Feeds:
Posts
Comments

Posts Tagged ‘linear models’

Photo credit: kevindooley, Flickr

As statistical tools for assessing probability and making predictions of the future keep getting more and more common, it may follow that we fall in love with the simple linearity that they typically represent. Especially human characteristics and behavior that are to begin with fundamentally difficult to verbalize, theorize and in the end quantify, are however most typically measured and predicted through linear models. Although this makes aspects of the human existence more accessible in reasonably reliable and valid ways, it is a mistake to assume that the whole can be captured in such a straightforward manner. As this type of information about human characteristics increases and becomes more accessible, the importance of relevantly using and assessing it becomes highlighted.

Information about customers can also be gathered, combined and modeled to a great extent to build detailed and intricate pictures of their needs. However, a product or service designed only based on statistical predictions and segmentation may not capture immeasurable variables strongly influencing behavior such as context, feeling or intuition. Intuition contains vital information for human decision-making but as it is primarily nonverbal, its importance is difficult to assess. In fact, it is difficult to even discuss, which may be the reason it is often deemed worthless. Intuition or emotion are rarely singled out as important factors for example in corporate decision-making. When money is in question, we feel the need to make exact predictions. Statistics provide ways to make assuring-looking predictions and feel security through containing wild human characteristics into neat linear depictions of behavior.

So what to do when there is clear need to create more understanding and more somehow tangible information about human behavior, be it in recruitment, sales or marketing? Test results and averaged statistical data will only make a normalized prediction that at least in part typically misses the personal, unique point. Using test results singly or assuming customer needs only based on segmentation will inevitably lead to mistakes at least at some individual point. The answer is simple: ask people. A human mind is paramount when it comes to combining information and making decisions in context. Customers are the ones to ask when you want to know more about their needs, desires and factors influencing purchase. Interaction and communication makes statistical information come to life and really serve a purpose.

The true value in services like psychological assessment, that aim at creating understanding about human behavior, does not only lie in methods but in human reasoning that puts it all together and communicates interpretation to others. Statistical methods cannot yet fully obtain the complexity and nonlinearity of human decision-making or dynamics of personality. This is not to say that they are useless – well-built methodology provides a way to concretely process information about people, offers a common language for communicating about personal qualities and in the hands of an expert helps create a wholesome picture of a person. The mistake is to use methodology without expertise, or make decisions based on it without listening to feeling or intuition. These two types of information, verbal/quantifiable and nonverbal/intuitive complement each other in enlightened thought.

Advertisements

Read Full Post »