The problems are complex, but the solution is not.
http://www.singlesexschools.org/evidence.html
[div class="excerpt" style="margin-left:0; border:3px solid #000066; color:#33000; background:#F1F2F7; width:550px;"]Researchers at Stetson University in Florida completed a three-year pilot project comparing single-sex classrooms with coed classrooms at Woodward Avenue Elementary School, a nearby neighborhood public school. For example, students in the 4th grade at Woodward were assigned either to single-sex or coed classrooms. All relevant parameters were matched: the class sizes were all the same, the demographics were the same, all teachers had the same training in what works and what doesn't work, etc. On the FCAT (Florida Comprehensive Assessment Test), here were the results:
Percentage of students scoring proficient on the FCAT
boys in coed classes: 37% scored proficient
girls in coed classes: 59% scored proficient
girls in single-sex classes: 75% scored proficient
boys in single-sex classes: 86% scored proficient.
Remember, these students were all learning the same curriculum in the same school. And, this school "mainstreams" students who are learning-disabled, or who have ADHD etc. Many of those boys who scored proficient in the all-boys classes had previously been labeled "ADHD" or "ESE" in coed classes.
2008 update: in a recent report on NBC Nightly News, Professor Kathy Piechura-Couture of Stetson University , reported that over the four years of the pilot study, 55% of boys in the coed classrooms scored proficient on the FCAT, compared with 85% of boys in the all-boys classes. Same class size. Same curriculum. Same demographics.
Frequent suspension often results in the student dropping out.
One significant reason for the success of this model is that boy misbehavior which would result in a suspension in a co-ed environment isn't reacted to as harshly in a single-sex classroom.