If you don't count, your park won't count: estimating the number of city park users can be difficult, but is important to receiving funding.

How many people visit the parks in your city? Do they go once a year for a festival, or every day to walk the dog? Do they prefer a park with a playground or one with benches by the lake? What would make their experiences better?

With a few exceptions, your mayor does not know the answers to these questions. The parks department may have calculated how many people came for that big summertime concert, and it almost certainly knows how many folks sign up for yoga classes, softball leagues or rounds of golf. But those paying customers are only a tiny fraction of residents and visitors who make general use of the entire park system. The million-dollar question is what is the system's "regular" usage--walkers, picnickers, jugglers, basketballers, readers, Frisbee throwers, playground climbers, runners, tennis players, cyclists, boombox listeners, kite flyers, skaters, sunbathers, bird watchers, people watchers, squirrel feeders--on a sunny Saturday in June, a grey week-day in November and everything in between?

Part of the problem is that it isn't easy. This isn't like a movie theater with tickets, a county fair with a turn-stile, or even a national park with an entry booth. How do you count pedestrians coming freely into a park system from an infinity of entrances and engaging in a multitude of activities spread across thousands of acres?

The other part of the problem is that many park managers aren't all that interested in knowing the answer. A profit-making business counts its customers (and surveys them, which is something different) so that it can make a variety of decisions that might increase its profitability. Most park managers feel that because they aren't in the profit business, counting is an expense and a headache that they don't need.

This attitude is wrong. The concepts of "profit" and "investment" in the private sector are matched by the concepts of "benefit" and "appropriation" in the public sector. Getting a sufficient appropriation from the city council is as dependent on strong numbers, as its equivalent on the private side. Alan Tate in his book Great City Parks writes, "User counts are the only form of profit and loss account that exists in park management. It is an object lesson in the patient, persistent and professional application of sound business principles in the public realm." Numbers help managers assess the success of operation, give clues as to how they can perform better and provide benchmarks for excellence and goals to achieve. Numbers can even be profitable in unexpected ways--in Portland, Ore., the Nike Corporation has pledged a substantial sponsorship contribution to the parks department if the city can demonstrate that it has doubled attendance at its parks.

Here's the clincher: checking the verb "count" in the thesaurus yields four telling synonyms: "enumerate," "estimate," "have influence" and "be important." Important things are tabulated; things that are unimportant aren't. If park users are ever to have influence, we need to start counting.

Counting vs. Surveying

Counting and surveying park users are not the same thing. Counting is rigorous, quantitative and essentially looks back at the past; surveying represents a softer; qualitative look into the future. Both are important means, not ends; they are tools for better park management, but counting is the more telling. It's like the difference between the pre-election polls and the election itself, or the Associated Press poll of college football coaches and the bowl games. Ironically, it's easier to survey than to count.

Writers and researchers on urban park systems have been calling for greater user data collection for some time. For instance, the Trust for Public Land's The Excellent City Park System identified "user satisfaction" as one of the main tenets of a successful parks program. Knowledge of how, when and where people use parks is essential in guiding managers in directing staff time, funding and many other decisions.

Surveys

Surveys are often administered through the mail or telephone, or occasionally in the parks themselves. In a study of the nation's 50 largest cities, the Trust for Public Land found 11 that conducted user surveys, most of them in conjunction with their strategic planning process. The most frequently asked questions involved suggestions for new facilities, and almost every survey asked for an overall rating of park and recreation services.

Surveys are good mechanisms for getting need, satisfaction and trend data for parks. In comparison with physical counts, surveys are also fairly easy and quick to administer, and they can be relatively inexpensive, particularly if they are a subcomponent of a full-scale resident satisfaction survey undertaken by a city auditor department (as is done in Portland, Ore.).

