Printer Friendly

Participation in the next generation of agriculture conservation programs: the role of environmental attitudes.

E. Jane Luzar [*]

Assistant dean and former Graduate Research Assistant, College of Agricultural and Life Sciences, University of Florida, Gainesville, Florida 32611-0270


This research provides a behavioral analysis of wetland owners' decision to participate in the Wetland Reserve Program (WRP). The study identifies the WRP as an incentive based land use program that relies on voluntary participation. The role of environmental attitudes in voluntary environmental program participation is addressed using the theory of reasoned action. A probit model is used with data collected from a mail survey of Louisiana wetland owners. Results of the analysis indicate that positive attitudes toward the WRP, acreage of wetlands owned, the level of information about the WRP, ownership of farmed wetlands, involvement in environmental organizations, and income level had a significant and positive influence on the decision to offer to participate in the WRP. Higher levels of education, and number of people living in the household (potential heirs) had an adverse effect on the likelihood to offer wetlands for enrollment in the WRP. (c) 1999 Elsevier Science Inc. All rights reserved.

Keywords: Wetlands reserve program; Incentive-based mechanisms; Environmental attitudes

1. Introduction

The search for environmental management tools able to efficiently address current environmental problems, including the negative role often played by agriculture, has resulted in increased consideration of alternative management strategies in environmental policy directed at agricuture (Anderson & Leal, 1991; Carlson, Zilberman, & Miranowski, 1993). Market based environmentalism is one alternative especially appealing to economists and to policy makers reluctant to extend or expand regulatory management of the environment. Contrary to sometimes ad hoc command and control regulatory policy, free market environmentalism relies on market forces for environmental management. Specific environmental management instruments used within this framework are called incentive-based mechanisms (IBM) or market-based incentives. This approach to the environmental management of agriculture, and especially its conservation efforts, is representative of the next generation of agricultural policy.

2. Incentive-based mechanisms

Incentive-based mechanisms can take various forms. They either rely on the "polluter pays" principle or provide economic compensation to elicit a targeted environmental behavior. IBM strategies influence the behavior of individuals, households and firms by creating a system of penalties and rewards. IBM require clearly defined property rights and make each resource user, owner, or trustee bear the full cost and/or reap the entire benefits of her actions. IBM strategies include green taxes, deposit-refund schemes, emission fees, and individual transferable quotas (Dudek & Palmisano, 1988; Stewart, 1988; Tietenberg, 1990).

In agriculture, market-based environmental strategies have been mainly used to promote desirable behavior through the use of positive economic incentives. For example, in agricultural land policy, conservation and restoration practices are increasingly elicited through the transfer (TDR) or purchase of development rights (PDR). Under a PDR system, the private land owner "voluntarily sells the development rights (also known as conservation easements) and receives compensation for the development restrictions placed on the land" (Daniels, 1991, p. 421). Development rights are. Under TDR arrangements, incentives are provided for temporary development restrictions. TDRs and PDRs were initially created to protect farmland against rapid urbanization and were mainly used as state and local land use policies. Provisions of the 1985 and 1990 Farm Bills revived and reoriented the focus of TDRs and PDRs from state level farmland protection to national conservation and restoration practices. For example, the Conservatio n Reserve Program (CRP) provides compensation, i.e., green payments, for temporary development rights restrictions on highly erodible agricultural land, whereas the Wetlands Reserve Program (WRP) provides green payments for either permanent or long term easements for wetlands.

The Wetlands Reserve Program, designed after evaluation of the CRP, purchased permanent rather than temporary easements on wetlands. Under the provisions of the WRP, the government was authorized to purchase easements from owners of eligible land who voluntarily agreed to restore and protect farmed wetlands and eligible adjacent acres (USDA, 1992). In this phase of the WRP, the USDA purchased permanent easements from participating wetlands owners offering farmed wetlands, prior converted wetlands, and riparian areas linking wetlands. The USDA paid landowners a fair market value for the enrolled acreage and up to 75% of the costs of wetland restoration to approved wetland conditions. Under the supervision of the Natural Resource Conservation Service (NRCS), the enrolled acreage from this phase of the WRP must be kept in a wetland condition in perpetuity (Carey, Heimlich, & Brazee, 1990).

The implementation of incentive-based conservation programs such as the WRP constitutes another step in the evolution of agricultural environmental policy. The CRP, the Water Bank, and the WRP programs provide positive economic incentives to farmers willing to adopt conservation or restoration practices. However, due to the voluntary nature of incentive-based programs, acceptance and thus participation in these programs is not automatic. Resistance to the use of markets as an environmental management tool may for example, hamper the implementation of incentive-based programs. Reasons for the unwillingness to allow market forces in environmental resource management may alternatively stem from the strong ideological and political attachment to regulation existing in the U.S. (Stewart, 1988) or from a mistrust of market forces (Kelman, 1981).

