Archive for the ‘Validity of 360 Processes’ Category
One question that has been at the core of best practices in 360 Feedback since its inception relates to the conditions that are most likely to create sustained behavior change (at least for those of us that believe that behavior change is the ultimate goal). Many of us believe that behavior change is not a question of ability to change but primarily one of motivation. Motivation often begins with the creation of awareness that some change is necessary, the accepting the feedback, and then moving on to implementing the change.
One of the more interesting examples of creating behavior change began when seat belts were included as standard equipment in all passenger vehicles in 1964. I am old enough to remember when that happened and started driving not long thereafter. So using a seat belt was part of the driver education routine since I began driving and has not been a big deal for me.
The reasons for noncompliance with seatbelt usage are as varied as human nature. Some people see it as a civil rights issue, as in, “No one is going to tell me what to do.” There is also the notion that it protects against a low probability event, as in “It won’t happen to me. I’m a careful driver.” Living in Nebraska for a while, I learned that people growing up on a farm don’t “have the time” to buckle and unbuckle seatbelts in their trucks when they are learning to drive, so they don’t get into that habit. (I also found, to my annoyance, that they also never learned how to use turn signals.)
I remember back in the ‘60’s reading about a woman who wrote a car manufacturer to ask that they make the seat belts thinner because they were uncomfortable to sit on. Really.
Some people have internal motivation to comply, which can also be due to multiple factors such as personality, demographics, training, norms (e.g., parental modeling), and so on. This is also true when we are trying to create behavior change in leaders, but we will see that these factors are not primary determinants of compliance..
In thinking about seatbelt usage as a challenge in creating behavior change, I found study from 2008 by the Department of Transportation. It is titled “How States Achieve High Seat Belt Use Rates” (DOT HS 810 962). (Note: This is a 170 page report with lots of tables and statistical analyses, and if any of you geeks want a copy, let me know.)
The major finding of this in-depth study states:
The statistical analyses suggest that the most important difference between the high and low seat belt use States is enforcement, not demographics or funds spent on media.
This chart Seatbelt Usage in US, amongst the many in this report, seems to capture the messages fairly well to support their assertion. This chart plots seat belt usage by state, where we see a large spread ranging from just over 60% (Mississippi) to about 95% (Hawaii). It also shows whether each state has primary seatbelt laws (where seatbelt usage is a violation by itself), or secondary laws (where seatbelt usage can only be enforced if the driver is stopped for another purpose). Based on this table alone, one might argue causality but the study systematically shows that this data, along with others relating to law enforcement practices, are the best predictors of seatbelt usage.
One way of looking at this study is to view law enforcement as a form of external accountability, i.e., having consequences for your actions (or lack thereof). The primary versus secondary law factor largely shifts the probabilities of being caught, with the apparent desired effect on seatbelt usage.
So, back to 360 Feedback. I always have been, and continue to be, mystified as to how some implementers of 360 feedback processes believe that sustainable behavior change is going to occur in the vast majority of leaders without some form of external accountability. Processes that are supposedly “development only” (i.e., have no consequences) should not be expected to create change. In those processes, participants are often not required to, or even discouraged from, sharing their results with others, especially their manager. I have called these processes “parlor games” in the past because they are kind of fun, are all about “me,” and have no consequences.
How can we create external accountability in 360 processes? I believe that the most constructive way to create both motivation and alignment (ensuring behavior change is in synch with organizational needs/values) is to integrate the 360 feedback into Human Resource processes, such as leadership development, succession planning, high potential programs, staffing decisions, and performance management. All these uses involve some form of decision making that affects the individual (and the organization), which puts pressure on the 360 data to be reliable and valid. Note also that I include leadership development in this list as a form of decision making because it does affect the employee’s career as well as the investment (or not) of organization resources.
But external accountability can be created by other, more subtle ways as well. We all know from our kept and (more typically) unkept New Year’s resolutions about the power of going public with our commitments to change. Sharing your results and actions with your manager has many benefits, but can cause real and perceived unfairness if some people are doing it and others not. Discussing your results with your raters and engaging them in your development plans has multiple benefits.
