stating at what point the graph peaks/levels off/basically a significant piece of data from the graph. This can be read off either the x- or the y-axis. For example, if temperature (degrees Celsius) were plotted against rate (arbitrary units) for an enzyme-catalysed reaction, you could say "Peaks at 38 degrees Celsius" or "peaks at 6.3 units" and either of these would be sufficient for the mark. Practise as much as you can with these questions.Explaining graphsFor these you need to give biological reasons why the shape of the graph is the way it is. In the above example you would say that below 38 degrees C, as temperature increases, kinetic energy of molecules increases, so more enzyme-substrate complexes are formed; 38 degrees C is the optimum temperature; above 38 degrees C, the forces holding the enzyme's tertiary structure together (hydrogen bonds, ionic bonds etc.) are overcome; enzyme denatures; enzyme-substrate complexes no longer formed. (That's far more than you'd need to say to gain all the available marks, but that's the idea.)Standard deviationThis could either be given in a table or as error bars on a bar chart. If one of the resources A, B or C includes standard deviation in the data there is bound to be a question where you will get credit for mentioning it. (The examiner's report on one of the ISAs even says so.)  If the standard deviations/error bars overlap this suggests that the results are not statistically significant. It could be that a > b (where a and b represent mean values of some variable) but if (a-SD) The larger the standard deviation, the more spread out the data.

Evaluating Conclusions

Reliability, accuracy and precision

Interpreting data