Over the weekend, seniors at Marjory Stoneman Douglas High School in Parkland, Fla. graduated. Four of those who were supposed to graduate were murdered by another student on Feb. 14. Their peers, some of whom graduated, turned that tragedy into a political movement aimed at changing the nation’s gun laws.

That effort entered a new phase on Monday, with the announcement that March for Our Lives would embark on a nationwide effort to register young people to vote — and then turn them out for the November midterm elections.

Taking on the NRA is a major challenge. Taking on the apathy and structural challenges experienced by young voters is only slightly easier.

Shortly after the 2014 elections, we created a graph showing turnout rates in California during that election. Provided by Political Data, Inc., the pattern is clear: Younger voters turned out at much lower rates than older ones. The lowest turnout rate was among 21-year-olds, about a fifth of whom voted. The highest rate was among 76-year-olds, about 70 percent of whom voted.

That 18- and 19-year-olds voted more heavily than 20- and 21-year-olds is not a statistical gray area. Young people who are first able to vote receive encouragement to register and cast a ballot that slightly older ones don’t. One reason 19-year-olds vote at lower rates than 18-year-olds, we can assume, is that 19-year-olds are first encouraged to vote when they turned 18 — a year before the first federal election.

Why older voters cast ballots more consistently is similarly not a mystery. Voting correlates to homeownership and income, since renting often means the burden of having to re-register to vote every few years and lower incomes can often mean less stable work schedules. Voting is also habitual. Young people haven’t gotten into the habit of voting.

But 2014 was also a midterm election, during which those who don’t usually vote — including and especially young people — were less likely to vote. Matt Hodges of Datadog graphed turnout by age in Ohio for the 2016 presidential election and a 2017 special election. That erosion at the left side of the chart is the drop-off in young voters.

This doesn’t only happen in Ohio or California. Census Bureau data (which has its problems) shows that in only two states did the percent of the electorate that was younger than 25 drop from 2010 to 2012, and in only one — Mississippi — did it drop from 2014 to 2016.

Young people turn out less than older voters — and that’s especially the case in midterm elections. On average, young people made up 2.6 percent less of the electorate in each state in 2010 than 2012 and 3.3 percent less in 2014 than in 2016.

The Parkland students have shown a remarkable ability to energize the country around the issue of gun control, but the real challenge has always been what happens in the November elections. Many elected officials have heard before that young people will come to the polls to ensure their defeat only to see that wave never materialize.

Shifting political rhetoric and public opinion is one thing. Changing electoral results may prove to be even harder.

Turnout estimates by age in each state since 2010

Data from the Census Bureau.