Please note that race and ethnicity are not key indicators of general purchase behavior, and thus African Americans and Hispanics are somewhat underrepresented in the sample.
Prior to 2010, this annual study was conducted using a mailback survey methodology. Readers should keep this in mind when making direct comparisons between 2010 and data published prior to 2010.
B. Questionnaire
In February, sample members of the TNS panel were invited to complete the sports participation survey online. Respondents were asked to identify up to three other household members ages seven or older, indicate their age and gender, and provide the number of days of participation for each sport.
In order to reduce a bias effect that is possible from the order in which the sports were listed, the list was shown in alphabetic order to half of the households, and in reverse-alphabetic order to the remaining half of households.
C. Returns
The overall survey response rate was 35%. The study results are based on over 25,000 individuals in these households who are age 7 and older.
In order to ensure returns were representative of the U.S., a 3,264-cell weight matrix was used. The demographic characteristics used in the weight matrix were: state of residence, household income, population density, and presence of children in the household.
The returns were then projected to 280,215,000 - the U.S. population age 7 and older. A 16-cell age/gender weight matrix was used for projecting the weighted sample to represent the actual age and gender distribution of the U.S. population.
III. STATISTICAL NOTES
It is important to note that in basing estimates of a population on a sample drawn from that population, a degree of sampling error is possible. While a number of biases can affect the extent of sampling error, it is to a great degree dependent on the size of the sample.
The larger the sample, the more closely the estimate represents the total population. Sampling error also affects samples of the same size. A sample drawn from a population will result in an estimate for a characteristic of the population, the proportion of archers for example. If another sample were drawn, it would be likely to include different members of the population. This would yield a slightly different estimate of the proportion of archers.
To allow judgments and decisions to be made with confidence when using samples, we can determine a confidence interval for the estimate.
A confidence interval is simply a range. A 95% confidence interval for the proportion of archers in the population is a range in which the estimates for the proportion of archers would fall 95% of the time, were the study to be done repeatedly. For example, in a particular year, the proportion of respondents who stated they participated in archery may be 2.3% (after weighting and projection). Thus, if we were to estimate the proportion of archers in the United States using 100 similar studies, 95 of those studies would yield an estimate within 95% confidence interval of 2.2%-2.4% (+/-0.1%) from the stated proportion.
The confidence interval for measurement of the proportion of participants in each sport in "Series I" does not exceed +/-1.0%.
IV. DATA INDICATED IN THE REPORTS
A. Participants
A participant is defined as an individual seven years of age or older who participates in a sport more than once a year for all sports in "Series I" except aerobic exercising, bicycle riding, exercise walking, exercise with equipment, running/jogging, swimming, and weightlifting. For these fitness activities, participation is defined as six times or more during the year.
Note that this study also reports separately the number of individuals who participated "1 Day Only during the year” or "1-5 Days during the year" for each sport. These are not included in "Total Respondents in 2010" for each sport since they are not "participants" as defined above.
The definition of "Frequent", "Occasional" and "Infrequent" participants varies for each sport and is indicated in the report. Generally, these categories are defined as the top 25%, the middle 50% and the lower 25% of participants. For sports where the incidence of participation is low, participants may be divided into "Frequent" and "Occasional" only. In the Summary Tables, “Did Not Participate” refers to those who did not participate in any of the sports/activities in Series I.
B. Means & Medians
Following the frequency distribution, the total participant days and the mean and median number of participant days is reported for each sport. The mean (average) represents total participant days divided by total participants. The median represents the mid-point, i.e. 50% played more and 50% less than the median.
The number of participant days for certain sports should be divided by an estimate of the number of participants who play together on a team or in an event in order to calculate the number of games, matches, etc. played.