I have been having fun! A friend recently critiqued my favorite state's youth soccer program as being one of the poorest in the USA. I started thinking, what would be a good metric for evaluating a state's girls' soccer program. It occurred to me that the number of girls from the state on Division I schools' soccer teams, in proportion to the state's population, would be a pretty good indicator. So, I spent a couple of days collecting data. Boring for most, fun for me. I took two primary looks: One, at the top 125 teams in the 2013 ARPI (the most current full season ARPI) using those teams' 2014 rosters, representing slightly over 1/3 of the teams and pretty much representing all of the top DI teams in the USA over time; and the second, at the top 225 teams, pretty much representing all but the teams that really form the bottom of Division 1 and virtually never are legitimate contenders. (I hope I haven't insulted anyone.) Here are the results for the top 125 teams (next post will cover the top 225 teams):
Wouldn't it be better to look at the # of registered youth soccer players against the D1 number? Considering d2 and 3 players is much more work but would be more accurate. In some areas, d2 is a big deal and some entire State systems are d2 or d3 rather than d1. and there's NAIA and JuCo. Plus college soccer isn't big in some of these states so there aren't roster spots to even fill. Point is if you're trying to get at "producing top players" as a measure of State YSA/Club quality, the ratio of total youth players to elite players (d1 rosters or any college roster) would be interesting. ALSO - you could pretty easily look at YNT rosters over the past 3-4 years across age groups and see where the really Elite players are coming from.
I thought, while going through the data, that using only D1 creates a difficulty in some states. In those states, there are relatively few D1 schools; and they are big states. Once the rosters of the in-state D1 schools are filled up, going to an out-of-state D1 school means going a long distance away from home. In those states, I wondered about the extent to which good players would go to a nearby D2 or D3 or NAIA school rather than try to go D1 out of state. In some areas of the country, on the other hand, there are lots of D1 schools and the states are smaller so that going out-of-state to a D1 school is not such a big deal. I chose a metric that was relatively easy to implement and that fit with my own interest, which is D1 women's soccer. I think it's fair to say that the metric provides reasonable evidence of which states' YSAs/Clubs are producing good Division I women's soccer players, in proportion to the states' populations. In fact, it would be pretty hard to argue otherwise, except for the sample being a snapshot of only one year. The two tables I provided don't show which states are producing the "most" elite players, although I could adjust my groups such as to include only the top 10 teams, top 20, etc. Realistically, however, I felt the top 125 covers the strong conferences and the top teams from middling conferences, so it's a pretty good group to look at. My metric does not consider the numbers of YSA/Club players in the different states, although I think it's pretty safe to say that the D1 players are coming out of YSAs/Clubs. Whether the states' "per YSA/Club player" numbers would be different than the states' "per capita" numbers, I don't know. In the past, I have delved in to trying to identify the numbers of YSA/Club players in the different states, and I concluded that I would not have confidence in any numbers I might be able to find. If someone has reliable numbers for each state, however, and wants to give them to me, I would be happy to insert them into my spreadsheet and do a "per YSA/Club player" table. I sort of like the per capita approach myself because it's simple. What I did involved just under 6,000 players, which is a good sized sample. But it is just a one-year data sample, which is a problem. And, it depends on where teams fell relative to the cut-off points I used. So, it's a pretty "order of magnitude" type of look. Using YNT numbers, in my opinion, would be too small a sample to be very meaningful. Incorporating D2, D3, NAIA, and JuCo players would increase the sample size, but then one would have to decide how to value players at the different levels. At least D1 is a single level -- but of course, that's questionable, which is why I did one test stopping at #125 and the second stopping at #225 and didn't even do the teams ranked below #225. Overall, I think one has to take this simply as what it is. That's why I didn't editorialize about what it means. Those are the numbers for the exercise I went through. Get whatever meaning (or non meaning) out of them that makes sense to you. And, if you want to use a different metric, be my guest!
I think this is very thought provoking and do not mean to come off as critical at all. It was just interesting to see Oregon and Oklahoma on that chart near CA who had 10 times as many college players. Even comparing Oregon and Oklahoma I wonder how many HS athletes play women's soccer of the State's total athletes. That is, are more female athletes exposed to soccer opportunities in Oregon compared to Oklahoma? and so more 'elite athletes' choosing soccer over other choices. That is part of it for sure and I think Oregon would be regarded as a better "soccer State" than Oklahoma (no offense Sooners!) I was also thinking how could you really find out which States have well run youth soccer systems, those maybe 'doing more with less' so to speak. If you could look at total registered youth players against those on college rosters, that would be revealing. Using registered youth players would help capture "resources" as well since those players all pay a small registration fee that provides the revenue for State YSA/soccer efforts. Some Clubs/local YSA's are bypassing the State system now but every kid is registered to certify their age and paying an admin fee to someone at least for insurance purposes. As you mentioned, differences in the D1 soccer landscape skew that comparison. Some States are big DI importers and others big exporters for sure. I read once that in NJ for example, there are several times more college bound HS Srs generally then there are seats at NJ colleges. There are lots of regression formulas and a Master's Thesis (if not a Dissertation) in these questions for sure. My personal interest is in sports 'socialization' i.e. how much more likely is a female in 2014 to play competitive HS or college sports if the mother had played HS or college sports?
This might be the first generation where you can even ask that question. Joy Fawcett's daughter at UW might be the first daughter of a NAT player to play college ball.
I think there's another "socialization" question, which is how much more likely is a female to play competitive HS or college sports if the father had played HS or college sports, especially as compared to the answer to the same question 20 or 40 or 60 years ago. Based only on personal observation, I'm betting that if the data were available, the comparisons to the past would show very big changes. Or, maybe it's just wishful thinking on my part.
Cirovski is a freshman at Bucknell - Father winning National Championships at UMD and Mother was 1991 World Cup Winner- no pressure at the dinner table in that house I'm sure... I think there would be no question a strong correlation between the parents experiences, profession, and education level and the outcomes for their children. Not sure I've seen that exact kind of study for soccer in the US though.
I've been trying to locate good numbers for girls youth soccer participation by State. I can find US Youth Soccer Association data (2012) for combined boys and girls participation by State, but not just for girls. I cannot find "by state" numbers for US Club Soccer or for American Youth Soccer Association, which are the two other major groups with organized soccer programs, much less their numbers broken out for girls. Can anyone here find those numbers? Interestingly, the US YSA numbers show different levels of youth soccer participation, in relation to overall state population. I'm reluctant to use just those numbers, however, because the different "major" organizations have different levels of activity from state to state so that just using one group's numbers seems not to be a good representation of youth organized soccer participation state by state. This is the problem I alluded to in an earlier post that makes it hard to get good youth soccer participation numbers on a state by state basis.
Very interesting and thanks for sharing. I would guess that a state like Florida suffers by the denominator having a lot of retired people who aren't likely to play collegiate women's soccer. But I am share it is quite variable among states as far as sources of data for those "at risk" to play college soccer
Brian Goff posts those women's stats as part of his labor of love. (I continue to be amazed and thankful for all he does for us womens soccer fans!) Go to https://sites.google.com/site/soccerrecruits/ and click on the "Archives" at the bottom of the page and then click on the "Stats" for the year you want. (He has 5 years up now - 2010 through 2014.) There you will see a breakdown by State and also by metropolitan area. A correction. Brian has 4 years (2010-2013) with the nice by state and by metro area broken down. For 2014, he has (inadvertently?) included his player listing twice. Generally speaking the top ten states are CA, TX, NJ, OH, NY, FL, VA, IL, PA, NC The most productive metro areas consistently are: Washington DC Chicago Dallas Philadelphia LA Orange County Long Island Cincinnati Atlanta
His work is one of the great resources for DI women's soccer fans. I actually started out using his announced 2015 recruits, but decided I wasn't confident those numbers would well represent what I was looking for, which is when I decided simply to go to teams' actual current rosters since those are definitive in terms of numbers of girls actually playing DI women's soccer (as slightly distinguished from numbers of recruits). I still have some requests for information out, as it would be nice to do a "per girl involved in club soccer" comparison of the states. From the responses I've gotten to requests so far, however, I'm doubting it's possible, or at least practical, to get good state by state numbers. Either the major national programs (US Youth Soccer, US Club Soccer, and AYSO) don't break them out for girls only or their breakouts are confidential. There's one researcher who might have good numbers, but I haven't heard from him yet. I might be able to use combined boys/girls numbers as a surrogate for the girls numbers on the assumption they'll proportionally be reliable for girls, but I don't think they'll be reliable unless I have numbers at least from US Youth Soccer and US Club Soccer. Absent those numbers, the per capita numbers based on state population probably are the best I can do.
More on Brian's breakdowns... The metro areas he has identified are pretty logical if one is knowledgeable of local geography and population centers. For instance, if you know the area, it makes sense to break out LA-Long Beach, from Orange county, from Riverside-San Bernadino, (and of course San Diego.) But to east coast types, it all looks like Southern California and if that were true, then the SoCal area would dwarf the perennially #1 metro area, Washington, DC. The same is true of the Dallas-Forth Worth metro area. Then there is a state versus metro area confusion exemplified by the Philly metro area which includes some of southern New Jersey. So how one chooses to parse the areas does make a difference when looking at larger regions such as states. For example, does NJ get credit for the spillover from the soccer hotbed area of Philadelphia, or does Philly metro take credit for all of the soccer players in adjacent suburban New Jersey? I think Brian has it right by posting both lists and letting the reader study up on geography.
Yes, I realized that when I was doing the state by state comparison because Washington was getting credit for players from Vancouver, WA, which adjoins Portland, OR with a river in between and with some Vancouver players who play club in Portland -- although Vancouver also has at least one excellent club itself. Brian knows the clubs really well, so he is able to do that, which is far outside of what I know. Unfortunately, to take the current array of DI women's players and match them with "competitive areas" as distinguished from states, one would have to use the DI teams' rosters and match them up with the player names and areas that Brian has, which would be an enormous task -- because some rosters list players' former clubs but most list their high schools.
It occurs to me that I am not sure where this is headed. I agree with Eddie K that there are so may variables to take into consideration with regard to how youth soccer and player development is handled in any particular state of region of a (larger) state. And the point I try to raise above is one of which state should get credit for the good numbers coming out of metro areas because in those cases there is bound to be a lot of cross border kids travel to play in the stronger clubs in a nearby state. (eg the home address of a player may be NJ, but they may have learned their craft in a Philly based club. Same with much of New England where some kids from New Hampshire, Vermont, RI, or Connecticut may end up traveling to play for the Stars of Massachusetts club, and this is especially so for DC where most of the prestige clubs are in the burbs in MD or VA.
You're exactly right. I think you have to take the tables I posted simply for what they are -- as the locations of DI players' states before they went to college. In a very gross sense, that might give a general sense of where the strong club programs are -- either there or in nearby areas in adjoining states -- but that's about the best that you can get from these numbers. There is someone who is doing research on the number of youth soccer players each state reasonably could be expected to have, as compared to the number the state actually has. I believe the purpose of that research is to let clubs know how they are doing in terms of reaching the numbers of players they reasonably could be expected to reach. I'm hoping that the numbers states actually have will become available from the researcher, although I'm not sure that stage of the research is done yet. Again, however, my understanding is that the research is state-based. In the meantime, as I said initially, my project mostly was for fun -- and to see if a comment I heard that Oregon has one of the weakest club programs in the country was correct. (Even from the data I gathered, the comment about Oregon clearly was not correct.) In addition, it confirmed something I had guessed, which is that Utah produces a high number of female soccer players relative to its population -- and Colorado too.
I see evidence of a lot of work here. I'm glad it was also fun! Just to be the devil's advocate, I am not sure why you feel you have addressed your friend's challenge about State programs? (1) The analysis assumes population homogeneity which certainly does not exist. Kids who have no interest in soccer do not develop into competent players, regardless of the merits of the development programs they ignore. Further, the relative population demographic is the youth cohort, not population at large. Folks who retire to Florida and skew the population to an older median are not trying out for college soccer teams. (2) Players ( 2,500+) who roster for a top 125 D1 team may yet not see many (any) playing minutes, and may in fact be quite poor players, relative to the pool of highschool seniors they are drawn from. Their presence on a D1 roster may be an artifact of their state of residence (if in-State) and a certain amount of serendipity. Many of these players are, IMO, indistinguishable from good highschool players who have had strong coaches, but who never experience a US youth development program and then go on to other things after highschool. (3) I am sure you will agree a one-year time frame is dodgy at best. I know it would be just that much more work to collect 3~5 years, but that would be a much more robust sample. To really address your friend's challenge, I think you need different metrics: (1) what percent of the State's relevant demographic ever participate in program soccer (sample size)? (2) what percent of State's program graduates actively seek a spot on a varsity team when selecting college (programmatically provided motivation)? (3) which States produced those players who attended (or were recruited and declined admission) one of the NCAA final 64 in the last 3~5 years, and of these, how many were products of the State's youth development program (player quality/program impact on same)? These are much tougher data sets to compile, but they do address your friend's challenge. As I review your analysis (assuming OR is the State in question) it just looks to me like the State rostered 59 kids (who knows where) on one particular year. That would put them in the top 20 States for for D1 placement that year, but certainly not the top 10, and it really says nothing about the State's development program. I think (in the spirit of statistical fun) it is still an open question! I know this was just a fun exercise, but the numbers don't really seem (to me) to get us beyond Soccerhunter's post (above) which unfortunately makes no mention of OR.
Your points about what would be a great study are, of course, correct. On the other hand, have you seen anything better than what I've produced? If so, let's get it posted here on BigSoccer. (Maybe someone wants to go back through five years of rosters and team rankings, but it's sure not going to be me, although I gladly will provide the rankings if someone wants to identify the states that some 30,000 players came from.) In the meantime, I'll issue a challenge: If the numbers I've produced, with the various caveats I and others have suggested, don't mean anything, then show us or come up yourself (this is a generic "yourself," not you in particular UNC4EVER) with something better. It seems to me that until someone comes up with something better, what I've come up with is the best you've got, as deficient as it may be. In the meantime, take it for what it's worth. Probably not a whole lot, but probably significantly more than nothing -- especially if it rattles peoples' purely subjective opinions about how states rank in producing quality female youth soccer players. By the way, UNC4EVER, I was surprised how many Cary, NC names appear on the various rosters (a lot). There seem to be a number of locales around the country that are producing a disproportionate number of players. A really good study would be of where those locales are and of what their development programs are. And, I think points made about Florida as a retirement center may have some statistical validity. Maybe somewhere in the state by state census data, there are breakouts of state populations by age. A better metric might be numbers of children below a certain age level, rather than using statewide population. Hmm, maybe that's something that's relatively easy to find within the recent census data. It would be easy to plug that into my program. Anyone want to come up with state by state populations of girls aged 7 to 17 (or whatever the right age group would be)? In the meantime, my numbers are where the current players for the 2013 ARPI top 125 and 225 teams come from (except for foreign players). That is fact.
I enjoyed looking at your results! I certainly hope you didn't find my comments annoying, or feel I was being critical of the work you did. If so, I apologize! A decadenal demographic breakdown of population by gender is available through census data, but I am certainly not motivated enough to dig it out and would not blame you a bit if you felt the same way! Yes, Cary NC does consistently produce a number of fine young players, as a result of a good Youth Soccer program and several consistently strong highschool teams. This speaks to your friend's taunt and my second point of critique, as some of our strong players are in the YS program and some are just the beneficiaries of strong highschool coaching. When the best highschools in the area square off against one another, the contests look a lot like weaker D1 games, yet many of the players do not participate in YS programs and many do not go on to roster to play in college. IMO its hard to make that connection between youth programs and D1 matriculation on a state-wide basis. I think you are really on to something when you suggest the paradigm might be more one of "hot-spots", but of course trying to identify those with any statistical rigor would be a task and a half! Regards!
No annoyance, I'm enjoying the discussion. I'm wondering if other posters from elsewhere are aware of whether they have good high school soccer players who are not playing club. I'm pretty sure that is not the case in Oregon and Washington, it would be very unusual here. I'm intrigued to go to the census data and see what I can find. It doesn't sound that difficult and I have a high tolerance for boredom when it comes to finding D1-related data -- it's tedious, but I never now what the results are going to be so it's like opening a present. (Maybe I need a real life?)
I'm not really sure what to do with these numbers. Anbody who knows anything about soccer generally knows where most soccer players come from. I guess I could use these numbers for a couple reasons: 1. If I want my Daughter recruited I need to sell my house, find a job and move! 2. If my team loses to a club team from one of these states the Coach can always use the excuse "Well they are from a Top Ten State that is highly recruited and have more D I players"! 3. Study up on populations! Damn I didn't know Utah was so successful with DI recruits.....it must be the Great Coaching they have there! 4. BlahBlahBlah.
Here's where you can start at US Census http://factfinder2.census.gov/faces/nav/jsf/pages/searchresults.xhtml?refresh=t
Wow, take time to contribute NOTHING! A tiny bit of social science on one thread, on one topic, on one website and you have to rain on our parade. Just skip over it bro. I think if you could get to some solid empirical evidence to back your State's soccer 'bragging rights", you'd be getting somewhere. If you don't think so, just read elsewhere. The internet is a big place.