Observing the evolution and challenges faced by Klout, a venture designed to measure the on-line influence of an individual or organization, has been nothing short of fascinating. Changes in the Klout algorithm (and its impact upon scores) can send the world of social media into a frenzy. The overriding opinion seems to be that you shouldn’t “mess” with the numbers, right?
But, that is absolutely wrong — in fact, you should.
I have a very different perspective on Klout’s struggle to develop into a meaningful measure. I find the struggle to be quite predictable. Probably because where I come from, when a new construct and its measurement are proposed, it often takes a very long time to determine true value and identify prudent uses in the real world.
On one hand, the outcry that resonates after a scoring revision is an excellent sign. It lets us know that Klout, at the very least, was actively being considered as one measure of influence. On the other hand, it becomes obvious that the scope of the development phase, may have needed to be more controlled to allow for necessary iterations occur.
In psychology, the development of a new construct is an important and often long-winded process. However, when you consider the importance of measuring concepts, such as intelligence and motivation, the development of that construct — and its valid measurement — are paramount.
All in all, you must tread quite carefully.
It may be useful to view Klout in reference to a few traditional elements of construct development:
- Does Klout have Face Validity? In other words, does the idea and its components seem to make logical sense.
- Does the measure demonstrate reliability? In other words, can the measure show consistency.
- Does Klout possess Content Validity? Do the components that make up the measure adequately represent the elements of influence.
- Does the measure have Construct Validity? When you look at scores on the measure, the scores should jive with other key markers of influence. (Convergent and divergent validity.)
There is another point to briefly consider here. Klout may be measuring “Potential to Influence” and not “Influence” itself. We simply do not know. As with other constructs, such as job satisfaction and its relationship with turnover — Klout scores may signal an impact on attitudes, yet the relationship with behavior, is not a causal one.
Time to Mature
All in all, Klout has to be allowed the time to develop fully. The algorithm should be subject to changes and iterations, as the organization sees fit, to adequately develop the measure.