Guidelines for interpreting data

First of all optimise the data. Data should be expanded to a scale where transitions of interest can be clearly seen in the context of the trace. It may be helpful to slope the data so that areas of flat response are shown as flat. The eye can interpret more easily from the horizontal. The process of sloping simply pivots the data graphically it is not a curve fitting or smoothing process so has no effect on transitions or calculations performed. Sometimes transitions are missed simplybecause they are very small compared to a major transition. Inspect the whole trace carefully if looking for small events. Remember, if a transition cannot be repeated it is imhkely to be real. If in doubt repeat the analysis. 

In addition, find out as much about the sample as possible, it is difficult to set up a good method or to give clear interpretations of a trace when working without full information. The sort of information commonly needed is an idea of melting point or decomposition temperature, but more fundamentally it is better to have some idea of what transitions to look for and why. For example  

What data are available from complimentary techniques, e.g. TGA  

Does the event look real Thermal events are not normally excessively sharp. 

Examine the sample pan and look for sample leakage, decomposition or bursting which may have influenced the trace. In some cases, it may be useful to reweigh the pan to see if sample has volatilised and escaped. 

---