In contrast to user counts, surveys also have the advantage of being able to incorporate non-users into the data collection and analysis. For instance, information about park users and non-users can be compared to census bureau information about the community at large to determine if all ethnic and age groups are being represented. Chris Walker, senior researcher at the Urban Institute, suggests using census data from the zip codes of users questioned in an interception survey. Computerized geographical information systems (GIS) can also be used to find the demographics of a quarter-mile buffer (or more or less) around the park. (For more information about GIS, see the article on page 52.)

On the other hand, surveys are not flawless. According to Walker, telephone surveys tend to be skewed toward higher-income individuals and toward frequent park users, because higher-income individuals are both more likely to use parks and to answer a telephone survey.

He also noted that people do not recall their experiences in parks very well, and respondents may overstate their use of the facilities. (A newer challenge is that more people every year are replacing their wired phones with wireless cellphones; these cannot be dialed by surveyers because each call is charged to the recipient.) Mail surveys also suffer from the sell-selection problem, exacerbated by the fact that they generate fewer total responses.

Of the 11 cities that perform surveys, none use the "interception" method, in which users are questioned in the park, although Portland plans to use this procedure in its next survey.

Finally, a telephone or mail survey cannot gather detailed information about visitation to specific parks. Most make an attempt to get broad patterns, but the information gathered from a survey cannot replace on-site observations and counts.

Counts

While straightforward in concept, counting park users is so challenging in practice that in recent years, it has become almost a lost art. Nearly none of the park departments contacted make an effort to count users beyond those that can be easily tallied through fee-paying services or gated facilities, such as swimming pools. Several managers interviewed were surprised at the suggestion of counting users in an open park, and believed that it cannot be done.

It wasn't always the case. As far back as 1871, officials with New York's Central Park were tallying those entering its 13 gates (including also horses, carriages and sleighs). Chicago and Philadelphia did so too, although their exact methods are not known. Julia Bachrach of the Chicago Department of Planning and Development attributes the interest in counting to the rise of the new fields of social science and social work. One of the reasons for creating parks, after all, was for the social betterment of urban factory workers, and social scientists were eager to measure the results. (In Chicago, Bachrach notes, meticulous numbers were kept on attendance at the "field houses"--recreation centers--because they were a new development for a park system, which previously had focused only on passive use.)

Counts, of course, can yield much more information than simply a single participant number. They can also take note of gender, age and ethnicity as well as park user activities. Gender and age are particularly significant because they have strong correlation with perceptions of safety. If a particular park count yields a high proportion of males between 18 and 45, it is likely that the park is frightening or intimidating to much of the population. Any park whose proportion of females is significantly below 50 percent probably needs detailed analysis and attention.

It is relatively easy to count park users if they come primarily by car. For instance, Rick Rowe of Virginia Beach Parks Department has the groundskeepers in his larger parks count cars twice a day and uses a multiplier to arrive at an estimate. However, he makes no attempt to count users in the smaller neighborhood parks, and it is clear that this method does not work in dense urban areas where most people travel to parks on foot or by public transportation.

Counting trail users on a linear corridor such as a rail-trail can be equally as daunting. There are multiple entry points to trails, but once users are "in" the park, they all pass a given point and can be tallied either by people or by equipment. In Indianapolis, Dr. Grey Lindsey has done extensive counts along the Monon Trail using an infrared beam that operates 24 hours a day. By assigning observers to occasionally double-check counts, it is possible to determine a formula to statistically correct for double-counts (when two users pass in front of the beam at the same moment). After four years of counts, Lindsey can easily spot (and discard) data anomalies due to malfunctioning of the infrared beam.

Of all the free, multiple-entry parks in the U.S., Bryant Park in New York City has the most comprehensive system for counting park users. Every day, a park groundskeeper walks through the 6-acre park at 1:15 p.m. (chosen as the peak point in the lunch rush) with two click-counters, one for tallying men, the other for women. He also makes note of the weather, the temperature, whether the main lawn is open and if there are any special events taking place in the park. (If there is a large event, a manager separately estimates the attendance.)