Another source of accountability can (and should) come from your coach, if you are fortunate enough to have one. I have always believed that the finding in the Smither et al (2005) meta-analysis that the presence of a coach is one determinant of whether behavior change is observed is due to the accountability that coaches create by requiring the coachee to specifically state what they are going to do and to check back that the coachee has followed through on that commitment.
Over and over, we see evidence that, when human beings are not held accountable, more often than not they will stray from what is in their best interests and/or the interests of the group (organization, country, etc.). Whether it’s irrational (ignoring facts) or overly rational (finding ways to “get around” the system), we should not expect that people will do what is needed, and we should not rely on our friends, neighbors, peers or leaders to always do what is right if there are no consequences for inaction or bad behavior.
©2012 David W. Bracken
My good friend and collaborator, Dale Rose, dropped me a note regarding his plans to do another benchmarking study on 360 Feedback processes. His company, The 3D Group, has done a couple of these studies before and Dale has been generous in sharing his results with me, which I have cited in some of my workshops and webinars. The studies are conducted by interviewing coordinators of active 360 systems. Given that they are verbal, some of the results have appeared somewhat internally inconsistent and difficult to reconcile, though the general trends are useful and informative.
Many of the topics are useful for practitioners to gauge their program design, such as the type of instrument, number of items, rating scales, rater selection, and so on. For me, the most interesting data relates to the various uses of 360 results.
Respondents in the 2004 and 2009 studies report many uses. In both studies, “development” is the most frequent response, and that’s how it should be. In fact, I’m amazed that the responses weren’t 100% since a 360 process should be about development. The fact that in 2004 only 72% of answers included development as a purpose is troubling whether we take the answers as factual or if they didn’t understand the question. The issue at hand here is not whether 360’s should be used for development; it is what else they should, can, and are used for in addition to “development.”
In 2004, the next most frequent use was “career development;” that makes sense. In 2009, the next most frequent was “performance management,” and career development dropped way down. Other substantial uses include high potential identification, direct link to performance measurement, succession planning, and direct link to pay.
But when asked whether the feedback is used “for decision making or just for development”, about 2/3 of the respondents indicated “development only” and only 1/3 for “decision making.” I believe these numbers understate the actual use of 360 for “decision making” (perhaps by a wide margin), though (as I will propose), it can depend on how we define what a “decision” is.
To “decide” is “to select as a course of action,” according to Miriam Webster (in this context). I would build on that definition that one course of action is to do nothing, i.e., don’t change the status quo or don’t let someone do something. It is impossible to know what goes on in person’s mind when he/she speaks of development, but it seems reasonable to suppose that it involves doing something beyond just leaving the person alone, i.e., maintaining the status quo. But doing nothing is a decision. So almost any developmental use is making a decision as to what needs to be done, what personal (time) and organizational (money) resources are to be devoted to that person. Conversely, denying an employee access to developmental resources that another employee does get access to is a decision, with results that are clearly impactful but difficult to measure.
To further complicate the issues, it is one thing to say your process is for “development only,” and another to know how it is actually used. Every time my clients have looked behind the curtain of actual use of 360 data, they unfailingly find that managers are using it for purposes that are not supported. For example, in one client of mine, anecdotal evidence repeatedly surfaced that the “development only” participants were often asked to bring their reports with them to internal interviews for new jobs within the organization. The bad news was that this was outside of policy; the good news was that leaders saw the data as useful in making decisions, though (back to bad news) they may have been untrained to correctly interpret the reports.
Which brings us to why this is an important issue. There are legitimate “development only” 360 processes where the participant has no accountability for using the results and, in fact, is often actively discouraged from sharing the results with anyone else. Since there are not consequences, there are few, if any, consequential actions or decisions required. But most 360 processes (despite the benchmark results suggesting otherwise) do result in some decisions being made, which might include doing nothing by denying an employee access to certain types of development.
The Appendix of The Handbook of Multisource Feedback is titled, “Guidelines for Multisource Feedback When Used for Decision Making.” My sense is many designers and implementers of 360 (multisource) processes feel that these Guidelines don’t apply because their system isn’t used for decision making. Most of them are wrong about that. Their systems are being used for decision making, and, even if not, why would we design an invalid process? And any system that involves the manager of the participant (which it should) is creating the expectation of direct or indirect decision making to result.
