There are different assays to study cell viability and proliferation. MTT is one of them and is accepted as a golden standard in this field by many authors.

As alternative methods you can use for example neutral red uptake cytotoxicity assay (NRA) and crystal violet staining (CVS). These three methods (MTT, NRA, CVS) have different cellular/molecular targets and mechanisms of action – that is why in some cases the results obtained by them could be very similar, but in other cases they can differ. If you study the influence of a treatment (a chemical agent for example) on cell viability/proliferation – the nature and biological behaviour of the compound and can explain at least partially the different data obtained between various methods.

But is some cases we can obtain different experimental data even using one and the same method. The reasons can be different – the presence of mysoplams can affect the results obtained; the cells has to be seeded at one at the same density; the monolayer should not be too dense; the compounds tested should be well dissolved and properly diluted; there can be some problems of pipeting; the protocols has to be followed adequately and the methods should be carried out at the same conditions, etc. Good luck!

@ Babak: Thanks. I am actually investigating the compounds with different incubation times. One particular compound has more than 90% killing at 48 hours, around 70% killing at 24 hours and hardly any killing at 6 and 12 hours. I am a bit confused. However, the most important thing is my readings at specific time (like 48 hours) has large fluctuations. Fortunately, I use six well for MTT assay so I can delete outliers and still have scientifically acceptable 3-4 cells to take the average. 

That is the reason why I was concerned so much about the variation.