Terry Anderson remains optimistic about learning analytics, but the two concerns he cites in this short post are more interesting (quoted):
- The first challenge stems from the ownership of this data... This is especially egregious when the data collected and analyzed is not made available to those who created it.
- Secondly, as researchers at Microsoft Research danah boyd and Kate Crawford noted, learning analytics currently suffers from the delusion 'that large data sets offer a higher form of intelligence and knowledge... with the aura of truth, objectivity, and accuracy.' (but the) belief in 'truth, objectivity and accuracy' is far from proven.
The thing with big data, to my mind, is that it is shallow data. It captures details about many users, but only their use of one or a few applications. Getting deeper knowledge would require more and more egregious breaches of privacy. But so long as the data remain shallow, it remains prone to confirmation bias, as researchers find in the data what they data were designed to show.
Today: 3 Total: 3 [Direct link] [Share]
