Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study
I am a lifelong project manager with some 40 years of project management experience under my belt, most coming from construction project management.
Over the years, I had noticed that some people are just naturally better project managers than others. Looking back over the years, it didn’t seem to matter whether they were engineers, nor did it matter if they were men or women, and having spent most of my life working around the world, it didn’t seem to matter what country they called home or what language they spoke or how they worshiped their God. It also didn’t seem to matter what astrological sign they were born under, nor did it appear obvious that those who were naturally good carried lucky talismans.
And most certainly, it didn’t matter if they did or did not have their PMP, PRINCE2, MBA or PhD behind their name!!
So what was that elusive “something” that made some people just “natural” project managers?
As I began my PhD research, trying to answer the question “Is project management a profession? And if not, what is it?” I intended to include part of that research a chapter on behavioral attributes, but as with most projects, time and quality constraints won out and I had to “descope” and the part that got descoped was the research about behavioral attributes.
Behavioral Attributes Rise From The Dead
But that was only a temporary diversion, and now, PhD in hand, I am resurrecting my interest in the behavioral attributes.
To start with, I relied on previous research done by my good friend and mentor, R. Max Wideman. Max chose to use Myers Briggs, and his research proved not to be sufficiently detailed for the work I had in mind. http://www.maxwideman.com/papers/profiles/myersbriggs.htm
So my quest for something more granular finally turned up a Dr. Dan Harrison, and his Harrison Assessment. http://www.harrisonassessments.com/ Unlike Myers Briggs or Kiersey, the HA Instrument tested for some 155 different behavioral attributes. Furthermore, HA has a feature that measures the CONSISTENCY of the responses, which provides and accurate measure of how truthful the respondent is being, or are they trying to game the system. Having found what I was looking for, I contact Dr. Dan and he suggested I contact his regional representative, Mr. John Suermondt, john@harrisonassessments.com and work with him at least in setting up a pilot. John is originally from the Netherlands, a former commercial diver, now living in Perth, Australian and a really dynamic and cool global kinda guy. And with over 19 years working with Harrison Assessments and was just as excited as I was to pilot this.
All of the participants in the pilot study came from people in the various in-house classes that I teach for our Fortune 500 clients. These classes were either the PMP or CCC/E Prep, or in my graduate level university classes at either ESC Lille Masters of Science in Project Management http://esc-lille.audaxis.com/en/Programmes/MS_MSc/Project_Management_Supply_Chain_Organisation/Specialised_Master_in_Project_and_Programme_Management or the University of Western Australia’s Masters of Energy Systems or the Masters in Petrochemical Engineering degree. http://www.blendedlearning.ecm.uwa.edu.au/
Project Manager Pilot Group
What I did was select a pilot group of 28 practitioners who were deemed “successful” project managers. In order to be deemed “successful” they had to pass three tests:
- They had to hold the job title of “Project Manager” in their company
- They had to have demonstrated to me in the classroom environment that they had exceptional leadership skills (top 5% of the class) and
- They had to have at least 5 years of working experience
This initial pilot study group of 28 consisted of:
- exactly half men, half women;
- about 30% were Muslim
- about 10% were Hindu or Budhists
- remaining 60% were Christian
- 9/28 = 32% Asian
- 6/28 = 21% North American
- 5/28 = 18% Australian/New Zealand
- 5/25 = 18% European (including Eastern Europe, Northern Africa and Turkey)
- 3/28 =11% Central or South America
The industries they represented were:
- Oil, Gas or Mining 9/28 = 32%
- Telecommunications or IT 9/28 = 32%
- HR, Sales or Marketing 5/28 = 18%
- International Development 3/28 = 11%
- Finance 2/28 = 07%
The test was facilitated by John and administered on line (it only takes about 20 minutes) and in all cases, it was conducted in English, although the instrument has been translated into some 15 languages.
Results
Tune in tomorrow for the second part of this series, with the results including predictors, desireable traits, and undesireable traits for project management success!
Behavioral Profiles of SUCCESSFUL Project Managers
- Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study
- Behavioral Profiles of SUCCESSFUL Project Managers- Results
Jun 30th, 2009 at 9:45 am
Twitter Comment
RT @pmstudent: Behavioral Profiles of SUCCESSFUL Project Managers- research study: [link to post] #pmot #projects
– Posted using Chat Catcher
Reply
Jun 30th, 2009 at 11:04 am
Twitter Comment
PM Student: Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study [link to post]
– Posted using Chat Catcher
Reply
Jun 30th, 2009 at 11:22 am
Very much looking forward to reading the research.
