[spirochete]

Does running a reaction on larger scale, ie going from 1 g to 20 g, always result in at least a slightly lower yield?  Is there any reasonable way to compare yields from different procedures when they are run on such different scales?

[azmanam]

I typically see the opposite effect.  consider: mis-measuring a reagent by 0.1 mL or 1 mg on a 0.025 g scale is a large percent error.  If you're measuring a strong nucleophile/base, then a large percent error in measuring reagents can lead to a lot of byproducts/poor yield. 

taking the same care on a larger scale means a much smaller percent error.  0.1 mL or 1 mg off on a 20 g scale is almost negligible and will often not corrupt the reaction much.  In my observation, yields go up with scale.  as long as you're as careful with purification as you are on small scale.

[alphahydroxy]

Yeah, typically I would see an increase in yield when scaling up...

I think going from 1g to 20g is reasonably comparable, much moreso than say 50mg compared to 5g.

Lower yields may be a result of practical issues - things like imperfect stirring or dissolution, or perhaps requiring a bit more time to go to completion.

[ARGOS++]

Dear spirochete;

I don't like to make it more complicated as it really is, but in quite a number of such situations not only you observe also a lower yield when enlarging the "scale". Mostly not in the lower "Gram-Scale".

There is much more involved during "Scale Up" as only multiplying the number of moles and voluminas with the same factor;  therefore Mr. EugeneDarkin is able to tell/sing you the melody about "Scale Up"  and "Dimensionless Numbers", what is an own specialized field in science. You can also start with searching the Internet for.

So be not surprised if you once get a higher and in another reaction you get a lower yield in %.

Good Luck!
                    ARGOS⁺⁺