Managing Agricultural Risk: Examining Information Sources Preferred by Limited Resource Farmers

Introduction

The most useful asset a farmer can have to help with the management of risk is good information. There are many sources of information available to farmers. However, the most appropriate place to look for information depends on the type of risk with which the farmer is concerned (Nelson, 1997; Harwood, Heifner, Coble, Perry, & Somwaru, 1999). Among the most common risk factors that farmers face are weather, crop and livestock diseases, pests, adoption of new technologies, fluctuating prices, and government programs and policies.

In an effort to help farmers mitigate risk, the U.S. Department of Agriculture and other organizations like the national crop insurance service have offered a wide range of different risk management tools such as crop insurance, futures, options, basis pool and forward contracts to farmers. However, the adoption of these and other agricultural risk management tools by farmers, in general, and limited resource farmers, in particular, has been slow.

Previous research (Coble, Knight, Patrick &, Baquet, 1999; FSC, 2000; Tiller, 2000; Roe, 1998) suggests that the slow adoption of agricultural risk management tools is related to lack of knowledge and understanding about them. For instance, a survey by the Federation of Southern Cooperatives of black farmers in Alabama, Georgia, Mississippi, and Texas found that less than 44% of the producers had received risk management training (FSC, 2000).

The main reason given for low participation in such training programs was that many agencies, including land-grant universities, do not give adequate technical assistance to farmers on such tools as crop insurance (FSC, 2000). In a survey of producers growing major field crops in Indiana, Mississippi, Nebraska, and Texas, Coble, Knight, Patrick, and Baquet (1999) found that less than 34% of the producers had attended any risk management education or other training programs.

Among limited resource farmers, however, the reasons for the slow adoption of risk management tools go beyond the lack of knowledge (Dismukes, Harwood, & Bentley, 1997). This group of farmers produces products (fruits and vegetables or livestock) that are generally not covered by insurance products. Furthermore, information gathering is costly, and small and limited resource farmers may not be interested in laying out those costs if the benefits of them are comparatively small. Large farmers, on the other hand, can justify information-gathering costs because it is a public good, so economies of scale enter into the calculation.

It is important that farmers know about the various risk management tools available to them so that risk acceptance is a result of choice rather than the lack of awareness of the availability of the alternative risk management tools/sources.

There are also numerous inconsistencies with respect to the factors that influence farmers' attitudes toward specific information sources (Gloy, Akridge, & Whipker, 2000). For instance, age and experience were important characteristics in determining information preferences in studies by Ford and Babb (1989), Schnitkey, Batte, Jones, and Botomogno (1992), Gloy, Akridge, and Whipker, (2000). But they were unimportant in studies conducted by Pompelli, Morfaw, English, Bowling, Bullen, and Tegegne (1997), Foltz, Lanclos, Guenthner, Makus, and Sanchez (1996), and Ortmann, Patrick, Musser, and Doster (1993).

Measures of farm size were related to both attitudes toward, and the use of, information sources in studies by Ford and Babb (1989), Ortmann et al. (1993), and Foltz et al. (1996). Schnitkey et al. (1992) and Ortmann et al. (1993) found that the farm's use and attitudes toward different information sources varied by enterprise type. Other factors that have been found to influence attitudes toward information sources are experience with the information source (Pompelli et al., 1997), experience with technology such as computers (Schnitkey et al., 1992; Ortmann et al., 1993), and farmers' skills in different functional management areas (Ortmann et al., 1993).

In view of these inconsistencies, the study reported here sought to understand small and limited resource farmers' perceptions of the usefulness of information received from a variety of information sources and identify factors that explain the variation in farmers' attitudes toward these sources. The insights gained from the study should help improve the efficiency with which Extension and agricultural educators develop targeted outreach activities that will ensure that farmers receive adequate information in a format they can appreciate and understand.

