Posts Tagged ‘Statistics’

Why do leaves change colors?

October 4, 2009 8 comments

We are smack in the middle of my favorite time of year, autumn in New England. Cool crisp mornings, right temperature days and a painted landscape. Thinking that I needed to contribute to my blog, I wanted to figure out how to best work in (and share) a New England landscape.

mitch lieberman - champlain valley fall 2008

In an attempt to keep up on my reading, I find it difficult to escape the cacophony of ‘advice’ about how you (as a business) must jump on the Social Media bandwagon and of course how you should go about it as well. I do believe in the power, but these initiatives do have a cost (people, technology…) What seems to be lacking are the scientifically based details which illustrate the return on these investments of time and money.

Yes, there are some proof points (the exciting stories which everyone points at), and we all have a visceral sense on where things are going – but each business is different. As we undertake these ventures (adventures?), we need to be sure which ones are working, and why they are working and make mid-course corrections as necessary. Anecdotal evidence will only get us so far…

Putting the two together

Autumn in this part of the world is quite colorful, nature puts on quite a show. The question is why do leaves change colors? Allow me 30 seconds to make sure we are all on the same page, in regards to what is actually happening:

During the growing season, chlorophyll (responsible for energy production and the green color) is continually being produced and broken down by leaves. In the Autumn, the chlorophyll production slows down and then eventually stops altogether. Other pigments, which have been present all along, now have a chance to show themselves.

Why does this happen?

If you look at the first chart, it is clear that it must be because the temperature gets colder which causes the production of chlorophyll to slow, then stop.  Here is a chart (scientifically done in Photoshop)  that shows a graph of the average daily temperature in October (for Burlington, VT) against the percentage of leaves which have changed color (Peak being 60-100%):

Proof, right?  There is a clear and direct correlation between the daily high temperature and amount of colors displayed by the leaves.  But, wait, what about the chart below? I did a similar exercise (equally scientific)  charting the amount of daylight, as the calendar moves through the month of October.

The amount of daylight hours against the percentage of leaves which have changed color:

The real answer is believed to be that hours of daylight is the determining factor, but that temperature plays an influencing role. But, this was not really the purpose of this note.

Causality and correlation are where many of us get ourselves in trouble. The first chart shows a correlation between temperature and the leaves, as does the second chart – but only after further experimentation which isolates one from the other can you determine cause and effect – causality. You can find many examples of correlation, some presented in a joking manner, some trying to actually show proof of cause and effect.

As you begin to experiment in Social Media, or Social CRM – be very clear on what you are measuring and what other variables might be impacting your assessment. 100% proof may or may not be attainable – if this were the barrier, we might not try anything, this is not my point. My key point is to go in with an open mind, but clear and measurable objectives.