As Molly Rockett recently observed, the percentage of women in American state legislatures has stalled at 25% for nearly a decade since making impressive gains in the 1960s and 1970s.

After 1992's "Year of the Woman" election, the number of women in Congress increased at a steady pace for more than a decade (from the 103rd Congress through to the 110th Congress). Since the 2008 election, however, growth appears the have plateaued at 20%.

If we plot historical data on the number (and percentage) of women in the United States Congress and in state legislatures, we generate graphs that resemble different stages in a sigmoid function or “Standard S curve”, which has a stretched S shape. ​

An S curve is a common type of function, in which an upper and lower limit bound the curve. Most real-life examples have a starting value of zero, but the upper limit and slope of the curve (the width of the S) vary greatly. Adrian Bejan of Duke University explains that the S curve “is so common that it has generated entire fields of research that seem unrelated – the spread of biological populations, chemical reactions, contaminants, languages, information and economic activity.”

Interestingly, the charts showing the percentage of women in Congress and in state legislatures  closely resemble different stages of two S curves, both with lower limits, naturally, of 0% and an upper limits--a mathematical glass ceiling of sorts--of 25% women. In the chart on Congress, the first part of the S is apparent (it is a very wide and stretched out S) until about the 95th Congress (1977-1979).  The trend in state legislatures over time also has an S-shaped trend, though it is steepest about 10 years earlier than is the trend in Congress. The upper limit there also appears to be 25%.  
​
What could explain this trend? Why would the upper limit on these curves be somewhere around 25% and not around 50% where we would expect it to be based on the population? How can we fix it?

Why is was progress faster in the 1980s and 1990s?

The S curve is frequently divided into three phases: the learning phase, the growth phase, and the stability phase.  During the learning phase, gains are slow.  The first female state legislators were not elected until 1894 (in Colorado).  It took until 1917 for a woman to enter the U.S. House of Representatives, and until 1922 for the first woman to take a seat in the U.S. Senate.  During the learning phase, there were few women to serve as role models for other women seeking office, and the odds were stacked against the few that ran, and so progress was slow.  

Very gradually, more women ran and won seats in their state legislatures. By the 1960s, cultural and generational change sprouted the growth phase, in which progress was, relatively, quick.  Over time, more and more women were able to run and win, which inspired more women in turn to run and win. Women candidates learned how to use available resources to run for office, or in other words, “mastered the skills” needed to run and win in greater numbers. By the mid-1970s, the increase in women with state legislative experience meant that there was a larger selection of qualified women candidates to run for Congress.  Correspondingly, the spawned relatively rapid growth in the number of women in Congress.

Why did the number of women in legislative office level off?

It appears that in both graphs, we are approaching the stability phase, where the percentage of women in legislatures has begun to flatten.  The number of women in Congress is directly related to the number of women who win state legislative seats, as these women in lower office are the ones who gain the experience and the credentials to run for Congress.  As the number of women in state legislatures leveled off in the early 2000s, it stands to reason that the number of women in Congress would follow suit.

That answers half of our question: but what first caused progress in the state legislatures to slow?

Interestingly enough, the disparity between the two political parties in the number women who are elected to state legislative helps explain why the mathematical glass ceiling stands at around 25%.  

Until the early to mid 1990s, the percentage of Democratic women was within about 2% of the percentage of Republican women in state legislatures.  Today, that difference is  4-6%.  Assuming that the split between the parties hovers between 10 points above and below 50% (currently it is 56% for Republicans) and that the  Democratic Party elects two women for every Republican woman, even if the Democratic Party achieved parity, legislatures would only be 38% women, which is still far from parity.  The limit, therefore, is far below 50%, which is why the curves documenting progress toward parity appear to have such low upper limits.  ​

Until the early to mid 1990s, the percentage of Democratic women was within about 2% of the percentage of Republican women in state legislatures.  Today, that difference is  4-6%.  Assuming that the split between the parties hovers between 10 points above and below 50% (currently it is 56% for Republicans) and that the  Democratic Party elects two women for every Republican woman, even if the Democratic Party achieved parity, legislatures would only be 38% women, which is still far from parity.  The limit, therefore, is far below 50%, which is why the curves documenting progress toward parity appear to have such low upper limits.  

