Also, in some neap exam, apparently a sample cell is not made out of glass or plastic because they ansorb infrared radiation. But arent the spectra for IR derived from comparisons between a reference cell and the sample cell? The fact that glass and plastic absorb IR radiation would then have no effect?
If your experiment is only to determine a relative quantity, then what you propose should be acceptable.
But, if you are doing very sensitive experiments (e.g. looking at exotic compounds by studying small changes in their UV absorption), the absolute peak heights become important, because they tell us information about quantum yields and other fancy things, which in turn tell us about orbital overlaps, the stability of excited states, and the relaxation pathways. For questions about developing the next generation solar panels, this is incredibly important.
Comparing two almost-transparent cells would give us the relationship between absorbance vs. concentration, but it wouldn't easily tell us how much light the cell has scattered/absorbed. This is especially tricky to deal with if we don't have a totally transparent solvent (e.g. if we use a special organic solvent that also absorbs IR, it will be hard to separate the attenuation caused by the solvent [more technically referred to as the "chemical matrix"], and the attenuation caused by the cell). Therefore, we try our best to use totally transparent cells, so that we can measure the absolute absorbance without worrying about an unknown amount of cell attenuation (there is always a small amount, but we minimise it where possible).
The reason why we use reference vs sample cell is mainly due to possible absorption by the chemical matrix. As far as water is concerned (effectively transparent in the visible spectrum), it seems puzzling why we have to be so careful. But when we actually want to look at the UV spectrum or the IR spectrum, the effects of the chemical matrix/solvent can be quite large.
Re: unidirectional scattering and linearity in spectroscopy,
Going a fair while back here lol. But does this apply to AAS?
Yes. This applies to all of the "usual" types of spectroscopy.
If you continue on with chemistry into Honours/Masters, you might eventually learn about special kinds of spectroscopy in different kinds of polarised light. There, scattering becomes a bit more tricky to interpret. Rest assured, spectroscopy methods in VCE all belong to the usual and simple kinds of spectroscopy.