INTERPRETING SPORTS PARTICIPATION RESEARCH

Sports participation statistics are of interest to a great number of individuals and organizations who regularly seek information on populations of sports participants: how many people play, how often, demographic profiles, geography, cross-participation, attitudes, motivations and other characteristics. These numbers — both on individual participant sports or aggregate sector data — find a wide variety of practical and theoretical applications: sporting goods market research, business plans, sports marketing surveys, health club research, exercise trend forecasting, public health assessments, sports injury epidemiology, tourism, advertising, and academic inquiries into the nature of cultural change are but a few examples.

Sports participation rates, physical fitness research and recreational trend data are found in daily newspapers, radio, TV, and now, on the most powerful information regenerator in history — the Internet. Most often, the numbers are pursued and reported conscientiously — with care and accuracy. Sometimes, the search and presentation is less fastidious; and on occasion, numbers are tweaked in support of a preconceived storyline. But very often, survey data are simply misunderstood or misinterpreted. There are four basic guidelines that govern sensible interpretation of survey research in the sports marketing context:

1. A Survey Is Not a Census

There are two ways of measuring sports participation behavior in the entire U.S. population: a complete census of all Americans, or a consumer survey which examines a representative sample of the population, and is able (with varying degrees of accuracy) to make reasonable estimates of its sports-related characteristics.

In an absolutely perfect world, the U.S. Department of Commerce would oblige ASD with a battery of sports participation questions added to the next Decennial Census. And if we accepted other census imprecision, we might be able to state unequivocally for example, that there are exactly 34,857,014 Runners in the U.S., because we counted every single one.

Since a survey represents only a sample of a larger universe, it is subject to statistical uncertainty known as sampling error. This means that if 100 identical surveys (with the exact same format and questions) were conducted simultaneously throughout the U.S., the results of each would be somewhat different. But the differences would fall into a predictable range (larger sample sizes generally produce smaller differences) and we could talk about a "confidence level" associated with any or all of these surveys. Our limited resources allow only a survey in which one respondent in the sample represents 17,584 people on the ground, so we can offer only an estimate of 34,857,000 Runners that can vary by +/-5.6% at the "95% confidence level." This means that if we've done everything else right (and leave aside other types of non-random survey bias), the odds are about 19-1 that there are somewhere between 32,905,000 and 36,809,000 Runners in the U.S.

2. To Be "Real," Year-to-Year Change Must Be Statistically Significant

If in 2003, two of only ten Americans interviewed across the country stated that they were Snowboarders, we could not — with even a minute degree of confidence — issue a news release declaring that 20% of the U.S. population participates in Snowboarding. If in the year 2004 (through a similarly futile exercise) we found that three of ten went Snowboarding, we could not proclaim that Snowboarding participation in the U.S. had jumped by 50% between 2003 - 2004.

Offending survey samples are not quite this insubstantial, but "trend" sightings in sports participation are often just as egregious. Companies and industry groups are eager to document "growth" in their sectors; the media are in perennial need of statistical validation for preordained news themes; and research providers, through a lack of rigor, patience (or even sophistication) seem to oblige with a remarkable aversion to significance testing.

Such abuses (often unwitting) are commonly illustrated in widely publicized surveys of 1,000 Americans, featuring the well-known, but often misleading +/-3% "margin of error". A survey of this magnitude may produce a grand total of 30-35 Snowboarders — which may represent a "massive gain" of 50% over the prior year — but which also happens to be statistically insignificant, and therefore only suggestive. While Snowboarding happens to be a "real" trend (with statistically significant change over time documented in larger surveys), growth in this activity cannot be proven here. In the same study, a 10% increase in a much larger activity such as Golf might be borderline at the 95% confidence level, not a "real" increase at all, and perhaps even the statistical equivalent of a 5% decline — which might have been the result of an identical study conducted on the same day with a different sample. For demographic sub-groups, the sample size requirements become even more demanding; we need larger samples to make reliable statements about "trends" in women's or age-group participation, for example.

3. All Sports Participants Are Not Equal

In 2001, according to ASD, there were 29.4 million people 6 years or older who played Golf at least once. This vast sport population forms a wide continuum of Golfers and so-called Golfers — ranging from the reluctant duffer who played a single round on an obligatory business outing, to the retired zealot who lives in a golf course condo, travels on golf vacations, belongs to a private club and plays no fewer than 200 times a year. Yet both men enjoy equal status in the general participant population. Of our 29.4 million Golfers, 7.5 million (26%) played fewer than 4 times in the past year; 8.6 million played at least 25 times (an arbitrary definition of "frequent" golfer); and nearly 1.9 million (6%) spent over 100 days on the links.

Of 34.9 million "Runners," only 10.3 million logged in excess of 100 days during 2001, qualifying for the designation of "frequent" participant in this activity. 100 days of Running per year — Herculean by Couch Potato standards — is derided as minimalism by Serious Runners, who in addition to year-round training regimens of 30-50 miles per week, buy 4-5 pairs of running shoes per year, participate in road races/marathons, and subscribe to Running magazines. Whereas 6.4 million people ran on at least 150 occasions, (still not the province of elite athletes) the exclusive fraternity of "Serious Runners" numbers perhaps 2 - 3 million.

4. Analytically, Frequent Participants Are Surrogate Product Buyers

In general, roughly 20% of all participants in any sports or fitness activity constitute the core or "heavy users" of products and services. These dedicated amateur athletes are the backbone of the sporting goods industry — representing a disproportionately high percentage of sales and repeat business. Very often, these "core participants" are market leaders and trendsetters emulated by the great mass of casual players. Analytically, frequent sports enthusiasts are surrogate product buyers. Using approximately the top quartile of participation frequency in each sport to define "core" participants, we usually find these groups to be excellent proxies for sporting goods customers — both in demographic/psychographic profiles and absolute numbers.

For example, we can point to a rough equivalence between the 4.3 million frequent soccer players (52+ days per year) and the combined memberships of USYSA, AYSO and SAY Soccer. A projection of 3.0 million frequent baseball players (52+ days) dovetails neatly with an aggregate membership of 4 million in Little League, PONY, Babe Ruth and other major organizations. 3.5 million frequent Tennis players (25+ days) aligns with annual sales of roughly 4 million tennis racquets. Similarly, the estimate of 1.4 million frequent Skiers (15+ days) is consonant with the sale of a million pairs of skis; and in the same sphere, 1.2 million snowboarders reflect the purchase of over a half million boards.

Marketers of athletic footwear, sporting goods, athletic apparel, and fitness products need to profile the sport's participants in order to understand the nuances of participation behavior. A knowledge of the prime consumer is just as important as market share, price points, retail distribution and other sports marketing intelligence.