Thursday, September 03, 2015

Women in Engineering Leadership

By: Gary J. Salton, Ph.D.
Chief: Research and Development
Professional Communications, Inc.

This study finds that women are poorly represented in engineering management. However, it does not appear to be due to gender selection bias.  Rather the pool of women at the professional engineering level consists of substantial portions who are not well-aligned with culturally mandated demands of engineering leadership. The engineering culture has proven highly effective in meeting engineering goals. It is very unlikely that the this culture can be radically adjusted to accommodate women without damage to the core function. A rational strategy for increasing the management representation of women will involve attracting and retaining women whose information processing profile more closely reflects the role responsibilities of the various managerial levels.

Click to view video
A study on the Engineering Personality (Salton, 2014 see Footnote #1 for reference) was able to trace characteristic engineering behaviors to the nature of the work done. It was also able to establish that engineering exceeded other professions in their cultural commitment. 

Click to view video
The study of Women in Engineering (Salton, 2015 see Footnote #2 for reference) found that women have the capabilities necessary for success in engineering.  Their low participation rate appears to be due to the engineering culture identified in the Engineering Personality study.

Click to view video
The study Engineering Insights (Salton, 2015 see Footnote #3 for reference) extended the research to consider the hierarchy of engineering management. It found that all levels in engineering are sensitive to the same variables, weight them in about the same manner and target the same kind of output. This creates a strong culture without the “relief valves” typically found in other professions.

Click to view video
This study looks at the status of women in engineering leadership positions. The goal is to see if women in leadership positions differed in any significant way from their male leader counterparts. A video summary of this research is available on YouTube by clicking the icon on the right.

The low representation of women in engineering limits the number of women who could rise to a leadership position. Table 1 shows the number of women in our database at each level of management. Levels labeled “Marginal Testability” do not have a sufficient number of women to do a definitive analysis (Footnote #4 for explanation).

Table 1

“I Opt” styles are used to measure the probability that a person will choose a particular kind of response on the next decision. The “I Opt” style” measures the kind of input a person seeks/accepts, the probable character of the intended response and the mechanism for getting from input to output. Information availability sets the range of behaviors that a person can issue. (Footnote #5 for “style” explanation reference). 
For example, attention to detail necessarily limits the decision horizon since detail expands exponentially with time. Focusing on general aspects (i.e., ignoring detail) lengthens the horizon at the cost of a loss of precision. Obviously, different decision horizons can lead to markedly different choices (e.g., long vs short term profit maximization). A host of other behaviors can be accurately predicted from style preferences (see Footnote #6 for reference to a non-exhaustive listing of predictable behaviors).
Table 2 shows the percent difference between men and women in their average “I Opt” style strength as well as the probability that the difference is just random variation  (Footnote #7 for explanation). It should be kept in mind that these are women actually working in engineering and not women taking engineering degrees. The number and “I Opt” style character of female engineers who have chosen not to enter or to leave engineering is unknown. However, leaders are drawn from the pool of working engineers and thus the number of working female engineers are what is relevant to this study.

Table 2

“I Opt” styles represent the first choice a person is likely to make when confronting an unfamiliar situation. They set the initial tone of a relationship. They are what get a person noticed (or ignored) by those making promotional decisions.  Styles influence career advancement.
Table 2 shows that women differ significantly from men at the professional level (see highlighted yellow boxes).  This suggests that a significant proportion of the women in these categories differ from men at a style strength sufficient to move the averages. These women are more highly committed to methodical action (i.e., LP style) and less attentive to more speculative idea generation (i.e., RI style). As noted in the study on Women in Engineering (see Footnote #2) this kind of difference can create an adverse environment.

The managerial levels of engineering show no significant difference between men and women in their approach to decision making. They share the same decision horizon, have roughly the same risk profile, favor the same character of response and so on.  In other words, men and women are being promoted to higher levels based on the same decision making criteria.

This same kind of male-female decision making consistency was found in an earlier Gender at the Executive Level study (Salton, 2006. See Footnote #8 for reference). The various levels of management appear to demand that incumbents adopt an information processing perspective suited to the kind of issues that the level typically addresses. In other words, no special allowances are being made to increase managerial participation along this dimension (See Footnote #9 for a cautionary footnote).

No single information processing strategy applies to every situation. When a favored strategy does not apply people shift to their next most favored strategy.  The specific rank ordering of favored styles establishes a behavioral theme that readily visible to all involved.  “I Opt” terms these characteristic long term behavioral themes as “patterns.”

Styles (short-term decision strategies) characterize an individual among those who have only occasional or intermittent exposure to their decision making.  Patterns (long-term decision strategies) temper these initial judgements among those who are able to witness repeated decisions.  Since time is required to witness repeated exposures “I Opt” patterns are deemed to be long-term strategies. 

In terms of hierarchical promotion, styles typically cause a person to be noticed as adept in a particular function. Patterns are usually used to assess that person’s likely performance in a particular role within that function. Table 3 shows the pattern differences between men and women at the various hierarchical levels..