The results are graphed and used for all kinds of analysis. As a word of caution, Jerome Barth, director of park operations, notes that these numbers provide only a noontime snapshot of one park (New York has 1,699 others), and that because of the time involved the work is expensive. Barth can afford it because the park is run by the Bryant Park Restoration Corporation, a private non-profit that operates independently of the New York City Parks Department.

It should be noted that, technically, what's done in Bryant Park is called a census. A simpler and less expensive type of count is a sample. Sampling, which is how transportation departments are able to generate usage numbers for every street in a city, involves a two-step statistical process of counting many locations and then determining a set of ratios between the main locations and all the subsidiary spots. Once the ratios are set, a single count can yield relatively accurate estimates for all the others.

Then there are estimates. Several park departments conduct telephone or mail surveys and then use the information to extrapolate a guess at the numbers of users. Fort Worth, Texas, for example, learned that 66 percent of respondents use a park at least once a year, yielding an extrapolated usership of 364,000 people. Chicago's survey revealed that 91 percent visited a park at least once a year; that projects out to more than 2.5 million Chicago residents.

Unfortunately, this number has two nearly fatal flaws: it does not count out-of-town visitors and other non-residents, and it doesn't count multiple users, or what is known as user-days. A user-day is a much more realistic measure of a system's use--someone who goes to a park for a concert once in the summer counts as one; a daily park jogger counts as 365. This is why a place like Forest Park in St. Louis can register a user-day number 12 million --far higher than the population of its metropolitan area.

A Complete Park User Assessment

The city that probably does the best overall job of assessing its park visitation is Portland, Ore. Portland has consistently examined its park system on an annual basis, and it is now undertaking an even more comprehensive assessment of park users' habits and attitude.

Since 1995, the city auditor has produced an annual report on overall government performance. As a part of that, the Bureau of Parks and Recreation is audited based on staffing, spending, workload, citizen satisfaction and ability to reach major goals. (Portland also identified six other comparable cities--Charlotte, Cincinnati, Denver, Kansas City, Sacramento and Seattle--against which to compare data.) In fiscal year 2003, the audit involved a survey about all city services, which was sent to 16,000 resident households. (Its cost in mailing, printing and data entry was $28,000, not including the staff time required in writing the questionnaire.)

Beginning in 2004, with the addition of Robin Grimwade to its staff, the Bureau of Parks and Recreation is taking park monitoring to a new level. Grimwade, who came to Portland after an accomplished career of monitoring national park usage in his native Australia, is committed to using a range of methodologies in his new city.

These are:

1. Observational surveys. Observers will be stationed in parks to collect numbers of users for a given day, time and season and repeat it at regular intervals.

2. Intercept surveys. Paid staff will interview people who are in parks to find out how they got there, why they came, how long they stayed, and the importance of the park's different attributes and facilities. The survey will initially be carried out seasonally, then every three to five years in order to track changes. If no changes are noted the frequency will be reduced to once every 10 years.

3. Focus groups. Certain individuals identified from the intercept surveys will be invited to participate in an in-depth focus group discussion about park use habits and desires for future park system directions.

4. Telephone survey. Finally, a random sample of 1,500 residents--park users as well as those who never enter parks--will be contacted by telephone to get opinions about park issues. (As of yet, the content of this survey had not been finalized.)

The purpose of gathering this information is to assist Portland in marketing its parks. By generating a "psychographic profile" of park users--including even such facts as where they shop and what TV programs they watch--Grimwade hopes to improve the agency's communication, increase park attendance and help users have a satisfying park experience. The icing on the cake will be that Grimwade and his staff will also know their park by the numbers, and hopefully be able to turn those facts and figures into funding.

Peter Harnik is director of the Center for City Park Excellence at the Trust for Public Land in Washington, D.C., and author of Inside City Parks (Urban Land Institute, 2000). Amy Kimball, a student at the Yale School of Forestry and Environmental Studies, was an intern with TPL in 2004. Harnik can be contacted at peter. [e-mail] or at [phone].