Theoretical Approach

The traditional approach of modeling behavior under risk is through the use of the expected utility approach. Utility theory provides a means of monitoring how people perceive risk and of measuring subjective values by taking advantage of an individual's perception of risk (von Neuman & Morgenstern, 1944; Luce & Raiffa, 1957; Myerson, 1979). The application of utility-theory methods does not require that decision makers have any explicit idea of probability or make explicit mathematical calculations (Rapoport, 1966:30). They need only make decisions based on their subjective perception of probabilities. It is assumed by this method that a decision maker's preferences are complete, transitive, and continuous (von Neuman & Morgenstern, 1944; Luce & Raiffa, 1957; Myerson, 1979).

Completeness means that a decision maker can compare any alternatives under consideration. Transitivity means that a decision maker who prefers A to B and B to C will also prefer A to C. Continuity means that a decision maker's utility increases continuously such that if A is preferred to C, any option B that is ranked between A and C can be represented by a randomized combination of A and C. Provided that a decision maker's preferences meet these requirements, researchers can use utility-theory methods to monitor preferences and to model decision making.

Economists taking an explicitly deductive approach tend to rely for its validity more on the theory's axiomatic foundations than on empirical demonstrations (Perry 1998; Paris & Caputo 1993). When economists do test utility theory, it is often in experiments (Kahneman, Knetsch, & Thaler, 1990; Cubitt & Starmer, 1998; Bosch-Domenech & Silvestre, 1999; Butler, 2000). Some experimental economists have focused on violations of utility-theory assumptions. Many of these limitations were detailed in a seminal article by Kahneman and Tversky (1979) in which they noted common violations of utility theory such as unequal weighting of losses versus gains, overweighting of certain outcomes over probabilistic ones, and failure to consider common features of prospects relevant to the calculation of their value.

Other researchers have built upon this foundation (Tversky & Kahneman, 1992; Cubitt & Starmer, 1998; Butler, 2000; Morrison, 2000). In contrast to critical experimental studies, non-experimental studies by agricultural economists (Bar-Shira, 1992; Smith & Mandac, 1995; Elamin & Rogers, 1992; Zuhair, Taylor, & Kramer, 1992) tend to support the fit between utility theory and people's actual behavior. For instance, Bar-Shira (1992) found that, when a feasible solution to a land allocation problem for farmers exists, risk aversion coefficients can be assessed and people behave in accordance with utility-theory predictions.

Despite various limitations, utility theory appears valid when its assumptions can be met, and violations of assumptions can often be overcome with modifications to utility functions (Kahneman & Tversky, 1979; Tversky & Kahneman, 1992; Butler, 2000). As noted by Morrison (2000), despite the limitations of utility theory, "a clearly superior model has not yet been identified." In view of the above we use utility theory to examine the factors that are correlated with how farmers rate the usefulness of the information sources available to them.

Data

The first step of data collection involved identifying the different information sources about agricultural risk management tools that are available to farmers, in general, and limited resource farmers, in particular, and the criteria used to evaluate information sources. This was achieved by contacting Extension agents using snowball sampling (Malhotra, Shaw, & Crisp, 1996) where each agent was asked to recommend others who could help further. An extensive search of the Internet and libraries also led to the discovery of different products and sources of information available to farmers. A brief summary was written about each information source and then categorized using evaluation criteria into:

• Risk management experts,

• Printed materials,

• Computer-based,

• Marketing associations,

• Radio/TV, and

• Advice/face to face contacts.

The evaluation criteria were:

• Cost,

• Readability,

• Relevance,

• Balance view,

• Depth of content,

• Range of content,

• Presentation,

• Ease of access,

• Ease of use,

• Timeliness,

• Accuracy, and

• Feedback.

The qualitative data from the Extension agents were used extensively in the design of the survey questionnaire that used scale type questions to identify and assess farmers' evaluation and ratings of selected information sources. The survey instrument was examined for content and face validity by four research associates in the Small Farms Research Center (SFRC) at Alabama A&M University. Then, the questionnaire was administered using two methods: mail and outreach.

