Plausible new ideas don't come along frequently which makes them interesting

 Hmm.  This seems consequential research.  Intriguing, interesting, and risky.

From Network Structure and Performance by Ilse Lindenlaub and Anja Prummer.  I had never considered this comparison before.  As always, subject to replication.  From the Abstract.

We develop a theory that links individuals’ network structure to their productivity and earnings. While a higher degree leads to better access to information, more clustering leads to higher peer pressure. Both information and peer pressure affect effort in a model of team production, with each being beneficial in a different environment. We find that information is particularly valuable under high uncertainty, whereas peer pressure is more valuable in the opposite case. We apply our theory to gender disparities in performance. We document the novel fact that men establish more connections (a higher degree) whereas women possess denser networks (a higher clustering coefficient). We therefore expect men to outperform women in jobs that are characterised by high uncertainty in project outcomes and earnings. We provide suggestive evidence that supports our predictions.

I have always argued that there is a tradeoff between quality control (efficiency) and innovation (effectiveness) and this theory seems to be aligned with that observation.

For enterprises, there is a high premium to consistency and standardization.  If you input X you should always get engineering maximum Y as an outcome.  Any shortfall or variance from that leads to problem solving to bring the process into alignment with maximum feasible outcome.

This leads to high efficiency and in general more compelling profit margins, hence the strong pressure towards process compliance.  It works well where exogenous factors are steady and consistent.  Factors such as inputs, demand, regulatory context, cultural norms, cost structures, inflation, etc.  

So efficiency maximization is good in a stable environment.

But all systems have to have some variance or diversity in order to evolve over time.  Anything deviant from the norm is a cost but it also introduces the opportunity to discover a better, cheaper, or faster approach.  At the same time that we want 100% process compliance, you better also have some bare minimum of deviance.  You need to sacrifice some efficiency in order to obtain some effectiveness.

The more the external environment changes, the more you need evolution, the more you need effectiveness (evolving fit for purpose).  It might be cheaper, more consistent and more efficient to have a single aluminum provider but you probably should have 2-3 aluminum suppliers.  Sure, there will be less compliance and consistency (and efficiency) but you will gain resilience and effectiveness.

That seems to me to be in many ways analogous to the higher degree of connections versus higher clustering of connections distinction they are making.

Higher degree of network (more nodes more widely spread with lighter average investment per node) enables a broader range of information to flow more quickly - valuable attributes in an uncertain and evolving context.  And analogous to high effectiveness (capacity to adapt to changing contexts).

Higher density of networks (fewer nodes more closely related and with higher investment per node) is the analog to efficiency where you are seeking to establish compliance and conformity (valuable in stable modes).

Male networks are characterized in their findings as a higher degree - they are broader, more variant and more lightly invested in (per node).  Effective at incorporating new information from exogenous conditions.

Female networks are characterized as denser, more compact and with higher investment per node.  I.e. Efficient at forcing compliance.  

For any process, you need both higher degree networks as well as more clustered networks, each serving effectiveness and efficiency respectively and the balance between the two types of network is conditioned on the relative variance, uncertainty and volatility of conditions in the exogenous environment.

This is also highly consistent with Claudia Goldin's work.  Goldin has documented that wage differences between men and women are substantially driven by different predilections towards risk, uncertainty and adaptability.  Women, from personal choice and decisions around family welfare, tend to opt towards businesses which are highly predictable, plannable, process oriented industries such as pharmacists.  

The more stable the work, the less well compensated it tends to be.  Stable processes tend to be commodity processes.  It is not discrimination that drives gender imbalances in compensation.  It is choices around working conditions.  

Men tend to opt for more higher risk, more volatile, faster evolving sectors.  The compensation is higher in part because of the greater risk and the need to constantly adapt and evolve (a psychological and time investment).  

These two models are exemplified in Sweden.  Culturally and legally the country is highly egalitarian.  But women tend to choose to work in the government sector which is highly regulated and slow to evolve.  Men tend to work in the much riskier private sector where things tend to evolve rapidly.  

Linenlaub and Plummer are introducing a further wrinkle.  Do women develop social networks that are dense and clustered because they choose to work in stable sectors and do men develop broader more varied networks because they are working in more dynamic sectors?

Or, do men and women have some sort of genetic disposition to those different styles of network creation and is it that genetic disposition regarding networks which then drives them towards different sectors?

I have no idea.  But it seems like a very plausible factor which I have not seen much researched at all.