Tags

, , , , , , ,

When I started out with psychographic text analysis a few years ago sentiment analysis was the craze. If I told anyone about the idea to analyze the psychological traits of someone online most (action oriented) people directly asked me about if I could do sentiment analysis. Some people still do. The thing is that I´m tremendously more interested in the psychological state of a person, e.g. the mood, than what the opinions they express about something. Today, finally, psychological text analysis is starting to come on the radar for a broader audience primarily due to the hedge fund Derwent Capital´s innovative approach to financial prediction using mood analysis of Twitter and sociologists researching the correlation between time of the day and mood on Twitter.

Not everyone will be happy with this development - in some cases for good reason I believe. Foto: Flickr/futureatlas.com

There is of course lots of innovation going on in the area since especially Google and Facebook target ads based on characteristics that fall into the domain of psychographics (interests), rather than good old demographics. Their approaches are, for what I can judge, not based on any a priori model, but instead based on large amounts of data points. The difference from a scientific point-of-view is whether you use an inductive or a deductive methodology. Inductive methods require lots of indata and lots of statistical computing power, something Google and Facebook excel at. They´re so good at it that some people even think they will put an end to deductive methodology in science.

Deductive methodology requires insightful imagination first – and then lot´s of indata and statistical computing power to validate the theory. The inductive methodology might end up with a model if the data shows consistent patterns, but doesn´t really need to bother with that. The deductive methodology, however, stands and falls with the predictive capacity of the model choosen.

I´ve chosen a model of the mind inspired by philosopher Ken Wilber to form the basis for my deductive approach to predict interest (or attention) online. The upside is that it allows for multiple models to form a coherent whole and that there is some research about how language reflects parts of the model. The rest I have to invent myself, which of course is the downside – it will need lot´s of more people with more knowledge and resources than me to develop and validate the methodology. I find it very interesting a project however, so I´ll carry on towards at least a large scale test of my hypothesis that interest can be predicted by the value words used online.

My methodology involves three psychographic levels of detail (and complexity) and my guess is that they will be of public interest progressively over the coming years. All of the three levels of psychographic text analysis will be researched and applied within psychology (individual profiling, health monitoring, intelligence profiling etc), cultural studies (culturomics, social network analysis etc), social sciences (financial prognostication, public opinion research etc) and in micro-level interest prediction (ad and content targeting etc).

1. Mood – where we are today. Lots of more applications to come, especially if the financial analysis Derwernt Capital and AmalgaMood is doing turns out to work.

2. Values – “serious society” is not there yet, but with the foundation laid out by small and sometimes far-out research initiatives such as the Human Memome Project.

3. Personality type – personality type or temperament doesn´t seem to be very popular in academia today due to the shift of focus towards the succesfull neuroscientific explanation models. My guess is that at the end of that line of study personality types will emerge again – and then those will be used to segment people psychographically online for the same reasons as above.

Time will tell if I´m guessing right.