For the mail survey, the questionnaire was sent out to a random sample of 228 small farmers in north Alabama counties. Of these, 36 usable questionnaires were returned, and 68 questionnaires were returned as undeliverable. Thus, the mailing of the survey resulted in an adjusted usable response rate of 21%. At the two outreach conferences (organized by the SFRC and the Alabama Cooperative Extension System), participants were asked to fill out the questionnaire. Overall, a total of 83 questionnaires were collected at both conferences.

Survey responses from the mail and the two outreach conferences were statistically compared on key variables relating to demographic and farmer characteristics, and no statistical differences (p > .05) were found either between the respondents to the mailing or between the outreach conferences. The analysis combines mail and conference survey responses to yield a sample of 117 small farmers. Although combining the responses from the two different data samples makes the overall sample non-random and vulnerable to other problems such as selection bias, the initial statistical tests showed no significant differences between the samples.

Information Source Evaluation

The survey asked farmers to select the most useful source of information and rank these sources using a Likert type scale (ranging from 1 for not useful to 5 for very useful). The responses from this question are used to construct the dependent variable (USFLNS), which measures limited resource farmers' ranking of the sources of information consulted. Overall the most useful information sources (Table 1) were printed materials (magazines, newsletters, and fact sheets) followed by face-to-face advice by other farmers and risk management experts (training courses/seminars, brokers/advisers) by order of preference. Others were computer (Internet-based education modules, e-mail), books, risk management associations (marketing clubs), and radio/television programs. The findings in Table 1 are consistent with previous findings (Suvedi, Campo, & Lapinski, 1999; Roe, 1998) that media, consultants, Agfacts, and, to a lesser extent, field days are the main information sources used by farmers.

Table 1.
Number and Percent of "Most Useful" Information Sources

Information Source	Count	Percentage
Printed Materials [magazines, newsletters, fact sheets]	41	35%
Face-to-face advise by other farmers	37	32%
Risk management experts [training course/seminar, broker/adviser]	31	26%
Computer [internet-based education modules, e-mails]	28	24%
Books [detailed reading on own]	23	20%
Risk management associations/marketing clubs	18	15%
Radio/television programs	16	14%

It appears from the results in Table 1 that one of the better ways to help limited resource farmers manage agricultural risk is their access to printed materials like periodic newsletters, fact sheets, and other practical material.

A snowball effect will also ensure that the more farmers are reached through initial efforts, that more other farmers will get the information, because communication with their peers seems to be one of the best sources of information at their disposal.

Econometric Approach

The empirical model examines how the ranking of the usefulness of risk management information sources are correlated with limited resource farmers' characteristics. The questionnaire asked farmers to "indicate how useful the sources of information are in helping [them] to make decisions." (The reliability coefficient, Cronbach' alpha, is estimated at .89, meaning that the data are seemingly measuring the same latent construct). In addition, the questionnaire captured personal data, including age, educational level, and ethnicity, as well as data about the farm: farm tenure, ownership structure, farm sales, and type of production (Table 2).

Because the information source rankings are qualitative and discrete in nature, an ordered probit model was estimated. The ordered probit regression produces the maximum likelihood estimates of coefficients that predict a farmer's ranking of the information sources. The underlying variable, the actual rank expressed by the farmer, is continuous and unobservable; only the values chosen as most closely representing farmers' actual ranking is observed.

Table 2.
Variable Definitions

Variable Name	Variable Description
	Ranking of Level of Usefulness of Risk Management Information
Dependent Variable
USEFULNESS	=0 if the information is not useful at all
	=1 if the information is somewhat useful
	=2 if not sure whether the information is useful
	=3 if the information is useful
	=4 if the information is very useful
Independent Variable
OWN	=1 if farmer owns the farm; 0 otherwise
FULL-TIME	=1 if full-time farmer; 0 otherwise
MARKETING PLAN	=1 if farmer has a marketing plan; 0 otherwise
INSURANCE	=1 if farmer has crop insurance; 0 otherwise
PRODUCTION	=1 if farmer produces row crops
	=2 if farm produces livestock
	=3 if farmer produces fruits and vegetables
	=4 if farmer produces products other than the above
AGE	=1 if age is 39 years or below
	=2 if age is between 40-49 years
	=3 if age is between 50-59 years
	=4 if age is 60 years or above
ETHNICITY	=1 if white
	=2 if black
	=3 if Hispanic
	=4 if American Indian
	=5 if Other
SALES	=1 if farm sales are less than $5,000
	=2 if farm sales are between $5,000 and $9,999
	=3 if farm sales are between $10,000 and $19,999
	=4 if farm sales are above $20,000
EDUCATION	=1 if farmer completed high school or less
	=2 if farmer attended college
	=3 if farmer attended graduate school