Attitudes towards the use of markets in environmental management may also play a determining role in the effective implementation of future IBM programs in agriculture. However, to evaluate the role of attitudes in participation decisions, especially those involving IBM's in the context of environmental management, it is necessary to move beyond the simple four item Likert agreement scale frequently used to represent and measure attitudes (Upmeyer & Six, 1989). Clarifying the conceptual and empirical relationship between attitudes and observed behavior provides a basis for theoretically appropriate model specification and empirical evaluation. This study addresses the attitude-behavior relationship in an empirical analysis of WRP participation, shedding light on the role attitudes play in environmental program participation decisions. It also provides a conceptually appropriate method for measuring attitudes within the framework of mail surveys.

3. The attitude-behavior relationship

A dominant part of the social psychology literature focused on behavioral research has established the role of attitudes as predictors of behavior (Ajzen & Fishbein, 1980; Ajzen, 1988; Fazio, Powell, & Williams, 1989; Heberlein, 1989; Upmeyer & Six, 1989). Fundamental to this social psychology literature connecting attitudes and behavior is the theory of reasoned action proposed by Ajzen and Fishbein (1980). The theory of reasoned action considers individuals as rational agents using at any given point in time the information available to them. The theory argues that an individual's intention is the prime determinant of his behavior or action. Intentions are determined by an attitude toward the behavior and a subjective norm. Attitudes refer to "a person's judgement that performing the behavior is good or bad" (Ajzen & Fishbein, 1980, p. 6). More formally, Ajzen and Fishbein (1980) define attitude as an individual's assessment of a psychological object. A person's perception of social pressure exerted on him to perform a behavior constitutes his subjective norm (Ajzen & Fishbein, 1980). Beliefs associated with an individual's attitude are his behavioral beliefs. Normative beliefs are defined as "beliefs underlying a person's subjective norm" (Ajzen & Fishbein, 1980, p. 7). Although one can hold a multitude of beliefs about a given behavior, research suggests one typically can only concentrate on a limited number of beliefs, referred to as salient beliefs (Ajzen & Fishbein, 1988; Miller, 1956; Ajzen, 1991).

Fig. 1 provides a schematic representation of the theory of reasoned action. Behavioral intention, the precursor of actual behavior, is shown as a function of the individual's attitude towards the behavior as well as the individual's subjective norm. One's attitude towards a behavior is determined by two components, salient behavioral beliefs and the subjective evaluations of those beliefs (Ajzen, 1991). Similarly, one's subjective norm is determined by his salient normative beliefs and his corresponding motivation to comply (Ajzen & Fishbein, 1988).

The model derived from the theory of reasoned action can be expressed as follows (Upmeyer & Six, 1989):

B = [w.sub.1]BI + ([A.sub.b])[w.sub.2] + (SN)[w.sub.3]. (1)


B = Overt behavior

BI = Behavior intention

W1 = Empirical weight attached to BI

(Ab) = attitude towards a behavior B, defined as [sigma] Bi Ei where,

Bi = Belief that a behavior will lead to outcome I

Ei = Evaluation of expected outcome I

w2 = Empirical weight attached to Ab

(SN) = Subjective norm, defined as [sigma] NBi MCi where,

Nbi = Perceived expectation of referent I

Mci = Motivation to comply with referent I

w3 = Empirical weight attached to SN

In this mathematical formulation of the theory of reasoned action, behavior is expressed as a linear function of behavioral intention, attitude towards the behavior considered, and subjective norm. Each of the explanatory variables is weighted by an empirically determined coefficient. Nothing in the theory of reasoned action precludes one from selecting alternative functional forms, including nonlinear functions, to describe the relationship between attitudes, behavioral intention, subjective norm, and observed behavior.

Extensions of the theory of reasoned action have been explored both empirically and conceptually. For example, Ajzen, (1988) adds perceived behavioral control to the original Fishbein--Ajzen model, resulting in the theory of planned behavior (Fig. 2). The theory of planned behavior allows a better evaluation of human behavior in cases when individuals do not enjoy full volitional control, i.e., when participation decisions are voluntary and completely under an individual's control. More formally, perceived behavioral control refers to "beliefs regarding the possession of requisite resources and opportunities for performing a given behavior" (Madden, Ellen, & Ajzen, 1992, p. 4). Because it allows the inclusion of additional explanatory variables, the theory of planned behavior is more flexible than the original theory of reasoned action. For example, to account for the financial constraint of the decision maker, available financial resources can be included as a proxy for actual control (Beck & Ajzen, 1991).