So Dale’s question to me (remember Dale?) is how would I suggest wording a question in his new benchmarking study that would satisfy my curiosity regarding the use of 360 results. I proposed this wording:
“If we define a personnel decision as something that affects an employee’s access to development, training, jobs, promotions or rewards, is your 360 process used for personnel decisions?”
Dale hasn’t committed to using this question in his study. What do you think?
©2012 David W. Bracken
I don’t know why I feel compelled to respond to what I see are unreasonable positions (primarily in LinkedIn discussions). But I do, and this blog gives me a vehicle for doing so without taking up a disproportionate amount of air time on that forum.
So what got me going this time? A LinkedIn discussion (that I started on the topic of 360 validity) got diverted into the topic of “proper” use of 360 feedback (development vs decision making). The particular comment that got me going was, “I believe these assessments should be used for development – full stop.” (Virtually 100% of 360 processes are used for development, but the context indicates that he meant “development only.”) Having lived and worked in London for a while, I realized (or realised) that the “full stop” has the same meaning as “period,” implying end of sentence and, with emphasis, no more is worth saying. By the way, I am using this person only as an example of the many, many individuals who have expressed similar dogmatic views on this topic.
There are probably a few things that are appropriate to put a “full stop” on. That would be an interesting blog for someone, e.g., would we include the Ten Commandments? “Thou Shall Not Kill. Full stop.” Hmmm… but then we have Christians who believe in capital punishment, so maybe it’s only a partial stop (or pause)? Like I said, I will let someone else take that on.
Are the physical sciences a place for “full stops?” Like, “The world is flat. Full stop.” “ The Sun revolves around the Earth. Full stop.” Just this last week, we were presented with the possibility that another supposedly immutable law is under attack, i.e., “Nothing can go faster than the speed of light. Full stop.” Now we have European scientists who have observed neutrinos apparently traveling faster than the speed of light and are searching for ways to explain and confirm it. If found to be true, it would challenge many of the basics of physics, opening the door to time travel, for example. The fact that some scientists are apparently challenging the “full stop” nature of the Theory of Relativity is also fascinating, if only for the reason that they are open to exploring the supposedly impossible. And, by the way, are begging for others to challenge and/or replicate their findings.
I firmly believe that the social sciences have no place for “full stops.” To me, “full stop” means ceasing to explore and learn. It seems to indicate a lack of openness to considering new information or different perspectives.
I suspect there are many practitioners in the “hard” sciences who question whether what we do is a “science” at all. (I think I am running out of my quota of quotations marks.) Perhaps they see our work with understanding human behavior as a quest with no hope of ever having answers. That’s what I like about psychology. We will never fully know how to explain human behavior, and that’s a good thing. If we can explain it, then we probably could control it. I think that is a scary thought. BUT we do try to improve our understanding and increase the probabilities of predicting what people will do. That is one of the basic goals of industrial/organizational psychology.
(I have been known to contend that what we do is harder than rocket science because there are no answers to what we do, only probabilities. The truth is that even the hard sciences have fewer “full stops” than even they would like. I just finished reading a book about the Apollo space program, Rocket Men, and it is very interesting to know how many “hard stops” that used to exist were bashed (e.g., humans can’t live in weightlessness, the moon’s crust will collapse if we try to land on it. Insert “hard stops” appropriately), how much uncertainty there was, and how amazing the accomplishment really was. I also learned that one of the reasons the astronauts’ visors were mirrored was so that aliens couldn’t see their faces. Seriously.)
Increasing probabilities for predicting and influencing employee behavior requires that we also explore options. I can’t see how it is productive to assert that we know the answer to anything, and that we shouldn’t consider options that help us serve our clients, i.e., organizations, more effectively.
On top of all that, the most recent 3D Group benchmark study indicates that about one third of organizations DO use 360 data for some sort of administrative purpose, and that almost certainly understates the real numbers. What do we tell those organizations? That they should cease doing so since our collective wisdom says that there is no way they can actually be succeeding? That we cannot (or should not) learn from what they are doing to help their organizations make better decisions about their leaders? That a few opinions should outweigh these experiences?
I don’t get it. No stop.
©2011 David W. Bracken
My friend, Joan Glaman, dropped me a note after my last blog, (http://dwbracken.wordpress.com/2011/08/30/thats-why-we-have-amendments/ ) with this suggestion:
“I think your closing question below would be a great next topic for general discussion: ‘Under what conditions and for whom is multisource feedback likely to be beneficial?’”