Coincidentally, I just arranged for Jennifer Tucker, author of the new book: Introduction to Type and Project Mgmt http://response.cpp.com/?elqPURLPage=102 to speak on the July 14 call of a professional community I coordinate: Type Coaching SIG – your readers would be welcome to join: http://groups.yahoo.com/group/typecoachingsig/
I’m impressed with Jenny’s book and glad to read additional works on personality type, behaviors, and project management.
Reply
Jun 30th, 2009 at 1:17 pm
Dr Paul,
I think this is an extremely important Did you review PMI’s Project Manager Competency Framework.
http://www.pmi.org/Resources/Pages/Library-of-PMI-Global-Standards-People.aspx
I also found a number of interesting entries on google under project management competency assessments
It would be very interesting to compare your findings to these other frameworks/findings.
Reply
Dr. Paul D. Giammalvo Reply:
June 30th, 2009 at 6:53 pm
Hi Dennis,
Yes, I did review PMI Competency Framework, but because of PMI’s rather onerous restrictions on the use of their IP, I only reviewed it.
I much prefer the approach taken by the Global Alliance for Project Performance Standards (GAPPS) which has created a set of project manager competency standard and has put them into the “public domain” under “open source” or “creative commons” licensing. (See page 2) http://www.globalpmstandards.org/listing/project-manager-standards/
GAPPS will be meeting next in Washington DC in late October for the next workshop, where we will be releasing PROGRAM manager competency standards. Consistent with the “creative commons” approach, ANYONE is invited to attend.
BR,
Dr. PDG, Jakarta
Reply
Jun 30th, 2009 at 2:04 pm
Twitter Comment
Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study – [link to post]
– Posted using Chat Catcher
Reply
Jun 30th, 2009 at 3:05 pm
Twitter Comment
Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study – [link to post] (via @ejly)
– Posted using Chat Catcher
Reply
Jun 30th, 2009 at 4:16 pm
Twitter Comment
RT @pmstudent #PMOT Comment on Behavioral Profiles of SUCCESSFUL Project Managers [link to post]
– Posted using Chat Catcher
Reply
Jun 30th, 2009 at 5:03 pm
Twitter Comment
[link to post] 2/2 #PMOT
– Posted using Chat Catcher
Reply
Jul 1st, 2009 at 8:03 pm
Paul,
How did you separate correlation from causality?
Is there null hypothesis H(0)that these attributes are associated from “successful” PM’s testable by an inverse value test with an “unsuccessful” PM group taking the same assessment test and producing a statistically significant variance from the correlations.
The next statistical test question – what is the correlation between the “successful PM” and the attribute profiles in the assessment test? Is this correlation statistically significant and what is the confidence interval on that statistical correlation? Is there a two-sided test you could have performed to compare the means of the control group with the sample group? The hypothesis might be “there is a statistical difference between those in the groups exhibiting “successful” qualifications before applying the attribute assessment, from a statistically sound random groups taking the same test. Can there be a hand selected groups presenting the inverse of the “success attributes” to confirm that the pre-selection of “success” is actually correlated with the assessment criteria?
What population size is needed to derive a statistically significant population – for both the control group (alternative to H(0)) and the test group (your hand selected 28). Finally what was the statistical bias introduced by hand selecting the 28?
Reply
Dr. Paul D. Giammalvo Reply:
July 1st, 2009 at 8:39 pm
Glen,
Not sure what you mean by separating correlation from causality?
And if you read the last paragraphs carefully, you can see that we have not been able to test for the null hypothesis.
And no, we have not done ANY statistical analysis yet. This was just a pilot program- an uncontrolled experiment. What we are seeking is a company or organization sufficiently interested in this topic to provide funding and/or participants for further, more structured research.
While the population from the original pilot was mixed in terms of industry sectors, I suspect that there may well be differences from one industry to another, although that hasn’t been born out by subsequent observations.
As for the population size, ideally, 100 or so for each of the categories (Essential, Desirable and Undesirable) should provide a reasonably smooth distribution curve. But I think 20 for each category would provide meaningful data.