The theory of reasoned action is not limited to a specific type of behavior and has been applied to a wide and diverse set of issues, including analysis and evaluation of leisure participation (Ajzen & Driver, 1991; 1992), dishonest behavior (Beck & Ajzen, 1991), job searching (van Ryn & Vinokur, 1990), voting choice (Watters, 1989), and class attendance (Ajzen & Madden, 1986). Agricultural applications dealing with environmental attitudes have focused on adoption decisions, including farmers' adoption of soil conservation practices (Lynne & Rolla, 1988), and water conservation techniques (Lynne, Casey, Hodges, & Rahmani, 1994). Drawing on this social psychological framework for defining and measuring attitudes, the following section presents a behavioral model for wetland reserve program participation.

4. The behavioral model

In the behavioral model of wetlands owners' voluntary decision to offer acres of wetlands for enrollment in the WRP, offers of participation and noninvolvement in the program are the two alternative choices available. The evaluation and prediction of discrete choice behavior such as offers of participation in the WRP can be expressed in a general form as a function of economic and socio-economic variables. Modeling choice behavior within the expanded behavioral economic framework proposed here requires inclusion of an additional class of independent variables, psychological constructs, developed by using the theory of planned behavior.

To evaluate the choice behavior relevant to WRP offers of participation, the dependent variable, choice, can be expressed as a binary variable and estimated using probit analysis. Estimation of the probit model can be accomplished using a maximum likelihood approach which yields likelihood estimators that are consistent, asymptotically normally distributed, and asymptotically efficient (Amemiya, 1981; Capps & Kramer, 1985; Judge et al., 1988; and Maddala, 1991). As illustrated by the likelihood function, the estimation of probit models via maximum likelihood normalizes the variance to one for identification purposes so a homoscedasticity assumption is not required (Aldrich and Nelson, 1984; Greene, 1993; and Windmeijer, 1995). The log likelihood function is:

l = [[[sigma].sup.T].sub.i=1] [y.sub.i] ln F([x'.sub.i][beta]) + [[[sigma].sup.T].sub.i=1] (l -- [y.sub.i]) ln [l -- F([x'.sub.i][beta])] (2)

where the binary dependent variable, the explanatory variables, and the vector of parameters to estimate are represented by, [Y.sub.i], [x.sub.i], and [beta], respectively and F is the standard normal cumulative distribution function (Judge et al., 1988).

In addition to the parameter estimates, changes in probabilities, and their corresponding standard errors, a likelihood ratio test can be evaluated as a test of overall model significance. The likelihood ratio test, which has a [X.sup.2] distribution, is derived from the maximum of the log likelihood function of the unrestricted model and the maximum of the log likelihood function of a restricted model, assuming that all the parameters except the intercept are equal to zero. Because no one R2 measure is universally accepted as a goodness-of-fit measure for this class of model, a suggested approach is to base the evaluation of qualitative choice models on several measures (Amemiya, 1981; Laitila, 1993). As a result, a number of alternative R2 measures are presented for the following model, including the Aldrich and Nelson (1984), McFadden (1974), and Veall and Zimmermann (1992) R2's. Computation and a brief discussion of these alternative measures are presented in the endnote. [1]

Apart from indicating the direction of the influence of a variable on the participation choice, parameter estimates of the probit analysis do not have a direct economic interpretation. Thus, marginal changes in probabilities and their standard errors are also reported. The changes in probabilities, as computed in Judge, et al., (1988) are the partial derivatives of the probability function evaluated at each independent variable's sample mean.

5. Data and methods

The primary data used to empirically evaluate the behavioral model of WRP participation were collected via a mail survey of Louisiana's wetland owners. Louisiana was one of eight pilot states selected to participate in the first phase of the WRP, and one of 19 states participating in the second phase of WRP sign-ups. The design and implementation of the mail survey were consistent with Dillman's Total Design Method (TDM) that guides questionnaire design and mailing procedures for improved response rates and response quality (Dillman, 1991).

The survey assessed the respondent's awareness and level of knowledge about the WRP and identified the respondent's sources of information concerning the Wetland Reserve Program. It elicited information about the respondent's land, including total acreage, the acreage of agricultural land, and wetlands owned by the respondent. Information on the location and type of wetlands, the crops grown, labor, and per acre return on the wetlands was also elicited. The survey also requested information about the respondent's participation in the Wetland Reserve Program, including questions pertaining to intentions and effective participation in the WRP, the number of acres of wetlands offered, and the per acre compensation proposed or requested. This section of the survey also provided information on the major reasons for participation or noninvolvement in the WRP.

