Lies and damn statistics

Never look a gift horse in the mouth or trust an academic’s efforts at data interpretation.

In a recent study by academics at the University of Texas they found that:

1. For a period from 2000 to 2011 that

2. Large cap US listed firms (Russell 1000 companies) that

3. Were successfully sued by patent trolls

4. Spent, on average, $211m less on R&D and $49m less on acquiring ‘in process R&D’ (whatever that means) compared to companies that were sued, but unsuccessfully

5. In the years (actual number unspecified) following their being sued

6. Also, collectively, the firms that were unsuccessfully sued produced collectively 63.52 and 723.98 more citations. On average? they don’t say. And again they don’t say in what time period

They then went onto state that this data ‘strongly supports the idea that NPEs have a real and negative impact on innovation for United States firms’.

But where is the comparison to firms that weren’t sued at all?

What fraction of the average R&D and IP budget does $260m represent? Assuming it’s, say, 5 years of spend (the study went from 2000-2011), that equates to $52m per firm per year which is probably less than 10%, maybe less, of the total R&D and IP budget of the average of these Russell 1000 firms. I doubt that it is significant.

And what exactly did these companies do with the supposed $260m they didn’t spend on IP and R&D? Maybe they just used it to buy patents or buy into patent funds, and in doing so effectively transferred their R&D and IP development to more effective groups in other companies – large and small, start-ups and universities.

I cannot for a minute imagine that any CEO that is being sued by patent trolls says ‘Right, lets make ourselves more exposed to patent trolls by cutting right back on R&D and patent protection’.

People will see what they want to see in dodgy data.