Totally crazy off the side thing here (i'm not an astronomer either) but I've been reading recently that were not quite sure how even the universe is anyways so space could be stretched, warped, curved or otherwise deformed between us and another galaxy which could throw things further off too. The effect might be subtle or it might be larger but sometimes I wonder if we really know.
I hadn't thought about it but it is wild to consider that light from one side of a distant galaxy arrived thousands of years before light from the other side of that galaxy. Man that broke my brain for a moment. The universe is so very difficult to comprehend.
Actually, we have the "evenness" of the universe nailed down pretty well. If it weren't fairly even, we would expect to see differences in how things looked when we looked at different areas of the sky. Since we don't see anything of the kind (instead, the universe looks roughly identical whichever direction you look), we can safely assume the universe has pretty much the same curvature everywhere (other measurements show it to be flat or nearly so as well). We can also obtain some idea of the curvature by looking how distorted a galaxy appears. Since we have plenty of nearby galaxies to compare with, we can be pretty sure when we're looking at one that is being gravitationally lensed or what have you based on the statistics.
One comment on the error estimations, every University seems to have their own favorite techniques. I mainly used summed absolute values instead of summed squared values. On top of that everything had to be rounded by a rather arbitrary 15 significant unit rule. I have a suspicion why they did that though, the laboratory equipment was antiquated but at least it taught the error estimation rather nicely. Each of the techniques has its merits and drawbacks. Though I haven't used the 15 significant unit rule after graduation for sure.
15 sig figs? How goddamn accurate were your equipment and measurements?! That's just about the most amazing accuracy I've ever heard of (the most accurately determined value in physics, the g-factor of the electron, is known to something like 20 sig figs). I think I managed six or seven in my spectroscopy lab. For the absolute value sum, what kind of error were you calculating? The sum of squares is mathematically the only way to do it for different factors correlated in the way I described.
From the Optics point of view, it would be interesting to know if gravitational lensing exhibits similar kind of aberrations as the normal thin lenses do. Three arcsecs doesn't sound much, but on the other hand we are talking about astronomical stuff here and those are pretty big numbers. So I can't yet tell if from a long distance way it could be significant, but as I said Astronomy is a field of Physics that I skipped in the University.
A shift of 3", while noticeable, isn't necessarily that big. Most systems studied in astronomy are at least several arcminutes wide, so a few arcseconds is at least an order of magnitude smaller, sometimes two (compared to the angular diameter of a big galaxy cluster, for example). As for aberrations, maybe; I don't know. They would be described differently than optical ones, however, because a gravitational lens' power increases the smaller the impact parameter is, the opposite of what an optical lens will have.
Then a question about the spectrometric measurements. Some spectrometers that I have worked with required a calibration marker, a short spectrum that is definitely known, preferably slightly outside of the spectrum of the object to be measured. Does it work the same way in astronomy too, and if it does, what is the spectral reference marker used in astronomy? Is it in the object to be measured or a completely different measurement?
You always need a calibration spectrum to determine radial velocities (Doppler shifts). What is used depends on the accuracy required and the application. My current research is on improving planet detection, which currently uses iodine as a reference spectrum (if this results in a published paper, I can let the board know if people end up being interested enough). In order to eliminate systematic errors resulting from unequal optical paths taken by the reference and observed spectra, the iodine is placed in the optical path of the spectrometer, just after the light enters it. This superimposes the iodine's spectral lines, which are NOT shifted in any way (very important thing for a reference!), on the observed spectrum. Measuring the distances between all the iodine lines establishes a wavelength scale* for the spectrum. Then we can measure the positions of each line in the observed spectrum on the CCD, thereby getting their wavelengths, and compare this to the wavelengths as measured in a lab on Earth. This tells us the Doppler shift of the object, and thus its radial velocity with respect to us.
For less demanding applications, a thorium-argon emission lamp is used which has a separate optical path to the CCD. The analysis proceeds as above. However, large systematic errors are incurred because of the differing optical paths used by the reference and object's spectrum. Some spectrometers eliminate this by using fiber optic cables to equalize the optical path length of both systems, which brings the precision up to levels comparable to that given using iodine. Right now, the best precision achieved is something like 3 m/s, IIRC (it might be as low as 1 m/s).
For just determining what an object is made of, no reference spectrum is necessary when the exposure is taken. Because the pattern of lines will always be the same no matter the Doppler shift, the pattern in the observed spectrum is just compared to patterns in spectra taken in labs on Earth.
*The wavelength scale is just how many pixels correspond to a given difference in wavelength, so it could be given as something like 5 pixels/nanometer or something (don't know a typical one off the top of my head, sorry).
Third thing that bothers me now is the effect of dust on the apparent sizes of stellar objects.
Almost nil. There isn't enough of it out there to do anything really significant.