The environmental attitudes of the respondents were assessed using the theory of reasoned action. Within the framework proposed by Ajzen and Fishbein (1980), environmental attitudes towards participating in the WRP were measured via a series of questions eliciting behavioral beliefs and the subjective evaluations of those beliefs. For the associated subjective norm, normative beliefs, and the corresponding motivations to comply with these beliefs were elicited by the remaining questions in this section. Questions about the respondents' involvement in environmental and agricultural organizations were also included in this section. The survey also gathered socio-economic information, including respondents' age, income, gender, and level of education.

The survey sample included randomly selected Louisiana wetland owners. Of the 767 surveys mailed, a total of 174 completed surveys were available for analysis, a response rate of 23%. Although 133 respondents offered to enroll in the WRP, only 73 respondents eventually enrolled wetlands in the WRP. Wetland owners who ultimately did not participate in the WRP represented 58 percent of the respondents. For purposes of this analysis, wetland owners were classified into two categories based on the offer to enroll wetlands, 1 if wetlands were offered for enrollment, 0 otherwise.

The specific behavioral economic model, including attitude and subjective norm measures, used in the probit estimation is given as:


Definitions of explanatory variables are found in Table 1. LTOTAL, the variable measuring the total acreage owned by the respondent, was assumed to positively affect the participation decision. It was hypothesized that the more land a wetland owner possesses, the greater his ability to adjust his land use options and thus, the greater his willingness to participate in conservation programs such as the WRP. The same reasoning underlies the positive sign hypothesized for the estimate associated with LWET, the variable expressing the acreage of wetlands owned. A negative sign was hypothesized for the revenue variable, REVENUE. Increased or equal returns per acre of wetlands were hypothesized to negatively affect the inclination to participate in the WRP because they lower the opportunity costs of not getting involved in the WRP. Incomplete data prevented empirical evaluation of marginal benefits and opportunity costs of the alternative participation choices.

To capture the effect of the quality and amount of information received by respondents, two information variables were included. Variables representing the source of information and the level of knowledge about the WRP were hypothesized to positively affect participation. A wetland owner who learned about the WRP via the Louisiana Cooperative Extension Service was hypothesized to have a clearer understanding of the nature of the program and the options it offers. KNOWRP, the variable representing the respondent's source of information about the WRP, was therefore hypothesized to have a positive sign. A positive sign was also hypothesized for INFOWRP, the variable expressing the respondent's self-assessed level of information. The greater a wetland owner's level of information about the program, it was hypothesized that he would be more likely to offer to participate in the WRP.

Due to the declining trend observed in the real prices of most of the major crops grown in Louisiana (Zapata & Frank, 1995), an individual owning farmed wetlands was assumed to be more inclined to offer wetlands for enrollment in the WRP than an individual owning other types of wetlands such as bottomland hardwood forests who might anticipate an economic return on the timber. A positive sign was thus hypothesized for the variable WETFARM, specified in this model as a dummy variable, equal to 1 if the respondent owns farmed wetlands, 0 otherwise. No a priori sign hypothesis was formulated for CROP, the dummy variable representing the main crop grown by the respondents. Its inclusion in the model determined the relationship between participation in the WRP and soybeans, the main crop grown by wetland owners in this sample.

A negative sign was hypothesized for ENVORG, the variable representing the respondent's noninvolvement in an environmental organization. The variable ENYORG was specified as a dummy variable equal to 0 if the respondent belong to an environmental organization, 1 otherwise. The wetland owner's involvement in an environment in a wetland owner's household, i.e., the more dependents or potential heirs, the less likely he would be to offer wetlands for enrollment in the WRP due to the additional restrictions that the perpetual easement would impose on his heirs. A negative sign was thus hypothesized for DEPEND, the variable representing the number of people living in a wetland owner's household. It was also hypothesized that wealthier wetland owners were more flexible to explore alternative options for their property. Therefore, INCOME, the variable representing the respondent's income level, was hypothesized to have a positive effect on the decision to offer wetlands for enrollment in the WRP. To reflect charact eristics of this sample for which 65% of the wetland owners had an annual income of at least $55, 000, INCOME was defined as a dummy variable equal to 1 if the respondent's yearly income was at least $55, 000, 0 otherwise.