To refresh (or create) your memory, that question that Joan cites is from the Smither, London and Reilly (2005) meta analysis. The article abstract states:
“…improvement is most likely to occur when feedback indicates that change is necessary, recipients have a positive feedback orientation, perceive a need to change their behavior, react positively to the feedback, believe change is feasible, set appropriate goals to regulate their behavior, and take actions that lead to skill and performance improvement.”
Before we answer Joan’s question, we should have a firm grasp on what we mean by “beneficial.” I don’t think we all would agree on that in this context. Clearly, Smither et al. define it as “improvement,” i.e., positive behavior change. That is the criterion (outcome) measure that they use in their aggregation of 360 studies. I am in total agreement that behavior change is the primary use for 360 feedback, and we (Bracken, Timmreck, Fleenor and Summers, 2001) defined a valid 360 process as one that creates sustainable behavior change in behaviors valued by the organization.
Not everyone will agree that behavior change is the primary goal of a 360 process. Some practitioners seem to believe that creating awareness alone is a sufficient outcome since they do not support any activity or accountability, proposing that simply giving the report to the leader is going far enough and in fact discourage the sharing of results with anyone else.
If you will permit a digression, I will bring to your attention a recent blog by Sandra Mashihi (http://results.envisialearning.com/5-criteria-a-360-degree-feedback-must-meet-to-be-valid-and-reliable/) where one of her lists of “musts” (arrrgh!) is criterion related validity, which she defines as, …does the customized instrument actually predict anything meaningful like performance?” Evidently she would define “beneficial” then to not be behavior change but to be able to measure performance to make decisions about people. This testing mentality just doesn’t work for me since 360’s are not tests (http://dwbracken.wordpress.com/2010/08/31/this-is-not-a-test/) and it is not realistic to expect them to predict behavior, especially if we hope to actually change behavior.
Let’s get back to Joan’s question (finally). I want to make a couple comments and then hopefully others will weigh in. The list of characteristics that Smither et al provide in the abstract is indeed an accumulation of individual and organizational factors. This is not an “and” list that says that a “beneficial” process will have all these things. It an “or” list where each characteristic can have benefits. The last two, (set goals and take actions) can be built into the process as requirements regardless of whether the individual reacts positively and/or perceives the need to change. Research shows that follow up and taking action are powerful predictors of behavior change, and I don’t believe that it is important (or matters) to know if the leader wants to change or not. What if he/she doesn’t want to change? Do they get a pass? Some practitioners would probably say, yes, and point to this study as an indication that it is not worth the effort to try to get them to change.
I suggest that this list of factors that lead to behavior change are not independent of each other. In our profession, we speak of “covariates”, i.e., things that are likely to occur together across a population. A simple example is gender and weight, where men are, on average, heavier than women. But we don’t conclude that men as a gender manage their weight less well than women, it’s due to being taller (and other factors, like bone structure).
My daughter, Anne, mentioned in passing an article she read about people who don’t brush their teeth twice a day having a shorter life expectancy than those who do. So the obvious conclusion is that brushing teeth more often will make us live longer. There is certainly some benefit to regularly brushing teeth, but it’s more likely that there are covariates of behavior for people that don’t have good dental hygiene that have a more direct impact on health. While I don’t have data to support it, it seems likely that people who don’t brush regularly also don’t go to the dentist regularly for starters. It seems reasonable to surmise that, on average, those same people don’t go to their doctor for a regular checkup.
My hypothesis is that 360 participants who aren’t open to feedback, don’t perceive a need to change, don’t feel that they can change, etc., are also the people who are less likely to set goals and take action (follow up) if given the option to not do those things. In other words, it’s not necessarily their attitudes that “cause” lack of behavior change, but the lower likelihood that they will do what is necessary, i.e., set goals and follow through, in order to be perceived as changing their behavior. Those “behaviors” can be modified/changed while their attitudes are likely to be less modifiable, at least until they have had a positive experience with change and its benefits.
