I am performing an investigation on extraction of organic solutes, for example caffeine. With this extraction method, I want to find out whether the organic solvents used actually have an effect, or if the method has error, or the volume used is so small that it won’t affect the yield. As such, we are changing out the solvent.

If the result is significant (after a stats test) that the solvent matters, in my discussion I will include the factors which affected this: particle size, bond type and polarity etc. With these, I can actually create a graph and plot discrete numerical values. However, we can’t actually measure these values - we will take the literature values.

Is there a need to place these in separate hypotheses? Would there be resources to find the correct way to do this?

  • $\begingroup$ I feel a single hypothesis should be enough, as long as it uncovers each aspect and its relation to the dependent variable. $\endgroup$
    – Ronith
    Commented May 13 at 4:56
  • 2
    $\begingroup$ The best thing is to only change one thing at a time and keep all the rest as constant as possible. This makes for lots of experiments but the results will be unambiguous even if the result is a null result. $\endgroup$
    – porphyrin
    Commented May 13 at 7:21
  • $\begingroup$ While it might be suitable for the stats site, questions involving chemometrics are generally appropriate here. However please try to come up with a title that reflects the chemistry problem. Also, you might want to distinguish between two issues: partition and solubility (saturation concentration). It seems you refer to these in the first paragraph using other words. $\endgroup$
    – Buck Thorn
    Commented May 13 at 8:06
  • 2
    $\begingroup$ I don't fully understand the question, but it sounds like you might want to start with the Friedman test. Miller's Statistics and chemometrics for analytical chemistry could be worth reading, too. $\endgroup$
    – andselisk
    Commented May 13 at 9:06


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.