ATTITUDE, the variable representing respondents' attitudes towards enrolling wetlands in the WRP was computed based on respondents' behavioral beliefs and belief evaluations elicited in the mail survey. Similarly, the assessment of wetland owners' subjective norms was based on their normative beliefs and the associated motivations to comply with those beliefs. Questions eliciting behavioral and normative beliefs were scored using a 1 to 5 unipolar scale. As suggested by Ajzen (1991), questions eliciting belief evaluations and motivations to comply were score based on a --2 to +2 bipolar scale. The attitudinal measure was obtained by summing the behavioral beliefs weighted by their associated evaluations. ATTITUDE, the variable included in this model, is the logged value of the attitudinal measure. The subjective norm variable, SUBNORM, was constructed via a weighted sum of normative beliefs and corresponding motivations to comply. The subjective norm variable was not logged because of the presence of a large number of negative values. It was hypothesized that positive attitudes towards enrolling wetlands in the WRP would significantly reinforce the decision to offer to participate in an environment--preserving program such as the Wetland Reserve Program. Thus, a positive sign was expected for ATTITUDE, the attitudinal variable. Due to the perceived beneficial effect of the WRP on wetlands preservation and restoration, a positive sign was also hypothesized for the measure of subjective norm, SUBNORM.

6. Empirical results and discussion

Descriptive statistics for the variables used in the behavioral model are presented in Table 2. A large dispersion, as measured by the standard error, was observed for the continuous variables included in this model. To control the variability of the continuous variables, i.e., stabilize their variance, a nondecreasing monotonic transformation was performed so that continuous variables were expressed in logarithmic form. The model was then estimated using SHAZAM (White, 1995). Parameter estimates, asymptotic standard errors, changes in probabilities, the percentage of right predictions, and goodness-of-fit measures are reported in Table 3. The value of the calculated likelihood ratio test, 53.80 with 16 degrees of freedom, indicates that the model presented is, overall, significant at a 99% confidence level.

The corresponding [X.sup.2] statistic is 32.0. The percentage of right predictions is 88.11%. The prediction success rates for respondents who offered to participate and those who did not offer to enroll in the WRP were 96.52% and 53.57%, respectively. Alternative R2 measures, also reported in Table 3, range from an Aldrich and Nelson R2 of 0.27 to a Veall and Zimmermann R2 of 0.55.

High intercorrelations between variables, i.e., multicollinearity, is often observed in socio-economic data such as that used in this analysis. Therefore, to test for the presence of multicollinearity, diagnostic tests based on condition indexes were performed. Collinearity tests suggested that, for this sample, the explanatory factors selected to explain wetland owners' decision to offer to participate in the WRP were not correlated. The value of the largest condition index resulting from the principal components analysis performed was 3.54. As suggested by Belsley, Kuh, and Welsch (1980), condition indexes less than 10 indicate only very mild collinearity between the variables considered.

The variables LWET, JNFOWRP, WETFARM, ENVORG, EDUCATE, DEPEND, and INCOME are significantly different from zero at the 90 percent confidence level. The positive relationship between the variables LWET, INFO WRP, WETFARM, and INCOME, suggests that information about the WRP, ownership of farmed wetlands, as well as high levels of income positively affect wetland owners' decisions to offer to participate in the WRP. RESIDE, the variable representing the place of residence of the wetland owner, is positive and significantly different from zero suggesting that, for this sample, wetland owners residing in cities of more than 10,000 inhabitants are more likely to offer participation in the WRP than individuals residing in small towns.

The parameter estimate for ATTITUDE, the variable measuring respondents' attitudes towards offering wetlands for enrollment in the WRP, is positive and significantly different from zero at a confidence level of 90%. Thus, as hypothesized, the more positive a respondent's attitude towards offering wetlands for enrollment in the WRP, the more likely she would be to offer to participate in the WRP. The parameter for the subjective norm associated with the attitude towards offering to participate in the WRP, although positive, is not significantly different from zero. The nonsignificance of the SUBNORM parameter estimate suggests that, despite the perceived environment-improving quality of the WRP, social pressure does not constitute a relevant explanatory factor for the evaluation of wetland owners' offers to participate in the WRP. The decision to offer wetlands for enrollment in the WRP appears for this sample to be a privately based decision.

7. Summary and conclusions

Application of IBM's to environmental management faces many challenges. Due to the voluntary nature of IBM programs, acceptance and thus participation in these programs are not automatic. One reason for the limited use of IBM's is the pervasive mistrust of market forces in dealing with environmental issues. Attitudes towards the use of market forces in environmental policy may therefore play a determining role in the successful implementation of IBM. A better understanding of environmental attitudes, their formation, and translation into participation in a given market-based management program could facilitate a successful implementation of IBM as alternative environment management strategies in agriculture.