One last point of view about “beneficial.” Another definition could be change that helps the entire organization. That is the focus of the recent publication by Dale Rose and myself, where (in answer to Joan’s question) we state:
“…four characteristics of a 360 process that are required to successfully create organization
change, (1) relevant content, (2) credible data, (3) accountability, and (4) census participation…”
We go on to offer the existing research that supports that position, and the wish list for future research. One way of looking at this view of what is “beneficial” is to extrapolate what works for the individual and apply it across the organization (which is where the census (i.e., whole population) part comes into play.)
I will stop there, and then also post this on LinkedIn to see if we can get some other perspectives.
©2011 David W. Bracken
I used my last blog (http://dwbracken.wordpress.com/2011/08/09/so-now-what/) to start LinkedIn discussions in the 360 Feedback and I/O Practitioners group, asking the question: What is a “valid” 360 process? The response from the 360 group was tepid, maybe because the group has a more general population that might not be that concerned with “classic” validity issues (which is basically why I wrote the blog in the first place). But the I/O community went nuts (45 entries so far) with comments running the gamut from constructive to dismissive to deconstructive.
Here is a sample of some of the “deconstructive” comments:
…I quickly came to conclusion it was a waste of good money…and only useful for people who could (or wanted to) get a little better.
It is all probably a waste of time and money. Good luck!!
There is nothing “valid” about so-called 360 degree feedback. Technically speaking, it isn’t even feedback. It is a thinly veiled means of exerting pressure on the individual who is the focal point.
My position regarding performance appraisal is the same as it has been for many years: Scrap It. Ditto for 360.
Actually, I generally agree with these statements in that many 360 processes are a waste of time and money. It’s not surprising that these sentiments are out there and probably quite prevalent. I wonder, though, if we are all on the same page. In another earlier blog, I suggested that discussions about the use and effectiveness of 360’s should be separated by those that are designed for feedback to a single individual (N=1) and those that are designed to be applied to groups (N>1).
But the fact is that HR professionals have to help their management make decisions about people, starting with hiring and then progressing through placement, staffing, promotions, compensation, rewards/recognition, succession planning, potential designation, development opportunities, and maybe even termination.
Nothing is perfect, especially so when it comes to matters that involve people. As an example, look to the U.S. Constitution, an endearing document that has withstood the test of time. Yet the Founding Fathers were the first to realize that they needed to make provisions for the addition of amendments to further make refinements. Of course, some of those amendments were imperfect themselves and were later rescinded.
But we haven’t thrown out the Constitution because it is imperfect. Nor do we find it easy to come to agreements what the revisions should be. But one of the many good things about humans is a seemingly natural desire to make things better.
Ever since I read Mark Edwards and Ann Ewen’s seminal book, 360 Degree Feedback, I have believed that 360 Feedback has the potential to improve personnel decision making when done well. The Appendix of The Handbook of Multisource Feedback is titled, “Guidelines for multisource feedback when used for decision making,” coauthored with Carol Timmreck, where we made a stab at defining what “done well” can mean.
In our profession, we have an obligation to constantly seek ways of improving personnel decision making. There are two major needs we are trying to meet, which sometimes cause tensions. One is to provide the organization with more accurate information on which to base these decisions, which we define as increased reliability (accurate measurement) and validity (relevant to job performance). Accurate decision making is good for both the organization and the individual.
The second need is to simultaneously use methods that promote fairness. This notion of fairness is particularly salient in the U.S. where we have “protected classes” (i.e., women, minorities, older workers), but hopefully fairness is a universal concept that applies in many cultures.
Beginning with the Edwards & Ewen book and progressing from there, we can find more and more evidence that 360 done well can provide decision makers with better information (i.e., valid and fair) than traditional sources (e.g., supervisory evaluations). I actually heard a lawyer state that organizations could be legally exposed for not using 360 feedback because is more valid and fair than methods currently in use.
I have quoted Smither, London and Reilly (2005) before, but here it is again:
We therefore think it is time for researchers and practitioners to ask “Under what conditions and for whom
is multisource feedback likely to be beneficial?” (rather than asking “Does multisource feedback work?”).
©2011 David W. Bracken
This is the one year anniversary of this blog. This is the 44th post. We have had 2,026 views, though the biggest day was the first with 38 views. I have had fewer comments than I had hoped (only 30), though some LinkedIn discussion have resulted. Here is my question: Where to go from here? Are there topics that are of interest to readers?
