A Conversation Between Frank Landy and John D. Mayer
[Editor’s note] From February 9 th to 12 th, 2004, In preparation for a forthcoming review he was working on concerning emotional intelligence, Frank Landy and I exchanged a number of e-mails regarding the nature of emotional intelligence, and whether it was a new concept. Here is a conversation between the two of us based in part on an edited version of that correspondence. It is posted with both our permissions.
Dr. Landy is author of Work in the 21 st Century: An Introduction to Industrial and Organizational Psychology, with J. M. Conte. He is professor emeritus in Industrial and Organizational Psychology at Penn State University. Dr. Landy has been a visiting lecturer or researcher at such Universities as Stanford University, Stockholm University, and Gothenburg University. He has published numerous articles and textbooks. His research work has been funded by a number of government agencies, and he is presently a member of SHL, an international assessment and consulting firm.
Frank Landy: How is your model of emotional intelligence different than the initial speculation by Thorndike that there were many different ways to measure intelligence and that the typical measures of abstract intelligence of that time were insufficiently diverse?...Thorndike’s point was that there were an almost infinite number of intelligences depending on stimulus content and experience -- which accounted for neural connections, the core of his associationist theory. As an aside, this opens the door for arguments in favor of tacit knowledge (or practical intelligence or procedural knowledge -- all much the same thing) as the foundation for emotional intelligence.
John (Jack) Mayer: Ours is a different approach, because I/we are more traditional in our view of intelligence, believing there to be genuinely different (but usually correlated) intelligences for different domains of information. We are not critiquing the intelligence establishment with our theory. Of course, if anyone says there are an infinite number of intelligences, then all new work would be a subset of that person’s statement, but the generalization (infinite intelligences) isn’t the same as defining one, operationalizing it, and seeing what it does.
We take the term "emotional intelligence" very seriously -- that is, to understand it as parallel to "verbal intelligence," "spatial intelligence," "perceptual-organizational intelligence," and the like, and to clearly define what does and does not constitute emotional information in that context.
One reason it took Peter Salovey and me so long to move from our 1990 piece through our theoretical "increments" of 1993 and 1995 to our 1997 statement of the 4-branch model, was that we were seriously in search of new wine -- a new intelligence. Of course, emotional intelligence wasn't totally new wine -- but it was wine that had almost never been tasted before, because the important precursors to our work were relatively obscure to most psychologists.
Landy: Which precursors are you talking about?
Mayer: Precursors from clinical literature on alexithymia (and repression), empathic sending and receiving ability, from non-verbal receiving and sending ability reviewed by Buck in 1980 or so, from the cognition and affect literature on creativity and judgment, and from research on aesthetics. This is off the top of my head. Many of these precursors are outlined in our 1990 article, and also in a forthcoming article in Psych Inquiry. [Ed: published in 2004].
Landy: What if you regressed the Big 5 (Actually, more like the Big 6 now that there is some agreement in partialing conscientousness into detail-orientation and achievement motive, disposition (positive affect and negative affect), “g”, and tacit knowledge against the MSCEIT? Do you think there would still be variance (other than error) unaccounted for?
Mayer: I have yet to see another test with an average correlation with the MSCEIT higher than .35 or so (this includes all personality scales, intelligence scales, and other emotional intelligence scales, and ability-based emotional creativity scales I have seen, taken individually).
Landy: Take the two ".35's," (for the Big 5, and verbal intelligence). Assume they are independent, and that each has a 12% variance overlap with the MSCEIT (based on the .35 correlations). You're still explaining only 24% of the MSCEIT's reliable variance -- say a multiple R of .50 or so. That's not much using 6 ostensibly orthogonal tests to predict a test with an overall reliability above .90.
Mayer: Even regressing the Big Five against the MSCEIT doesn't do much better. For example, in Brackett's 2003 article with me, he showed that regressing all 5 of the Big Five against the MSCEIT yielded a multiple R of .38 (note that this is simply the R, rather than the R-squared). (By contrast, it was .75 with the EQ-i and .58 with the Schutte scale -- the latter scale was more highly predicted by a scale of well-being). So, yes, I believe the variance of emotional intelligence is genuinely different from that of earlier scales. I believe it to be potentially as discrete as the variance for, say, spatial intelligence. I don't think adding tests will do much in terms of accounting for more of the variance, other than capitalize on chance.
Landy: My larger question, of course, is that even if you do identify some unique variance, does it matter? As a differential psychologist, one might say it doesn’t have to matter, it simply has to be (as in the periodic chart of the elements). But is the desire that of the taxonomist (differential psychologist); that is, to add another classification?
Mayer: I am content with theoretical validity as per the periodic chart. At the same time, I think that as new criteria for emotional intelligence are identified, there will be socially significant incremental validity demonstrated in several areas (as indicated in forthcoming validity reviews). [Ed: in Psychological Inquiry, 2004].
Landy: As an applied psychologist, one must ask the “matter" question. Of course you believe that it matters, not that it “is.” That is certainly the thrust of all commercially available instruments. We are not at the 1893 Exposition in Chicago when Cattell and Munsterberg wowed the audience with this new thing called a psychological test.
Mayer: My general understanding of the evolution of human cleverness suggests to me that it is fairly unlikely that one could find a fairly broad (e.g., characterized now by up to 15 or so tasks), unifactorial area of human ability, such as emotional intelligence, that isn't important, and meaningfully so, for some purposes. Of course, we'll need to continue to uncover evidence for how, and I accept that responsibility (and invite others’ assistance in doing so).
Landy: Again, to be theatrical, so what? What does it matter what the correlation is between the MSCEIT and big toe-nail thickness?
Mayer: As we learn more about the correlates, I think they will have some definable importance.
Landy: Where does that leave us for now? New wine in no bottles?
Mayer: My argument is that with the ability model, we have new wine, and some new bottles. What we don't know as well as we would like yet, is what those who have been drinking this wine have been able to do, that others haven’t been able to do. But that is becoming clearer with the most recent research.
Landy: Well, should I concede the “old wine in new bottles” argument regarding emotional intelligence?
Mayer: Unfortunately, no, because that does reflect much of the activity that takes place in the field under the name, “emotional intelligence,” but that is mostly because many others using the term emotional intelligence are (from my standpoint) misusing it. Our 1997 statement of EI has nothing to do with the other EI theories out there (which are hodgepodges of earlier personality theories, with some overlap with our own model).
[The next section returns to the issue of whether EI can be distinguished from other intelligences -- Ed.].
Landy: I am not sure I see the distinction -- after all Thorndike and possibly Bob Glaser and J. B. Carroll anticipated many different intelligences.
Mayer: It is one thing to anticipate that a multitude of new unspecified intelligences will be discovered in the future -- that isn’t much different than anticipating that science will make future discoveries.
In contrast, our theory provides a coherent and specific description of a previously unstudied domain of human abilities. I think there are multiple definitions and approaches to emotional intelligence. I think our own best represents the new wine.
I haven't seen an earlier theoretical discussion that describes a chunk of human ability, that is similar to that of our 1997 paper (although, again, I grant the existence of plenty of important precursors; again, the closest and most important precursors were very obscure and remain so).
Please forgive me if have talked up our theory too much here, but I really can't see how the old wine issue applies in this instance, and would invite you to instruct me if you still believe it does after reading this.
Landy: Very well put. Thanks. I have enjoyed our dialogue and it has (and will further) influence the way I think and write about EI, I hope in a way that will please you. I am yet to be convinced, at least in the employment domain, that EI has incremental validity over other measures. On the other hand, I would not argue that EI devices might not be more efficient at predicting certain work behaviors but my endorsement would be based on an efficiency notion, not (yet) that EI measures something unique.
Mayer: I have enjoyed the dialogue as well and I wish you luck with your forthcoming chapter on the area.