Based on the limited understanding of the influence of attitudes in market-based program participation, this research explored the role of environmental attitudes in an incentive-based environmental management program, the wetland reserve program. Constructs derived from the theory of reasoned action included specific measures of wetland owners' attitude towards offering wetlands for enrollment in the WRP and the associated subjective norm. The subjective norm is, in this case, a measure of the social pressure exerted on the decision maker to enroll in the WRP.

As suggested by the Akaike information criteria, R2 measures, and percentage of right predictions, the behavioral model including psychological constructs derived from the theory of reasoned action significantly explained wetland owners' decisions to offer to participate in the WRP. Although the model correctly predicted over 90% of the participation decisions for wetland owners who offered to enroll in the WRP, it accurately predicted about 50% of the choices made by respondents who did not offer to participate. This low prediction success may be due to the limited number of respondents who did not offer acres of wetlands for enrollment in the WRP.

Explanatory factors such as the acreage of wetlands owned, the level of information about the WRP, ownership of farmed wetlands, involvement in environmental organizations, and higher income levels had a significant and positive influence on the decision to offer participation in the WRP. Respondents' education level, and number of people living in the household had an adverse effect on the likelihood to offer wetlands for enrollment in the WRP. Using measures derived from the theory of reasoned action, the model supported the hypothesis that pro-environmental, or positive, attitudes increase the probability to offer to enroll in the WRP. The addition of appropriately designed and measured psychological constructs may conceptually improve the traditional neoclassical economic approach to evaluating choice behavior by allowing the consideration of well established determinants of behavior such as attitudes.

Environmental management in the United States, including the agricultural sector, increasingly relies on the use of voluntary environmental management instruments such as incentive-based mechanisms. This growing interest in the use of incentive-based mechanisms offers a unique opportunity to efficiently correct agriculture-related environmental problems and bring agriculture more in line with mainstream environmental management. The effective implementation and success of future incentive-based environmental management programs rest on getting all the incentives "right," including economic incentives and attitudinal concerns. As a result, policy makers may consider, in the early stages of an IBM, campaigns geared towards increasing environmental awareness or improving predisposition towards a specific environmental management program as an additional instrument to foster the successful implementation of incentive-based programs. However, as information about incentive-based programs such as the WRP become mo re widespread among landowners, policy makers should consider the provision of additional economic incentives to significantly increase participation.

In conclusion, as agriculture moves into its next generation of environmental management, the role of noneconomic factors such as environmental attitudes may play a determining role in individual participation decisions. Our understanding of these attitudinal factors and their relationship to other decision variables will facilitate identification of the "right" incentives upon which to build successful incentive based programs.

(*.) Corresponding author. Tel.: + 1-352-392-1963; fax: + 1-352-392-8988. E-mail address: EJL@ufl.Edu (E.J. Luzar)


(1.) Aldrich and Nelson's [R.sup.2] ([[R.sup.2].sub.AN]) is computed as:

[[R.sup.2].sub.AN] = [lambda]/[lambda] + n

The [[R.sup.2].sub.AN] is based on the likelihood ratio test statistic. However, it does not adjust for the degree of freedom (Aldrich & Nelson, 1984). To achieve an upper limit of one, the [[R.sup.2].sub.VZ] is calculated by multiplying the [[R.sup.2].sub.AN] by a correction factor (Windmeijer, 1995). Veall and Zimmermann's [R.sup.2] ([[R.sup.2].sub.VZ]) is:

[[R.sup.2].sub.vz] = [lambda]/[lambda] + n 2L(0) - n/2L(0)

McFadden's [R.sup.2] ([[R.sup.2].sub.MF]), one of the most commonly used [R.sup.2] measures in qualitative choice models, is given by:

[[R.sup.2].sub.MF] = 1 - [L([beta])/L(0)]

When adjusted for degrees of freedom, McFadden's [R.sup.2] ([R.sup.2].sub.MFA) is written:

[[R.sup.2].sub.MFA] = 1 - [L([beta])/(n - k)/L(0)/(n - l)]

Although the [[R.sup.2].sub.MF] and [[R.sup.2].sub.MFA] lie within the [0, 1] interval, they can not be used as a measure of explained variation because, in addition to the second moments, they involve all the characteristics of the distribution (Laitila, 1993).


Ajzen, I., Attitudes, Personality, and Behavior. Chicago, IL. The Dorsey Press 1988.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes 50, 179-211.

Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behavior. Englewood Cliffs, New York: Prentice Hall

Ajzen, I., & Driver, B. L. (1991). Prediction of leisure participation from behavioral, normative, and control beliefs: an application of the theory of planned behavior. Leisure Sciences 13, 185-191.

