Classical Music Consumer Segmentation Study 2002 [United States] (ICPSR 35535)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

The Classical Music Consumer Segmentation Study was undertaken to improve understanding of symphony orchestra audiences and markets and to develop a conceptual model of the market place that can assist orchestras in "capturing additional market potential".

Study Design View help for Study Design

For the National Survey, a series of four focus groups were conducted in November 2000 with orchestra ticket buyers and prospects (non-buyers) in Charlotte and St. Paul. The overall purpose of the focus group research was to test protocol for the national telephone survey and to probe specific behaviors, attitudes and opinions related to classical music participation. Ipsos Reid conducted telephone interviews with a random national sample of 2,200 people in March 2001, using random-digit dialing. All respondents were asked a battery of core questions on classical music participation. A subset of 1,295 respondents who qualified as "potential classical music consumers" were asked a more detailed set of questions, in addition to the questions in the core survey. In February 2001 the national public telephone survey protocol was tested in Charlotte and St. Paul. A total of 700 interviews were conducted across the two markets. Results from the pre-test were analyzed, and the protocol was refined. Ipsos Reid completed the fieldwork. For all of the studies, an unknown amount of bias from non-response remains in the data. Other sources of bias must be recognized, including the acquiescent response phenomenon, in which respondents provide answers that they think the interviewer wants to hear. While the protocol was carefully designed to provide respondents with unbiased response choices, we cannot guarantee that respondents provided completely accurate answers. For example, in the national survey, respondents were asked if they have a favorite classical music composer and composition. Of the 43 percent of potential classical consumers who indicated that they have a favorite composition, 22 percent could not name one. Of the 24 percent of potential classical consumers who indicated that they have a favorite composition, 22 percent could not name one. It is not known if these respondents provided acquiescent responses or whether they could not spontaneously identify their favorite composer or composition. Another source of bias in the data relates to estimates of frequency of doing certain activities and is discussed in a footnote on page 75 (available in this collection). Other research suggests that some respondents "telescope" their activity (i.e. compress time, such that they might count an activity that happened 14 months ago in an estimate of activity over the previous 12 months). For the 15 Market Areas Survey, approximately 750 telephone interviews were completed in each of the 15 markets between August 2001 and March 2002. In sum, a total of 11,318 interviews were completed. The geography to be sampled in each market was determined through an analysis of the orchestra's actual customer records and was defined as the area from which the orchestra draws approximately 85 percent of its ticket buyers. Orchestras whose market areas were surveyed: Brooklyn Philharmonic Orchestra, Charlotte Symphony Orchestra, Colorado Symphony Association, Detroit Symphony Orchestra, Fort Wayne Philharmonic Orchestra, Kansas City Symphony, Long Beach Symphony Association, Louisiana Philharmonic Orchestra, New World Symphony (Miami-Dade County, FL), Oregon Symphony Association (Portland, OR), Philadelphia Orchestra, Saint Louis Symphony Orchestra, Saint Paul Chamber Orchestra, Symphony Society of San Antonio, and Wichita Symphony Society. During the survey's time, the telemarketing environment was increasingly difficult and prohibitively expensive to reach every household that was randomly dialed. The experience was that people with an interest in the survey subject matter tend to complete culturally-oriented surveys, despite the researcher's best efforts to complete the interview with eligible respondents selected at random. For example, people with higher education levels tend to complete general population surveys related to arts and cultural activities. As described above, generally accepted weighting techniques were used to reduce bias from self-selection. Ipsos Reid conducted telephone interviews with random samples of approximately 750 people in each of the 15 orchestra markets. All respondents were asked a battery of core questions on classical music participation. A subset of roughly 60 percent of respondents who qualified as "potential classical music consumers" were asked a more detailed set of questions, in addition to the questions in the core survey. The protocols were nearly identical to the national survey protocol with the addition of series of questions about respondents' connections to the local orchestra. Also, each orchestra was allowed a small number of discretionary questions. To reduce the effect of extreme observations on the calculation of mean values for some numeric variables, limits were set for acceptable answers. For Ticket Buyers Survey, a total of 1,500 machine-readable questionnaires were mailed to a random sample of each orchestra's subscribers (750) and single-ticket buyers (750) in each of the 15 orchestra markets between August 2001 and February 2002. Of the 10,098 valid responses received, 5,553 were from current subscribers, 1,657 from former subscribers who are current single-ticket buyers, and 2,888 from single-ticket buyers who are not former subscribers. The mailings occurred in three cycles between August 2001 and February 2002. Each orchestra was assigned to one of the three survey cycles. Prior to each cycle of survey mailings, advance notice postcards were mailed to all respondents, inviting their participation in the survey effort. Several weeks after the first mailing, a second survey package was mailed to everyone who had not yet responded. In order to improve the quality of the data, every form that was returned was edited by hand to correct markings that were either too light or incomplete and to erase stray marks on the forms that would cause scanning errors. Verbatim answers to open-ended questions were also entered at this time. To reduce the effect of extreme observations on the calculation of mean values for some numeric variables, limits were set for acceptable answers.

