I had this discussion again and again. Some minutes ago, I had it again. Now that’s too often. Here is a little correct — incorrect cheat sheet. I hope it helps making the world a better place. (I know it doesn’t.)
I’m asked to carry out an experiment / measurement for educational purpose.
- Browse the web for a step-by-step instruction and follow it step by step.
- Try to understand what’s to be measured and think of a smart way to achieve this. (Literature research can really help!)
I got a result.
- Go home and have some beer.
- Check the result for plausibility and repeat the experiment to become sure of reproducibility.
I get no plausible result or the results don’t fit together.
- Repeat the experiment until the end of day as often as possible.
- Search for possible errors in the experimental setup and repeat the experiment with them removed.
I have an awful lot of data showing a realistic mean value but there are some quite weired results in it and the standard deviation is ridiculously high.
- Choose three or four values that fit nicely together, make the highest one a little lower and the lowest one a little higher and drop the rest of the datapoints.
- Run an outliers test on the data list, drop outliers for further evaluation of the result but keep them mentioned in the report. (At least their number.) If the standard deviation is still too high, carry out another experiment and—if nothing helps—point out the high uncertainty in the discussion of the results.
I can’t do any better within my possibilities but simply can’t obtain a good result.
- Try to have a quick glance at the lab assistant’s notebook for the “correct” results or look it up elsewhere. If nothing helps: Guess some “better” result and write the report using this data instead of the actually measured values.
- Write the report as best as you can pointing out that the experiment failed and discuss why this might have happened.
The lab assistant doesn’t honor my honest report and only focuses on the missing result.
- Remember and be less honest next time.
- Stab down lab assistant.
But my colleagues, who cheat on their results, get good marks and the assistant asks me, why I can’t do as good as them.
- Learn from them to improve your marks.
- Talk to them. If that doesn’t help: Try cold water. If nothing helps: Stab down colleagues.
Why should I do all this?
- Who said you’d have to?
- Wouldn’t that be much more appropriate for you than studying science?