Ajzen, I., & Driver, B. L. (1992). Application of the theory of planned behavior to leisure choice. J Leisure Res 24, 207-213.

Ajzen, I., & Madden, T. J. (1986). Prediction of goal directed behavior: attitudes, intentions, and perceived behavioral control. J Exp Social Psych 22, 453-474.

Aldrich, J. H., & Nelson, F. D. (1984). Linear Probability, Logit, and Probit Models Sage University Papers Series 07-045, Sage Publications, Newbury Park, London, New Delhi

Amemiya, T. (1981). Qualitative response models: a survey. J Econ Lit 19(4), 1483-1536.

Anderson, T. L., & Leal, D. R. (1991). Free Market Environmentalism. San Francisco, CA: Westview Press.

Beck, L., & Ajzen, I. (1991). Predicting dishonest actions using the theory of planned behavior. J Res Personality 25(3), 285-301.

Belsely, D., Kuh, E., & Welsch, R. (1980). Regression Diagnosis. New York: Wiley.

Carey, M. B., Heimlich, R. E., & Brazee, R. J. A Permanent Wetland Reserve: Analysis of a New Approach to Wetland Protection. AIB-610 Economic Research Service, USDA, August 1990.

Carlson, A. C., Zilberman, D., & Miranowski, J. A. (Eds.) (1993). Agricultural and Environmental Resource Economics. New York: Oxford University Press.

Daniels, T. L. (1991). The purchase of development rights: preserving agricultural land and open space. J Am Planning Assoc 57(4), 421-431.

Dillman, D. A. (1991). The design and administration of mail surveys. Ann Rev Socio 17, 225-249.

Dudek, D. J., & Palmisano, J. (1988). Emissions trading: why is this thoroughbred hobbled? Columbia J Environ Law 13, 217-256.

Fazio, R. H., Powell, M. C., & Williams, C. J. (1989). The role of attitude accessibility in the attitude-to-behavior process. J Consumer Res 16, 280-288.

Greene, W. H. (1993). Econometric Analysis (second edition). New York, Macmillan Publishing Company.

Heberlein, T. A. (1989). Attitudes and environmental management. J Social Issues 45, 37-57.

Judge, G., Hill, C., Griffiths, W., & Lee, T. (1985). Theory and Practice of Econometrics. New York: Wiley

Kelman, S. (1981) What Prices Incentives? Economists and the Environment. Boston, MA: Auburn House Publishing Co.

Laitila, T. (1993). A pseudo-[R.sup.2] measure for limited and qualitative dependent variable models. J Econ 56(3), 341-356.

Lynne, G. D. (1995). Modifying the neo-classical approach to technology adoption with behavioral science models. J Ag Appl Econ 27(1), 67-80.

Lynne, G. D., & Rolla, L. R. (1988). Improving attitude-behavior prediction models with economic variables: farmers actions towards soil conservation. J Social Psych 128(2), 19-28.

Madden, T. J., Ellen, P. S., & Ajzen, I. (1992). A comparison of the theory of planned behavior and the theory of reasoned action. Personality and Social Psychology Bulletin 18(1), 3-9.

McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior. In P. Zarembka (Ed.) Frontiers of Econometrics (pp. 105-142) Academic Press, New York.

Miller, G. A. (1956). The magical number seven plus or minus two: some limits on our capacity of processing information. Psychologic Rev 63(2), 81-97.

Stewart, R. B. (1988). Controlling environmental risks through economic incentives. Columbia J Environ Law 13, 153-169.

Tietenberg, T. H. (1990). Using economic incentives to maintain our environment challenge. 33(2), 42-46.

Upmeyer, A., & Six, B. (1989). Strategies for exploring attitudes and behavior. In A. Upmeyer (Ed.) Attitudes and Behavioral Decisions. Springer--Verlag, New York, NY.

Van Ryn, M., & Vinokur, A. D. (1990). The Role of Experimentally Manipulated Self-Efficacy in Determining Job-Search Behavior among the Unemployed. Unpublished manuscript, Institute for Social Research, University of Michigan at Ann Harbor.

Veall, M. R., & Zimmermann, K. F. (1992). Pseudo-[R.sup.2]'s in the ordinal probit model. J Mathematical Socio 16 333-342.

Watters, A. E. (1989). Reasoned/Intuitive Action: An Individual Difference Moderator of the Attitude-Behavior Relationship in the 1988 U.S. Presidential Election. Department of Psychology, University of Massachusetts at Amherst. Unpublished master's thesis.

