Dennis D. McDonald (ddmcd@outlook.com) is an independent consultant located in Alexandria Virginia. His services and capabilities are described here. Application areas include project, program, and data management; market assessment, digital strategy, and program planning; change and content management; social media; and, technology adoption. Follow him on Google+. He also publishes on CTOvision.com and aNewDomain.

More Thoughts on Linkedin's One-Click Skill Endorsements




By Dennis D. McDonald

Click or tap image to download .pdf of this article.In last week’s blog post My Thumbs are Mixed on Linkedin’s New Skill Endorsements I expressed doubts about the new Linkedin one-click endorsement feature. While I like the granularity of endorsing for specific skills, I also have concerns about the validity of the data given the lack of control over how the data are generated, analyzed, and used.

Des Walsh commented on that post and said,


Thoughtful and nicely balanced comments, Dennis. As you say, it doesn’t substitute for “a heartfelt and knowledgeable endorsement (what LInkedIn calls a “recommendation”) by a trusted colleague, but at the very least it enables us to provide some recognition of people in our network for whom we may not feel quite ready, or knowledgeable enough, to write a more detailed recommendation. That would apply especially to those who have chosen to connect with people they don’t know well but see as having some specific skills, e.g. through their writing.

My response to this was that it does make sense to distinguish between “strong endorsements” and “weak endorsements.” Many connections via social media are between people whose knowledge of each other may be quite limited. Viewed this way a set of endorsements is one way to provide a Linkedin connection with a set of data that reflects what that person is known for without having to go into detail on quality or evaluation. I can see the benefits of that, as long as one is aware of the limitations of the data.

Alex Joseph on Google+ had a somewhat different take on my concerns about the one click endorsements:


Good points and agree the system is never perfect, but Linkedin is counting on the power of large numbers and signaling value by the peer-reference system. Some will always try to game the system, but it gives a directional sense - just like any academic credentials that denote a high probability of competence, not a guarantee.

Alex’ use of the “large numbers” concept is significant. Single click endorsements do offer more data to the mix of what can be discovered given the variety of “signals” available about an individual via Linkedin — job history, education, self-described specialties and accomplishments, etc. Where I stumble on the “large numbers” argument is that the value really depends on the quality of the numbers and the perspective of the person doing the analysis.

Viewed statistically, in any population of ratings for an individual made by a group there is bound to be a mix or distribution of reliability for the scores, with some scores made by people who really do know the connection and with others made by people who don’t really know. In general there is a tendency to think that such variations in quality or reliability will “average out” and that, given a data set of a certain size, the distribution of scores for a given skill will in fact reflect some level of relative competence in comparison for the group. For those individuals who have limited their Linkedin connections to people they actually know, that may well be the case.

In certain cases Linkedin’s value is not in terms of how it can be used to organize and characterize a group of individuals but in facilitating a connection between pairs of individuals. If it’s just two people the data about them needs to be accurate. Just as we want the names, experience, education, and other facts to be correct, we also want to make sure that the endorsements they have are also accurate. For these two individuals, “large numbers” are less relevant than the reliability and trustworthiness of the data. Given how Linkedin is set up that’s difficult to assess.

I’m not suggesting that accuracy and trustworthiness have never been concerns with Linkedin. After all, a large number of my own connections, built over the past several years, are with people I’ve never even met and for whom I would find it difficult to make an endorsement of a specific skill. There’s nothing wrong with that; that just means that for many of my Linkedin connections I would never attempt any skill endorsement.

The flip side of this is that I need to take seriously what I say via a one-click endorsement about the people that I do feel I know well enough to endorse. The value of this is that I am now in a position to focus my endorsements on the people I really do know, and that’s probably a good thing.

Copyright (c) 2012 by Dennis D. McDonald, Ph.D. Dennis is a Washington DC area consultant specializing in collaborative project management and new technology adoption. His clients have included the US Department of Veterans Affairs, the US Environmental Protection Agency, Jive Software, the National Library of Medicine, the National Academy of Engineering, Social Media Today and Oracle, and the World Bank Group. Contact Dennis via email at ddmcd@yahoo.com or by phone at 703-402-7382.

The "Open Movement" Continues: Big Pharma Company Opens Up Access to Clinical Trial Data

The "Open Movement" Continues: Big Pharma Company Opens Up Access to Clinical Trial Data

My Thumbs are Mixed on Linkedin's New Skill Endorsements