Archive for the ‘Environmental Data’ Category
Originally published in July 2005. Edited slightly to accommodate now-dead links.
From Jorn Barger’s Robot Wisdom Weblog I linked to a great article by Edward Tufte about how the findings of primary studies gradually lose their power and meaning, when they become repackaged and redistributed by secondary organizations (journalists, public relations firms, think tanks, nongovernmental organizations, governmental agencies, and on and on – see the posts from the previous couple of days):
In repackagings, a persistent rage to conclude denies the complexities, ambiguities and uncertainties of the primary evidence. A substantial selection bias also operates: news wins out over olds, as recency of evidence decides relevance of evidence.
You must see the original to get the subtext behind the phrase “rage to conclude”, which is a wonderful quote from Flaubert.
There’s no better way to start the New Year than getting a jump on it. I don’t know where I found this, but “Five Easy Lies” is a wonderful example of how to sow doubt and uncertainty about any set of data. I have written before about how sowing doubt and uncertainty with environmental data is a time-honored tradition of opponents of toxic substances regulation, climate change denialists or anyone desiring to use “too much uncertainty to make a decision” as a strategy for deferring any type of economically or politically painful decision. Stated more eloquently on SKAPP’s web site:
“Doubt is our product,” a cigarette executive once observed, “since it is the best means of competing with the ‘body of fact’ that exists in the minds of the general public. It is also the means of establishing a controversy.”
The lesson about how to talk about data that isn’t telling your story works better with the visuals (visit the site if you’re interested), but here’s a summary of the principles:
Select your cutoffs – focus on just the piece of the trend or the piece of the data set that’s most favorable to your side of the argument.
Talk about the trend of the trend – very useful for engaging people who are bad at math. If there is still an increasing (or decreasing) trend in the data that isn’t helpful to your side of the story, be sure to talk about the rate of the increase or decrease. If this fails, as the post notes, “[k]eep on differentiating until you find a curve that matches your needs.” If that fails, transform the data until is resembles something that’s helpful.
Talk about the different phases – focus on the changes in the data trends – any data set will have it’s moments when it will support your side of the story, even if the weight of evidence is against you. Make sure that noone looks closely at the magnitude of the different data trends.
Focus on outliers – there’s always a case that is not readily explainable based on the preponderance of the data, especially if it’s a noisy data set. This is more easily done if you ignore error bars or other measures of data uncertainty.
Sow confusion – combine any or all of the above to increase doubts about the data set.
I found this to be such a resonating statement:
Evidence is your friend. More evidence means more cutoffs to choose from, more trends to analyze, more phases to count, more outliers to discover, and more confusion to sow. Be careful to disguise the fact that you and not the data are the source of the confusion.
We’ll talk another time about how to criticize the methods use to collect environmental data, as a technique for sowing doubt.
Happy New Year.