Lastly, as I explained, we have done NO statistical analysis on any of the data, so I don’t know the bias introduced by hand selecting the candidates. However, isn’t that process EXACTLY the same one we use in most companies? We pick those who impress us most as being “competent” or “capable” and in most cases, that process is purely subjective?
To conclude, if you are interested in pursuing this further, we are seeking funding or a company willing to volunteer to participate further. We have had two thus far, but due to NDA’s the results are not ours to release. And we don’t know if those clients did or did not do any further statistical analysis.
John, anything you want to add?
BR,
Dr. PDG, Jakarta
http://www.getpmcertified.com
Reply
Jul 1st, 2009 at 11:20 pm
Paul,
A common statistical error in many studies like this is the confusion between a correlation between two data sets – yourhand picked subjects and the attribute test and the implication that there is a causal connection between their score and their ability to be good PM’s.
There may be a correlation – good PM’s have specific profiles. But that has little to do with the hypothesis that specific profiles are the source of good PM’s.
As for the “ideal” population size, you’ll need to define the Z-test sample size needed to acheive the desired confidence level. 100 is likley completely arbitrary in the absence of that starting analysis.
The work your doing has motivated me to bring down my research methods books from grad school decades ago. If you further interest in “experiment design,” I would be happy to recommend materials. After 31 year the processes are still the same for defining the population sizes, construtcing H and H(0), and establishing the confidence levels – even for social science experiments like this. My background was particle physics, so the confidence intervals were a “bit” tighter when we wanted to conjecture there was a “causal” connection between what we say and what the theory said we should be seeing from the detector.
Reply
Dr. Paul D. Giammalvo Reply:
July 1st, 2009 at 11:38 pm
Many thanks for your kind offer, Glen, but having earned my PhD, I believe I have sufficient expertise in designing an experiment and performing the appropriate statistical analysis.
If you want to be helpful, what we really need is an organization sufficiently interested to take this from the the preliminary and anecdotal evidence from the pilot and follow on applications, allowing us to do a much more rigorous experiment.
Thanks again for your interest and continued support and participation.
BR,
Dr. PDG, Jakarta
http://www.getpmcertified.com
Reply
Jul 2nd, 2009 at 1:31 am
Twitter Comment
RT @franklyPM Behavioral Profiles of SUCCESSFUL Project Managers- A pilot research study – [link to post] #pmot
– Posted using Chat Catcher
Jul 2nd, 2009 at 4:04 am
Twitter Comment
Comme c’est étonnant… [link to post]
– Posted using Chat Catcher
Jul 2nd, 2009 at 8:20 am
Paul,
Once your experiment has been designed with a H(0) and it’s defined correlation and confidence intervals, a control group profile, and the statement as to the causal connection between behavioral attributes and success of project managers, I would say there many firms would be interested in the quantitative results our firm being one.
In the absence is these, the experiment appears to be an anecdotal observation gathering exercise. Interesting perhaps, but not sure if the results are actionable outside the hand selected participants.
We work several programs where the selection and assignment of PM’s is a continuous issue. Assessment of suitability for the job is guided by our Organization Development practice. A quantitative instrument would certainly be more useful than our current approach if there can be shown a casual connection between the classification with the screening and project manager success.
Reply
Dr. Paul D. Giammalvo Reply:
July 2nd, 2009 at 9:05 am
Glen,
The initial experiment did exactly what it was intended to do. It was a pilot study, to see if there was any predictability, which there is. We have repeated the experiment on a larger scale, not once, but twice, by selecting people based on that profile and the top scorers from both companies proved to be “better” project managers than those who didn’t score as highly.
Scientific enough for you? Obviously not. A step in the right direction? We think so.
The reason for posting the article in the first place is to see if companies (such as yours) would be interested in doing a more detailed and controlled experiment. Nothing more than that.
So if your company is interested, are you going to wait for some other company to step forward to help us “prove” the null hypothesis before you invest $50 or $100k? Or are you willing to try a group of about 100 people, split 33% between top, middle and poor performing project managers, so we can get the data to run the kinds of statistical tests you want?
Enough for tonight, Glen…… Off for a long weekend of Yellowfin Tuna fishing for the Fourth of July weekend. No cellphone, no email…..
BR,
Dr. PDG, Jakarta
http://www.getpmcertified.com
Reply
Nov 29th, 2009 at 1:45 am
This looks like the same information from Andy Crowe’s book titled, “Alpha Project Managers: What the Top 2% Know That Everyone Else Does Not”.
Reply