Meanwhile, here is my pet peeve(s) of the week/month/year: I was recently having an exchange with colleagues regarding a 360 topic on my personal Gmail account and up pops ads on the margin for various 360 vendors (which is interesting in itself), the first of which is from Qualtrics (www.qualtrics.com) with the heading, “Create 360s in Minutes.”
The topic of technology run amok has been covered before here (When Computers Go Too Far, http://wp.me/p10Xjf-3G), my peevery was piqued (piqued peevery?) when I explored their website and saw this claim: USE VALIDATED QUESTIONS, FORMS and REPORTS.”
What the heck does that mean? What are “validated” forms and reports, for starters?
The bigger question is, what is “validity” in a 360 process? Colleagues and I (Bracken, Timmreck, Fleenor and Summers, 2001; contact me if you want a copy) have offered up a definition of validity for 360’s that holds that it consists of creating sustainable change in behaviors valued by the organization. Reliable items, user friendly forms and sensible reports certainly help to achieve that goal, but certainly cannot be said to be “valid” as standalone steps in the process.
The Qualtrics people don’t share much about who they are. Evidently their founder is named Scott and teaches MBA’s. They appear to have a successful enterprise, so kudos! I would like to know how technology vendors claim to have “valid” tools and what definition of validity they are using.
Hey maybe I will get my 31st comment?
©2011 David W. Bracken
I have a few events coming up in the next 3 weeks or so that I would like to bring to your collective attention in case you have some interest. One is free, two are not (though I receive no remuneration). I also have an article out that I co-authored on 360 feedback.
In chronological order, on May 25 Allan Church, VP Global Talent Development at PepsiCo, and I will lead a seminar titled, “Integrating 360 & Upward Feedback into Performance and Rewards Systems” at the 2011 World at Work Conference in San Diego (www.worldatwork.org/sandiego2011). I will be offering some general observations on the appropriateness, challenges, and potential benefits of using 360 Feedback for decision making, such as performance management. The audience will be very interested in Allan’s descriptions of his experiences with past and current processes that have used 360 and Upward Feedback for both developmental and decision making purposes.
On June 8, I am looking forward to conducting a half day workshop for the Personnel Testing Council of Metropolitan Washington (PTCMW) in Arlington, VA, titled “360-Degree Assessments: Make the Right Decisions and Create Sustainable Change” (contact Training.PTCMW@GMAIL.COM or go to WWW.PTCMW.ORG). This workshop is open to the public and costs $50. I will be building from the workshop Carol Jenkins and I conducted at The Society for Industrial and Organizational Psychology. That said, the word “assessments” in the title is a foreshadowing of a greater emphasis on the use of 360 Feedback in a decision making context and an audience that is expected to have great interest in the questions of validity and measurement.
On the following day, June 9 (at 3:30 PM EDT), I will be part of an online virtual conference organized by the Institute of Human Resources and hr.com on performance management. My webinar is titled, “Using 360 Feedback in Performance Management: The Debate and Decisions,” where the “decisions” part has multiple meanings. Given the earlier two sessions I described, it should be clear that I am a proponent of using 360/Upward Feedback for decision making under the right conditions. The other take on “decisions” is the multitude of decisions that are required to create those “right conditions” in the design and implementation of a multisource process.
On that note, I am proud to say that Dale Rose and I have a new article in the Journal of Business and Psychology (June) titled, “When does 360-degree feedback create behavior change? And how would we know it when it does?” Our effort is largely an attempt to identify the critical design factors in creating 360 processes and the associated research needs.
This article is part of a special research issue (http://springerlink.com/content/w44772764751/) of JBP and you will have to pay for a copy unless you have a subscription. As a tease, here is the abstract:
360-degree feedback has great promise as a method for creating both behavior change and organization change, yet research demonstrating results to this effect has been mixed. The mixed results are, at least in part, because of the high degree of variation in design features across 360 processes. We identify four characteristics of a 360 process that are required to successfully create organization change, (1) relevant content, (2) credible data, (3) accountability, and (4) census participation, and cite the important research issues in each of those areas relative to design decisions. In addition, when behavior change is created, the data must be sufficiently reliable to detect it, and we highlight current and needed research in the measurement domain, using response scale research as a prime example.
Hope something here catches your eye/ear!
©2011 David W. Bracken