Quantifying Observation

As I went through the lab reports that students have prepared for the solubility challenge, I was impressed by the huge task we scientists have of translating observations into meaningful numbers. I have done a solubility laboratory during first semester organic where students classify the solubility of a compound in a certain solvent at a given temperature as insoluble, somewhat soluble, soluble, or very soluble. This is, of course, very subjective and student “observations” tend to go all over the place. This solubility experiment was about quantifying solubility beyond an observation. The underlying supposition is that observations are cheap and unreliable. Numbers, however, can be checked and need to make sense. In this particular experiment, I could not publish about half of the students results because key data was missing or the numbers didn’t make sense the way they should. I believe that most of this was a result of my not giving my students a “recipe” to follow in doing the experiment. There own lack of confidence and tentative understanding of laboratory process worked against the collection of useful information. The discussion of numbers leads to the issue that the value of student research is tied to the ability to validate results (numerical or otherwise). Hopefully, the statistics people can help us make sense out of data proliferation which may not always be entirely accurate or particularly precise.



  1. Posted February 10, 2009 at 6:54 am | Permalink

    Brent – thanks for participating in this! Looking over the results it looks like many of the solubilities are well below what others have published. The students didn’t seem to record how they made “saturated solutions” – they just reported that they did. We have found that it requires time (10 minutes vortexing after the last addition) to get reproducible solubilities. For larger batches like you were using it might be easier to heat to get the compound to dissolve then let the solution cool to room temperature to ensure saturation. It was still a very valuable first attempt.

  2. Posted February 10, 2009 at 7:36 am | Permalink

    Yes,this is interesting, particularly in the light of how one might go about doing it again. Should we encourage classes to read through the existing material to get information, or let them run at their own way without guidance? Would the latter work better with at least two sessions so they can go back and re-do measurements rather than just find out they are probably a bit low? Nonetheless useful to have data even if it is just a lower bound.

  3. Posted February 10, 2009 at 8:38 am | Permalink

    It certainly might be helpful to have students look at all the different techniques reported in the full list of experiments
    then maybe have a pre-lab group discussion.

    If they looked at the measurements labeled DONOTUSE via Rajarshi’s query tool:
    they could look at all the reasons data were excluded and try to avoid the same problems. That would be a great example of using “failed” experiments productively.

Post a Comment

You must be logged in to post a comment.
%d bloggers like this: