I finally took the plunge into social media – not that long ago – when I realized that the only way to keep up with my local disc golf community, the Morgantown Mountain Goats, was through Facebook.
More recently, as I began doing research on disc golf, I discovered one more reason to join: Facebook may be the only available tool for estimating the number of disc golfers in the United States and learning more about how this group has changed over time.
Among adults who use the internet, 72 percent have Facebook accounts, a proportion that has remained about the same since 2014, according to a survey conducted by the Pew Research Center. A full 90 percent of young Americans use social networks. Almost all of this growth has occurred over the last ten years. Only seven percent of adults were on social media networks in 2005.
Given that disc golfers tend to be younger than the average American, it is likely that the great majority of them are on Facebook.
A quick search of Facebook groups offers evidence to support this claim. There are currently 2,870 distinct Facebook groups that have the word “disc golf” somewhere in their titles or group descriptions. Roughly 2,612 of these are based in the United States.
To better understand this population, I listed all of them in a single file, gave each one a unique ID number, and randomly selected 100 of them (see Sampling Procedure below for more details).
This sample, as I’ll argue more thoroughly in later posts, may be more representative of the U.S. disc-golf population than any other sample ever collected. It certainly contains a more diverse set of disc golfers than the samples collected by the PDGA, the State of Disc golf survey and numerous other small-scale surveys.
I’ll refer to my sample as “Disc Golf’s Facebook 100 (2016),” or the 100 for short.
A statistical analysis of the 100 can tell us several interesting things. In future posts, I’ll write about the average age and lifespan of Facebook groups, the geographical diffusion of these groups, and eventually I’ll use this sample to estimate the total number of disc golfers in the United States.
For now, I’ll begin with the issue of group size.
|Descriptive Statistics on Group Size|
|Total members in all groups combined||34,431|
|Percentage of total members in the ten largest groups||75|
|Mean number of members per group||344|
|Median number of members per group||51|
|Mean number of members after excluding outliers||174|
|Number of members in smallest group||5|
|Number of members in largest group||10,713|
Older disc golf groups and those involved in the sale of disc golf equipment tend to have many more members than newer groups located in small towns and cities. The ten largest groups account for 75 percent of the total members in the 100.
There are 34,431 members listed in all the groups combined. If all of the members in The 100 belonged to one and only one group, we could conclude that the disc golf community on Facebook is at least 899,338 members strong.
But many Facebook users are members of more than one group and some may be members of dozens of groups. I’m currently working on this problem and hope to correct for overlapping group memberships and post a solid population estimate soon.
The Ten Largest Groups in Disc Golf’s Facebook 100 (2016):
|Name||Number of Members||Year Created|
|1) Disc Golf Thrower Exchange||10,713||2013|
|2) $15 or less DISC GOLF DEALS||6,993||2012|
|3) Mile High Disc Golf Club||2,582||2009|
|4) San Antonio Disc Golfers||1,306||2008|
|5) Disc Golf History||1,177||2012|
|6) Disc Golf: Buy & Sell||818||(Unknown/Closed)|
|7) Charlotte DGC||747||2015|
|8) Devens Disc Golf||616||2009|
|9) Maricopa Disc Golf Club||478||2012|
|10) Tallahassee Disc Golf Association||447||2009|
The average group size requires a bit of interpretation. A few very large groups are skewing the distribution, which has inflated the mean. The distribution becomes more normal when the two highest and two lowest outliers are removed. After excluding these outliers, the mean (or “average”) number of members per group is 174. Of course, deleting groups from the distribution is also a problem. For this reason, the median, which is not sensitive to outliers, is the best indicator of average size of disc golf groups on Facebook.
So, if your favorite disc golf group has more than 51 members, it’s doing quite well. If not, maybe it’s time to recruit more disc golfers.
If you’d like to learn more about the details of this study, see below. If not, please follow my blog or drop by again soon. Next week (9/27/2016) I’ll be posting an article on the life and death of disc golf groups on Facebook. How long do most groups survive? It may be shorter than you think.
Thanks for reading. Don’t forget to like and follow us on Facebook.
The sampling procedure for Disc Golf’s Facebook 100 (2016) began on August 1 and was completed by August 16. The initial sample included nine international groups (9 percent), which means that the total number of American disc golf groups is roughly 9 percent smaller, or about 2,612. Given my focus on the U.S. community, I excluded the international groups and randomly selected nine more U.S. groups to reach the sample target of 100.
Another problem involved “closed” Facebook groups. Thirty seven of the 100 groups in my sample were closed, which meant, given Facebook’s privacy controls, I was not able to determine when these groups were created, who created them, or whether they were still active. For this reason, in an effort to include the closed-group population as much as possible while staying within the ethical guidelines of my academic discipline, I sent personal messages to the group administrators of each closed group, briefly described my research interest, and asked for permission to join the group.
After four days, 14 of the 37 closed groups allowed me to join and I collected their data. In a few cases, the administrator told me that the group was no longer active, and provided the approximate creation and latest post dates. I kept all the closed groups in the sample, because even closed groups provide some interesting, publically available details.
All scientific samples have limitations, including this one. The 100 is only representative of U.S.-based Facebook groups that display the term “disc golf” somewhere in their titles or group descriptions.
There may be many disc-golf related groups that only use alterative terms such as “frisbee golf” or “frolf.” These groups, by definition, would not show up in a Facebook group search using the term disc golf. This limitation, in addition to the problem of overlapping group memberships discussed above, would lead to a systematic undercount of the total number of disc golfers on Facebook. In future posts, I’ll tackle both of these problems, as well as consider whether or not “frolfers” should be categorized as disc golfers.
6 thoughts on “Is your favorite disc golf group on Facebook bigger than average?”
I think Morgantown Mountain Goats stats would be even higher if we didn’t include Auction, Company and/or Sales Groups. It would be interesting to see what percentile we would be amongst club pages.
Thanks for the comment Jimmy Mac!
To follow up on your interesting suggestion, I excluded seven groups from Disc Golf’s Facebook 100 (2016). Six of them were “Auction, Company and Sales” groups, and one was a general interest group (Disc Golf History). This left 93 groups that appear to be connected to a specific geographical location (almost all of them are connected to local disc golf clubs). Running the stats on this group offers a different picture of average group size and so forth. The distribution is still heavily skewed, but both the mean and median drop when the big trade/sales groups are excluded.
With 138 members, the Morgantown Mountain Goats, based on data from the 93 groups, would move from the 65th percentile to about the 68th percentile in terms of group size.
Any group connected to a local club or disc golf scene that exceeds 500 members is extraordinary. The Goats will get there!
Here are the new data:
Number of groups: N = 93
Std. Deviation: 319.341
Skewness: 5.480 (when greater than 1, distribution is far from symmetrical)