Suitability of the investigative procedure to answer the question being asked. e.g. an investigation to find out if the rate of a chemical reaction depended upon the concentration of one of the reactants would not be a valid procedure if the temperature of the reactants was not controlled (q.v. Variables: Control).
Data is only valid for use in coming to a conclusion if the measurements taken are affected by a single independent variable only. Data is not valid if for example a fair test is not carried out or there is observer bias.
e.g. Valerie and Vernon are investigating how the length of a cylinder of putty effects its electrical resistance. However, to make the putty a different length they roll it out more thinly! This means that the putty's length increases, but in addition they alter the width of the putty. This last factor also affects the resistance, so their data is not valid - they have not carried out a fair test.
A conclusion supported by valid data, obtained from an appropriate experimental design and based on sound reasoning.
In other words, a valid conclusion can be made only if the second paragraph above holds true.