This is a posting I've been thinking I should write for quite some time. I'm prompted as I again saw this graph of CO2 in parts per million, as measured at the Mauna Loa Observatory in Hawaii since the 1950's, used this week at the Royal Society of Edinburgh by the Lord Kreps.
The first graph is as published by NASA where they show vertical scale from about 315 to 400 parts per million. The second graph, based on "cherry-picking" the annual numbers, is on a scale from zero to 400 parts per million.
When one is told before seeing this graph that we have a problem with carbon and then we see the first graph, the concern is pretty much "proven" without discussion. If one were shown he second graph would not be more difficult to support the "problem" assertion? Neither graph shows the "ideal" number of carbon molecules we should have in parts per million, probably because nobody knows.
All this reminds me of that great book "How to Lie with Statistics".