White, K. J. (1993). Shazam User's Reference Manual Version 7.0 Second Printing: McGraw-Hill

Wiadmeijer, F. A. (1995). Goodness-of-fit measures in binary choice models. Econometric Reviews 14(1), 101-116.

Zapata, H. O., & Frank, D. Agricultural Statistics and Prices for Louisiana, 1988-1994. A E A Information Series, No 134, Louisiana State University Agricultural Center, Department of Agricultural Economics and Agribusiness. October 1995.
    Explanatory variables for the behavioral model of WRP participation
Variable    Definition
OFFERWRP  = 1 if the wetland owner offered acres
            for enrollment in WRP; 0 otherwise
LTOTAL    = Total acreage owned (expressed in log
LWET      = Acres of wetlands owned (expressed in
            log form)
REVENUE   = 1 if the respondent's yearly average
            net return per acre increased or
            stayed the same;
            0 otherwise
KNOWRP    = 1 if the respondent learned about the
            WRP from extension service; 0
INFOWRP   = 1 if the respondent has at least good
            knowledge about WRP; 0 otherwise
WETFARM   = 1 if the respondent owns farmed
            wetlands; 0 otherwise
CROP      = 1 if respondent grows soybeans; 0
ENVORG    = 1 if the respondent does not belong to
            an environmental organization; 0
GENDER    = Respondent's gender; 1 if male; 0
EDUCATE   = Respondent's education level; 1 if at
            least attended college; 0 otherwise
RESIDE    = Respondent's residence; 1 if
            respondent resides in a city of more
            than 10,000 people;
            0 otherwise
LAGE      = Respondent's age (expressed in log
DEPEND    = Number of persons living in
            respondent's household
INCOME    = Respondent's income; 1 if respondent's
            annual income is greater or equal to
            0 otherwise
ATTITUDE  = Respondent's attitude towards
            enrolling wetlands in the WRP
SUBNORM   = Respondent's subjective norm for
            enrolling wetlands in the WRP
[epsilon] = Error term
 Descriptive statistics for explanatory variables in the behavioral model
Variable Mean or % Standard Deviation
LTOTAL    6.09      1.5544
LWET      5.06      1.6487
REVENUE   0.48      0.2353
KNOWRP    0.60      0.4893
INFOWRP   0.59      0.4915
WETFARM   0.53      0.5002
CROP      0.50      0.5016
ENVORG    0.33      0.4870
GENDER    0.83      0.3702
EDUCATE   0.72      0.4482
RESIDE    0.29      0.4565
LAGE      3.88      0.6849
DEPEND    2.68      1.2753
INCOME    0.52      0.5006
ATTITUDE  2.05      0.9826
SUBNORM  -1.15     14.8381
N = 174.
     Probit maximum likelihood estimates of the behavioral model [a,b]
Variable Parameter Estimates                Changes in Probabilities
                Value        Asymptotic                        Value
                             Standard Error
LTOTAL        --0.246        0.156
LWET            0.200 [b]    0.122                             0.028
REVENUE         0.356        0.352
KNOWRP          0.415        0.357
INFOWRP         1.380 [b]    0.363                             0.194
WETFARM         0.775 [b]    0.357                             0.109
CROP           -0.049        0.327
ENVORG         -0.877 [b]    0.432                            -0.123
GENDER         -0.933        0.592
EDUCATE        -1.310 [b]    0.508                            -0.184
RESIDE          0.669 [b]    0.402                             0.094
LAGE           -0.792        0.771
DEPEND          0.357 [b]    0.156                            -0.050
INCOME          0.905 [b]    0.456                             0.127
ATTITUDE        0.358 [b]    0.179                             0.050
SUBNORM         0.019        0.012
CONSTANT        4.912        3.479                                --
         Standard Error
LWET     0.017
INFOWRP  0.050
WETFARM  0.049
ENVORG   0.060
EDUCATE  0.071
RESIDE   0.056
DEPEND   0.021
INCOME   0.064

(a.) N = 174; AIC = 60.815; Likelihood Ratio Test = 53.80 with 16 d.f.; Percentage of Right Predictions = 88.11%; [[R.sup.2].sub.AN] = 0.27; [[R.sup.2].sub.MF] = 0.38; [[R.sup2].sub.MFA] = 0.30; [[R.sup.2].sub.VZ] = 0.55.

(b.) Estimates significant at a 90% confidence level (critical t-statistic = 1.64).
COPYRIGHT 1999 JAI Press, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1999 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Luzar, E. Jane; Diagne, Assane
Publication:The Journal of Socio-Economics
Date:May 1, 1999
Previous Article:A socio-economic approach to market transactions.
Next Article:Gender inequality in wage earnings and female self-employment selection.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters