To begin with, consider what factors might cause someone to abandon research and go into management (note that this excludes the somewhat rare individual that continues research while assuming a management position, for example in order to have increased influence at setting research directions and/or because they also have a talent for effective management - this is rare since true research requires a degree of concentration that for most is impossible given the constant distractions associated with managing a project or department). There are the following possible factors: (1) having failed to produce significant research results in the past, and losing hope of ever doing so in the future; (2) having succeeded at producing significant research results, but for various reasons (e.g. becoming unable to generate fundamentally new ideas, deterioration of the ability to solve complex technical problems, a perceived exhaustion of potentially significant new directions in the field of research for which expertise was developed, etc.) deciding that this is unlikely to happen in the future; (3) a loss of interest in research; and (4) valuing money or power more than ideas. Now if there has been a loss of interest, or a change in values, then then there can be no future research success, regardless of success or lack thereof in the past. The second case seems more complex, but the end result is the same: there is a perception that there can be no future research success. The first case is the simplest: failure in the past and an expectation of continued failure in the future. Thus the various factors can be combined into a single factor: a strong expectation (either at the conscious level, as in the first two cases; implicitly, as in the second two cases; or possibly some combination of these) of failure at continued research.

Failure is always traumatic (even in the case where it may be a feeling of failure at some subconscious level), and the problem of this failure can be approached in two basic ways. The first way can be considered positive, and transmutes failure into success: by doing everything possible to facilitate the research of those being managed (that is, of those who - one would hope - have not abandoned research), the research manager achieves a degree of success that may be considered to be the sum of the successes of those whose research is being managed. This is a research manager of the first type. A distinguishing characteristic is that research directions are determined in a bottom up fashion (that is, although the research manager ostensibly sets research directions, this is done by means of consultation with and advice from the researchers).

The second way can be considered negative, and results not only from failing at continued research, but also from failing to recognize the possibility that this failure can be transmuted into success. Research managers of this second type, unable to recover from the trauma of failing at research in the past and/or strongly expecting to fail at research in the future, can easily be identified, since their actions are characterized by arrogance, vindictiveness, and greed. Research directions are always set in a top down fashion. Rather than facilitating the research of those being managed (having failed to recognize the potential for the transferral of success), they tend to do everything possible to inhibit independent research (due to a perception that research successes of those being managed only deepen their own failure at continued research).

One might naively expect that over time only research managers of the first type would succeed, and that those of the second type would eventually fail. In practice, however, the dynamics of the situation at any given time in a research organization are complex, and research managers of the second type may achieve a high degree of success. For example, they tend to excel at winning territorial disputes, at giving impressive (but, unknown to even more clueless higher levels of management, inherently meaningless) presentations, at convincingly selling vaporware of all sorts, at slashing head count (when the opportunity arises) in order to meet ill-conceived short term objectives, and finally, they are masters at obsequiousness toward those that they regard as above them in the organizational hierarchy.

The collection of factors favoring research managers of each of these two types essentially defines the research management culture. Simplifying, we may call a research management culture that exclusively favors research managers of the first type a culture of the first type, say C₁, and a research management culture that exclusively favors research managers of the second type a culture of the second type, say C₂. Any given research organization will have a research management culture, C(t), which is a time-varying mixture of these two types:

Given the above discussion, which may be considered as an extended definition, the changes in the management culture at IBM Research (with which the current author is unfortunately all too familiar) over the 25 year period 1980-2005 can be summarized as follows: r(t) is a monotonically decreasing function, with r(1980) ~= 0.85 and r(2005) ~= 0.05.

- John T. Robinson