Buzz (aka “sentiment”) – how do you measure it?

There is a glimpse – but only a glimpse – of the future in this post by the BBC’s Rory Cellan-Jones on new ways of measuring new media political comment on the leaders TV debates during the election. The bar graph below shows an example of Twitter sentiment during the the third debate, on the economy.

Lexalytics on 3rd party leaders debate

Lexalytics on 3rd party leaders debate

Two thoughts occur straight away.

1) The methods for measuring the flows of messages, tweets and posts are going to get much better pretty quickly. Since networks are potentially influential, people trying to track and explain opinion changes will have to analyse what happens in social networks such as Twitter. The early techniques listed just sound a bit primitive.

2) The basic reason behind opinion polling is that the science and experience behind it makes it more likely (but not certain) that it will be right. Cellan-Jones is puzzled that one measurement doesn’t match the opinion polls. It’s a fair bet that paid political spinners are responsible. Any means of communication, new or old, can be fixed to produce a distorted result by someone sufficiently energetic and determined. More careful sampling of new media opinion is more likely to produce better results.


Tags: , ,

Comments are closed.