Question: How do I make sure the "YELLOW LINE" threshold is the same for each measurement with the spectormeter (trying to maintain consistent conditions for experiments across multiple days)

I suspect there are two issues here: 1) the software interface and 2) mechanical instability. On #1, I believe the web-based software always needs to 're-zero' where to take the single-line of image data within the video frame. On #2, assuming you are using the paper-and-velcro PLab spectrometer, there simply is not sufficient mechanical rigidity in the design to provide the repeatability you require. (Might be helpful: https://publiclab.org/notes/stoft/03-02-2015/plab-3-spectrometer-upgrade-prototype )

I think stoft (see above) has some good suggestions.

Calibrating more often is also a good idea.

Better standards, Holmium oxide for example, would be ideal. They just aren't affordable. The cfl lamps have to do as a std.

@Ag8n true, calibrating before a measurement (or series of) helps assure the measurement data is referenced. As per wavelength references, while there are better standards, the bandwidth of web-cam / DVD spectrometers is, at best, about 1nm (while the 'typical' paper/velcro build might be no better than 5nm repeatability) so using 3 spectral lines from the simple CFL is generally sufficient.