The Giving Model

AUDIENCE 98’s Giving model is an statement of the interactions between givers’ motivations, mindsets, and means. The simple statement of the model in the main findings page does not convey either its finer points or the full extent of what we learned in its creation.

AUDIENCE 98’s Giving Model
Dependent Variable:
Natural Logarithm of Annual Household Gift

	Unstandardized Coefficients		Standardized Coefficients
	B	S.E.	Beta	t	Sig.
(Constant)	2.861	.099		28.941	.000
Time Spent Listening	.006636	.000	.123	4.738	.000
Loyalty	.001950	.001	.088	3.427	.001
Personal Importance of   Network Programming	.04081	.015	.063	2.672	.008
Personal Importance of   Local Programming	.03770	.013	.065	2.810	.005
Annual HH Income	.003261	.000	.271	11.219	.000
Actualizer	.108	.034	.076	3.158	.002
Joint Licensee	.114	.032	.079	3.570	.000
Station's Average Gift	.002090	.000	.117	5.219	.000

 

Reliance: Time Spent Listening by the listener to the supported public radio station is in hours per week. Loyalty of the listener is the percentage of all his or her listening to radio that is to the supported station.

Personal Importance: The Personal Importance of Local and Network Programming are measured on a six-point scale, with 6 being "agree definitely" that the station’s "network [or local] programming is an important part of my life. If it went away I would miss it."

Listener Characteristics: Annual Household Income is in thousands of dollars per year. Actualizer is dummy coded (0,1) to indicate whether the listener’s primary or secondary VALS 2 type is Actualizer.

Station Characteristics: Joint Licensee  is dummy coded (0,1) to indicate the radio station is licensed jointly with a public television station. Station’s Average Gift is in dollars; it is the sum of all respondents’ gifts to the station divided by the number of respondents giving to the station.

Definitions

Households. The Public Radio Recontact Survey asks, "How much did your household give to [station] in the year of your most recent contribution?" (emphasis added). Because the measurement is the household’s gift, AUDIENCE 98 aggregates the responses of listeners in the same household into a single response. Therefore the giving model is based upon the household rather than the individual, unlike any other analysis in AUDIENCE 98.

For a detailed discussion of "Households, Pseudo-Respondents, and the Attribution of Listener Support," see pages six through eight in the Public Radio Recontact Survey Database Toolkit.

Annual Gift. Again, the survey asks, "How much did your household give to [station] in the year of your most recent contribution?" (different emphasis added). The reported number is the sum total of gifts for the year for listeners who gave more than once, and should therefore be interpreted as an annual gift, not the amount of the most recent gift.

Current Givers. Both the "Givers" and "Giving" models created by AUDIENCE 98 focus on current givers only. The Public Radio Recontact Survey was fielded in March of 1997; current givers are those who said they "gave in 1996 or 1997." Therefore, a current giver is a person who lives in a household that has given to at least one public radio station in the last 15 months.

Control Variables

AUDIENCE 98’s Giving model acknowledges two station characteristics that make a difference in the size of listeners’ gifts. In statistical terms, we have "controlled" for the effects of these variables. This greatly strengthens the other findings to emerge from our model.

Joint Licensees. We find no evidence that joint licensees perform better as a group than other stations. However, listeners who give to joint licensees report slightly higher giving levels. We interpret this simply as listeners reporting their gift to the combined radio and television operation.

For instance, one household in Washington DC reports giving $60 to WAMU and $75 to WETA, a joint licensee. But gifts to WETA radio are also gifts to TV; for most listeners it would be impossible to apportion the gift across the two operations. Again, by controlling for this real and understandable confusion, the remainder of AUDIENCE 98’s size-of-gift findings is greatly strengthened.

The Station Itself. Each station differs from others in its tactics and ability to earn gifts from listeners. At some stations the development efforts may be more aggressive, the tactics more powerful, or the communities richer.

We find no readily apparent commonalities among stations with higer-than-average gift levels. But our model acknowledges these differences and is greatly strengthened as a result.

The control variable used is a calculation of the average station gift as calculated from the AUDIENCE 98 database itself. Using a component of the dependent variable (gift size) to predict the dependent variable introduces multicollinearity in the model. Analysis shows the multicollinearity does not significantly alter the remainder of the model.

This table shows the standardized coefficients (betas) for the model with and without the average gift variable. Note how the inclusion of the variable increases the model’s predictive power without disrupting the other independent variables.

	Model with Average Gift r2 = .152	Model without Average Gift r2 = .134
Time Spent Listening	.123	.125
Loyalty	.088	.085
Personal Importance of   Network Programming	.063	.059
Personal Importance of   Local Programming	.065	.055
Annual HH Income	.271	.267
Actualizer	.076	.068
Joint Licensee	.079	.086
Station's Average Gift	.117

– David Giovannoni
AUDIENCE 98 Core Team

Audience Research Analysis
Copyright © ARA and CPB.  All rights reserved.
Revised: September 01, 2000 12:38 PM.