Dear Agung,

What do you consider 'bad data'? How do you know whether your aerosol sample actually contains such 'bad data'?

Although many people talk about 'bad data', this concept is actually irrelevant and useless. Data are generated by a measuring instrument or by a computer program; as such they are the result of a purely mechanistic process. Data are neither good nor bad: they are just numbers.

However, those numbers may be easy or difficult to interpret: If you use an instrument incorrectly, or if your processing code includes bugs, the data may not inform you about what you are looking for. That is your problem, not a data problem.

I suspect that once you clarify what you really mean by good and bad data, you will also find the criteria to separate your data set into items that can be interpreted in the context of an implied model or theory from those items that cannot. Note also that you may learn more from the latter than from the former, at least on the observation protocol or the processing involved.

Good luck with your investigations. Michel.

I think is important we understand what you mean by bad data. I believe if your method of data collection followed standard procedure and the instrument used in the data collection is functioning properly, then you shouldn't talk about bad data. However, there may be outliers in your data due to certain reasons.