Table 3

At the managerial levels a majority of the categories register an overall insignificant level of difference between men and women. There are two instances where a marginal level of significant difference is visible. Both of these differences can be questioned.

The Project Management difference in the “Performer” category (i.e., a short-term horizon focused on immediate results) rests on a sample size of only 20 women.  On its own, sample size would not be a “killer.” However, the nature of the job is also highly variable.  Project manager can involve responsibility for a billion dollar infrastructure project or the design of a critical component part.  The combination of small size, widely discrepant job demands and marginal levels of significance call into question the “reality” of the difference found. Refined categorization and larger sample size could well cause the difference to disappear.

The difference in Mid-Level management’s “Perfector” category (i.e., longer-term horizon focused on complete understanding of new items) is different. It has a more substantial sample size of 84 women. Categorization remains an issue. Mid-Level management consists of jobs with a wide range of responsibility. For example, if just the General Manager (n=17 all males) were removed from the comparison the probability of no male-female difference would jump from 1.4 to 3.2 out of 100. Shifting seventeen people can more than double the uncertainty.

A detailed examination of the data underlying the pattern calculation was conducted. A significant “Perfector” difference persisted even when the comparison was confined to managers or directors.  The mid-level manager difference appears to be real. The difference is likely attributable to the information processing strategies used by the pool of women from which mid-level managers are drawn—the Professional Level.

Table 3 indicates that Professional level women and men differ on all four “I Opt” patterns.  In 3 of the 4 patterns that difference reaches “highly significant” levels.  This reinforces the position that the raw number of women engineers dramatically overstates the number actually available for promotion. Many of the professional level women engineers do not have the information processing patterns that match the culturally prescribed patterns suitable for higher levels. 

The result of this condition is that it will appear that women are being systematically discriminated against in promotional opportunities. In reality, those women who use strategies compatible with the demands of these higher level positions are being promoted and “fit in” with some precision (see footnote #10 for an analysis of the mid-level manager anomaly). 

This analysis points to at least some of the causes underlying the issues on condition of women in engineering. The women electing to stay in engineering after graduation are systematically different than their male counterparts. In addition to stereotype exposures (see “Women in Engineering” study) this difference methodically disadvantages them in career advancement. The cause (i.e., information processing election) is invisible. Women considering engineering are likely to attribute this to gender bias. This discourages others from entering the field. The raw number of women in the field is further reduced.  It is a bit of a vicious circle.

The Bureau of Labor Statistics registered 7.6% female managers in Architecture and Engineering while they made up 15.4% of the workforce. In contrast women made up 38.6% of management versus a total workforce of 43.7% across other professional occupations (Bureau of Labor Statistics, 2014; see Footnote #4 for reference).  There appears to be a serious shortfall of women engineering managers whether considered in absolute or relative terms.

This analysis indicates that the managerial shortfall has two interrelated causes. The first is due to the relatively few women who choose to enter and stay in engineering. This can be addressed in a variety of ways. The study of Women in Engineering (footnote #2) suggested that things like the compensation structure could be adjusted to better recognize the role responsibilities of women while still retaining gender equity. Attention to the physical work environment might also be adjusted to lessen the exposure to potentially gender hostile occurrences. These kinds of things in combination with the current promotional efforts would do much to attract and retain women engineers. 

The second issue involves the kind of the women being attracted and retained. Information processing styles and patterns are strategies used continuously in all areas of life. While strategies for temporarily adjusting them can be taught (see footnote #12 for reference and explanation) they tend to be very “sticky” and difficult to change. This means that attracting and retaining women with information processing strategy aligned with engineering management needs is key. The numbers tell us that engineering will promote women who meet its cultural mandates. Engineering is not likely to sacrifice its mission objectives or cultural mandates to meet abstract societal concepts of gender equality. 

What is needed is to attract a mix of women engineers whose way of viewing the world matched the demands of the profession. All styles should be welcomed—there is a role for everyone within engineering. However, a bit extra effort focused on women with strong analytical (i.e., HA) and idea generating (i.e., RI) capacities would help fill that portion of the pool from which management is drawn. This will probably require efforts by all involved—academicians, professional associations and current engineering management. 

The Engineering Insights (see footnote #3) study argued that engineering’s cultural norms are near ideal and rooted in the demands of the profession. That means adjusting cultural norms and standards is probably ill-advised. Rather strategies have to be devised that attract, retain and develop women with the requisite characteristics.  These strategies are not likely to come from the management structure of Chief Engineer on down. These levels all share a common information processing structure. This creates a common view and limits the range strategies which might be deployed. 

The Engineering Insights study does offer a potential avenue for remedying the condition. The Vice Presidential level is populated by individuals with a strong Relational Innovator (i.e., RI) component. These people have the predisposition to generate options and ideas which can create a favorable environment while maintaining the highly successful core engineering culture.

The strategy suggested by this research is for the Vice Presidential level to reach down and adopt the development of Professional Level talent as a major responsibility. This is likely to be a stretch for the VP. Their natural interactions will be with the various levels of engineering management. Adopting this responsibility will require them to extend that interaction down to a far more populous professional level (see footnote #13 for elaboration). 

