Noisy data

As an example, Convolution-type digital filters such a moving average can have side effects such as lags or truncation of peaks.

It can be caused by human error such as transposing numerals, mislabeling, programming bugs, etc.

If actual outliers are not removed from the data set, they corrupt the results to a small or large degree depending on circumstances.

If valid data is identified as an outlier and is mistakenly removed, that also corrupts results.

Fraud: Individuals may deliberately skew data to influence the results toward a desired conclusion.

In this example of an outlier and filtering, point t2 is an outlier. The smooth transition to and from the outlier is from filtering, and is also not valid data, but more noise. Presenting filtered results (the smoothed transitions) as actual measurements can lead to false conclusions.
This type of filter (a moving average ) shifts the data to the right. The moving average price at a given time is usually much different than the actual price at that time.