I am carring out practical work to find a mathematical equation between kVP and the backscatter factor.
This is not only a function of kVp, but rather of "beam quality", which can be described by kVp AND filtration (or by HVL). Additionally, backscatter for the same beam will be different for different materials. See Appendix VIII in the IAEA report 457 - one table is attached, you can download full report from the IAEA site:
http://www-pub.iaea.org/MTCD/publications/PDF/TRS457_web.pdf
EDIT: additionally, backscatter will depend on the field size.
it depends on the material the beam interacts with. if the material is of high atomic number, then the lower the kv, the more photoelectric interaction (absorption) and less the compton scatter. For low density matter such as fat, then the higher the compton scatter that will occur. general law of kv and scatter is that as kv increase so too does compton scatter and photoelectric interaction (absorption) decreases.
A short remark to Darrions post. Your description of compton scatter is ok, but you miss that compton photon distributions show an angular distribution and that classical scattering will also occur. So you and Taha depend on experimental data like in the really matching literature tip of Witold. And beam quality just desribed by HVL or kV and filtering are a shortening of the real more complicated situation. Backscatter factors additionally depend on the thickness of irradiated material.
In order to carry out your work, as long as the other factors mentioned here remain constant then obtaining a mathematical relationship should be relatively straightforward. Remember to control for field size, mAs, incident material, secondary incident material (eg is there a wall just behind your target), filtering and HVL.
Once all other factors are controlled for then simply varying kVp will allow measurements of backscatter to be undertaken.
Bear in mind also that the location of the detector will affect how much backscatter is detected (angle from the incident beam).
Depending on your resources you could do multiple experiments to assess these relationships individually, but also be aware that for some of these factors there is an interlinking effect, eg increasing kVp will increase backscatter but it will also affect the scattering angles as well so there will be two factors changing from the one exposure change.
Good luck.