Checking experimental data for consistency

Changes with time

any trend is bad; e.g. brightness: change in focus; area: sample changed