Sample View help for Sample

For the National Survey, the margin of error for the total sample is plus or minus 2 percentage points at the 95 percent level of confidence. For the subset of 1,295 respondents who qualified as "potential classical consumers," the margin of error is plus or minus 3 percentage points at the 95 percent level of confidence. For the 15 Market Areas Survey, the sampling methods were identical in each market. The margins of error for the local surveys were plus or minus four percentage points, at the 95 percent confidence level.

Time Method View help for Time Method

Universe View help for Universe

All adults in the United States.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Description of Variables View help for Description of Variables

The National Survey collected data from respondents on attitudes, behaviors, and opinions related to classical music participation. In addition to these questions, a subset of respondents who qualified as "potential classical music consumers" were also asked a series of questions about factors that affect their decisions to attend cultural programs and about their relationship with the classical music art form. The 15 Market Areas Survey collected data in each of 15 orchestra markets on attitudes, behaviors, and opinions related to classical music participation. In addition to these questions, a subset of respondents who qualified as "potential classical music consumers" were also asked a series of questions about factors that affect their decisions to attend cultural programs about their relationship with the classical music art form. The Ticket Buyers Survey collected data from orchestra subscribers and single-ticket buyers in each of 15 orchestra markets on attitudes, behaviors, and opinions related to classical music participation. Questions asked in these surveys were similar to the interview protocols for the national and local market surveys also conducted for this study.

Response Rates View help for Response Rates

The National Survey contained a response rate of 35 percent. For the 15 Market Areas survey, cooperation rates for the general population telephone survey are illustrated in Appendix Table N2 in the final report. Rates averaged 38 percent across the 15 markets. For the Ticket Buyers Survey, the overall response rate was 45 percent, and ranged from 35 percent to 60 percent for the individual orchestras. In sum, a total of 10,098 valid responses were received, including 5,553 from current subscribers, 1,657 from former subscribers who are current single-ticket buyers, and 2,888 from single-ticket buyers who are not former subscribers. Identical procedures were followed for each orchestra.

Original Release Date View help for Original Release Date

2015-03-19

Version History View help for Version History

Weight View help for Weight

For the National Survey, to help correct for bias from non-response, mathematical weights were applied to the national telephone survey final data set to align the results with several key demographic figures for all adults in the applicable geography: age, race, income, education, and gender. The data were weighted using the strata weighting method. Strata weighting was selected to reduce the effects of interaction between variables. Thirty-two different strata were created using dichotomized variables: Age (55 and over/54 and under), Education (High School or less/Some college or more), Income (Less than $75,000/$75,000 or more), Race (white/all others), and Gender (Male/Female). For each possible combination, the percentages of the five variables were multiplied together to form the population incidence. Dividing the population incidence by the sample incidence produced the initial weight for each respective stratum. Additional weighting was conducted to account for missing data. The final weight variable brought all demographic variables to within 1 to 1.5 percentage points of the population incidence. Generally, the weighting procedure had the effect of lowering most indicators of arts participation. For 15 Market Areas Survey, to help correct for bias from non-response, mathematical weights were applied to the telephone survey final data sets (national and 15 local) to align the results with several key demographic figures for all adults in the applicable geography: age, race, income, education, and gender. The data were weighted using the strata weighting method. Strata weighting was selected to reduce the effects of interaction between variables. Thirty-two different strata were created using dichotomized variables: Age (55 and over/54 and under), Education (High School or less/Some college or more), Income (Less than $75,000/$75,000 or more), Race (white/all others), and Gender (Male/Female). For each possible combination, the percentages of the five variables were multiplied together to form the population incidence. Dividing the population incidence by the sample incidence produced the initial weight for each respective stratum. Additional weighting was conducted to account for missing data. Each market averaged about 175 cases with missing weights after the initial procedure because these respondents did not answer all of the demographic questions and therefore were not included in the initial strata weighting. The final weight variable brought all demographic variables to within 1 to 1.5 percentage points of the population incidence. The Ticket Buyers Survey does not contain any weight variables.