What can be done?

In the 1920s, both parties decided to mandate one man and one woman from every state would serve on their national committees. Other rules in national and state parties have existed for more than half a century to ensure parity in party committees. However, these measures to facilitate gender parity were never extended to candidate recruitment practices. Adopting voluntary measures within the party organizations that give women an equal voice in candidate recruitment or even a (self-imposed) requirement that the party support equal numbers of male and female candidates could bring representation closer to parity, while staying well within the traditions the American political parties. Furthermore, the higher levels of the political parties could give grants to state and local parties that meet agreed upon goals for recruiting women to run for office.

Multi-winner districts (MWDs) also give more women the chance to serve in office.  First of all, both a woman and a man can win an election for a MWD, so the position of already powerful male party leaders would be less threatened and less likely to directly oppose the idea of a woman running.  Furthermore, people display a diversification bias in which they make more varied choices when they can make multiple selections than they do when they are choosing just one.  Voters may be more likely to vote for a woman when voting for a slate of candidates for the same office than they would when just voting for a single candidate. For political parties, this means that the party has incentives to offer a diverse slate of candidates so that as many different types of people (both in terms of gender and race) are represented as possible. Finally, MWDs help reduce the negative tone of elections, as candidates will be more valuable if able to work with the other candidates on the slate, which is much harder after a vicious election cycle where personal attacks leave animosity between those serving side by side.  As women are more reluctant to run negative campaigns, according to Jennifer Lawless and Richard Fox, a more positive campaign environment will encourage more women to run and therefore win.

Regardless of whether our assessment of why the number of women in state legislative office has stalled is correct, the most important lesson to take away from this discussion is that the rate of increase in the number of women in legislatures has slowed to a crawl, and it will be incredibly difficult to reach gender parity in elected office if we continue to use the same strategies as we have in the path. The key to achieving gender parity is to find ways to increase that mathematical glass ceiling to 50% and to increase the rate of change, or narrow the “S” shape. It is clear that past techniques will not be sufficient to achieve gender parity in elective office. Representation 2020’s structural reforms will help us smash the mathematical glass ceiling and bring us closer to parity more quickly. 

In the 1920s, both parties decided to mandate one man and one woman from every state would serve on their national committees. Other rules in national and state parties have existed for more than half a century to ensure parity in party committees. However, these measures to facilitate gender parity were never extended to candidate recruitment practices. Adopting voluntary measures within the party organizations that give women an equal voice in candidate recruitment or even a (self-imposed) requirement that the party support equal numbers of male and female candidates could bring representation closer to parity, while staying well within the traditions the American political parties. Furthermore, the higher levels of the political parties could give grants to state and local parties that meet agreed upon goals for recruiting women to run for office.

Multi-winner districts (MWDs) also give more women the chance to serve in office.  First of all, both a woman and a man can win an election for a MWD, so the position of already powerful male party leaders would be less threatened and less likely to directly oppose the idea of a woman running.  Furthermore, people display a diversification bias in which they make more varied choices when they can make multiple selections than they do when they are choosing just one.  Voters may be more likely to vote for a woman when voting for a slate of candidates for the same office than they would when just voting for a single candidate. For political parties, this means that the party has incentives to offer a diverse slate of candidates so that as many different types of people (both in terms of gender and race) are represented as possible. Finally, MWDs help reduce the negative tone of elections, as candidates will be more valuable if able to work with the other candidates on the slate, which is much harder after a vicious election cycle where personal attacks leave animosity between those serving side by side.  As women are more reluctant to run negative campaigns, according to Jennifer Lawless and Richard Fox, a more positive campaign environment will encourage more women to run and therefore win.

Regardless of whether our assessment of why the number of women in state legislative office has stalled is correct, the most important lesson to take away from this discussion is that the rate of increase in the number of women in legislatures has slowed to a crawl, and it will be incredibly difficult to reach gender parity in elected office if we continue to use the same strategies as we have in the path. The key to achieving gender parity is to find ways to increase that mathematical glass ceiling to 50% and to increase the rate of change, or narrow the “S” shape. It is clear that past techniques will not be sufficient to achieve gender parity in elective office. Representation 2020’s structural reforms will help us smash the mathematical glass ceiling and bring us closer to parity more quickly.