The delegation of this responsibility to the VP makes sense. In addition to the ideas and options, the VP also has the authority and resources to actually effect the changes that they deem warranted. The specific strategies are likely to be unique to each organization. A “cookbook” prescription is probably not needed. The strong idea-generating RI capacities of the engineering VP found in the Engineering Insights study suggest that there will be no shortage of potential remedies.

<1> Salton, Gary (April 2014). “The Engineering Personality.”
       A textual version can be found in our research blog at:
A video describing the research can be found on YouTube at
       It is also available in the Coffee Break Videos section of

<2> Salton, Gary (January 2015). “Women in Engineering.”
       A textual version can be found on our Research Blog at
A video describing the research can be found on YouTube at
       It is also available in the Coffee Break Videos section of

<3> Salton, Gary (July 2015). “Engineering Insights.”
       A textual version can be found on our Research Blog at
       A video describing the research can be found on YouTube at
       It is also available in the Coffee Break Videos section of

<4> The Student’s-t test was used to estimate significance. 
Technically there is no minimum sample size when using
this statistic. The primary effect of low sample size is the 

loss of statistical power. Power is the probability of saying 
that there is no effect when one is actually present (i.e., a Type II error).
The low sample size among senior Engineering Executive (n=9) 
and 1st Level Executives (n=6) suggests that the finding of no 
significant different in these categories should be taken as a tentative 
indication rather than a definitive finding.

<5> Salton, Gary (February 2008). "I Opt" Strategic Styles and Patterns. This video provides an orientational overview of how styles are used to represent information processing postures. A complete explanation of the theory and application of the technology is available in a ~17 hour online/telephone counseling certification program conducted on demand by Shannon Nelson (


<6> A battery of I Opt “snowflakes” that are predictive of observable qualities and which cover subjects like general behavior, learning, communication, emotional impact, corporate culture and general culture are available free of charge at
<7> The “p=” value in italics below the significance text is the decimal probability that the difference discovered between males and females is just a random variation in scores. The academic standard for marginal significance is p=.05 or a 5% chance that the difference is due to chance variation rather than a causal difference. In the case of “I  Opt” styles the differences in the LP and RI styles would qualify as “highly significant.”  
<8> Salton, Gary (November, 2006). Gender at the Executive Level.
A 9 minute video is available on YouTube at
A textual summary of the research can be found on the Google Research blog at;postID=116370873786509038;onPublishedMenu=posts;onClosedMenu=posts;postNum=31;src=postname
<9> It is worth noting that the sample proportions in this study cannot be used to infer causes of differentials in female versus male participation rates between various engineering levels. The data is roughly random within categories but not between them. There is no assurance that the categories are comparable in terms of migration from one level to another.  However, legitimate inferences can be made between males and females within a category.

<10> The marginal significance of the Perfector pattern discrepancy in mid-level management is due to a combination of style differences.  Women have less commitment than do men in both the analytical HA and idea-generating RI styles (short-term decision strategies).  Neither of these reaches statistical significance at the style level.  However, patterns multiplicatively combine styles. This combinatorial process is what elevates the “Perfector” pattern to significance at mid-level.  Practically, this illustrates the importance of small differences.  Women will be considered for promotion based on style. But they may tend to lose out to men who are likely to be seen as being better equipped to handle the variety of decisions required by the higher level job (i.e., a better aligned pattern strategy).
<11> Bureau of Labor Statistics, 2014;HOUSEHOLD DATA ANNUAL AVERAGES; 11. Employed persons by detailed occupation, sex, race, and Hispanic or Latino ethnicity. Accessed: June 29, 1015:

<12> Everyone has access to all information processing styles. People can elect to use one or another to facilitate particular transactions.  Booklets describing how this can be done are available in the “So You’re a …” series which can be seen at  However, people make thousands of unique decisions every day and it is impossible to guide each one through rational selection. There is no choice but to rely on a generalized strategy as the guide for the conduct of life. Since all aspects of life are involved adjusting the global strategy is a major undertaking that takes much time and effort to effect. 

Reliability studies have determined that changes in strategy are infrequent and when they do occur they tend to follow a predictable pattern that may not be aligned with a rationally selected target (e.g., preparation for engineering leadership). These studies can be seen on YouTube at:
“I Opt” Style Reliability Stress Test:
“I Opt” Pattern Reliability Stress Test:

Textual versions of these studies can be found on the “I Opt” Research Blog at:
“I Opt” Style Reliability Stress Test:
“I Opt” Pattern Reliability Stress Test:
<13> This interaction can take many forms.  Periodic luncheons with a rotating sample of younger professional engineers might be one course.  MBWA (Management by Walking Around) might be another. The basic idea is for the VP to become aware of and involved with the personal circumstances of the individuals at the professional level.  The technical elements of development are already tended to through on the job training and various Engineering Leadership Development Programs. The VP’s role is to make the cultural tweaks that cause women to want to stay long enough to use those technical abilities.