The estimated model is specified as:

USEFULNESS = Constant + Own + Full-time + Ethnicity + Education + Insurance + Marketing plan + age + Sales + Production. (See Table 2 for variable definitions)

Similar studies have found that the selected factors are usually correlated with how farmers perceive or rate information sources that they receive and also on whether farmers adopt new techniques or technologies (Jones, Battle, & Schnitkey, 1989; Isengildina & Hudson, 2001; Amponsah, 1995; Batte, Jones, & Schnitkey, 1990). The equation is estimated using the ordered probit procedure in LIMDEP (Greene, 2000). While considering data limitations, the results can assist Extension and agricultural educators in identifying information sources most preferred and useful to small and limited resource farmers.

Results

The dependent variable (USEFULNESS) is constructed to take into consideration the indicated sources of information for which each farmer has provided an evaluation. The different ratings for each source are combined into one value that gives a general idea of what farmers in general think about the sources of information that they consult. The results are presented in Table 3, including the log likelihood coefficient, the Chi-square, model's prediction success, and estimated marginal effects.

The measures of goodness of fit indicate that the model fits the data fairly well. The log-likelihood, which measures the significance of probit function, was significant at p< 0.01, suggesting that a relationship exists between information source ratings and the suggested independent variables. The model correctly predicted 71% (83 out of 117) of the responses. Similarly, the estimated model assumes that there are threshold values (µ1, µ2 and µ3) above which the rating of information sources goes to the next higher level. The estimated µ values for these threshold levels are statistically significant at p< 0.01 and positive, hence validating the use of the ordered probit model.

To further understand how the dependent variable (level of usefulness) is related to the independent variables, marginal effects are estimated and reported in the last column of Table 3. These effects are evaluated by assuming that a given respondent has the mean score for every independent variable; in other words, the respondent is average in every way. This technique enables to isolate the effect of a change in one variable given that all the others remain constant.

The variables that are significant at 5% level or higher are OWN, AGE, and MARKETING PLAN, implying that these variables are the strongest correlates with how farmers rate the usefulness of the risk management information they receive. To the contrary, variables related to ethnicity, production, full-time, and sales are less instrumental in influencing the way farmers rate/perceive the different sources of information.

As we further discuss the results, it is important to note that statistical problems such as multicollinearity are a common problem in economics and econometric modeling. Even though no evidence of such problems was noted in this analysis, given the qualitative nature of most of the explanatory variables, it is possible that the significance and possibly the signs of some parameter estimates may have been affected by such problems, albeit insignificantly.

Table 3.
USEFULNESS Model: Summary of Results

Variable	Coefficient	SE	P-Value	Marginal Effects
	USEFULNESS=4
CONSTANT	0.829	0.619	0.181	0.1583
OWN	0.886**	0.304	0.004	0.1692
FULL-TIME	0.228	0.228	0.319	0.0434
INSURANCE	0.348	0.264	0.187	0.0665
MARKETING-PLAN	0.573*	0.227	0.012	0.1094
PRODUCTION	-0.084	0.151	0.579	-0.0160
AGE	-0.256*	0.103	0.013	-0.0489
ETHNICITY	0.073	0.136	0.593	0.0139
SALES	0.036	0.113	0.752	0.0068
EDUCATION	0.206	0.152	0.177	0.0393
µ1	0.571**	0.179	0.001
µ2	1.657**	0.243	0.000
µ3	2.906**	0.303	0.000
Log likelihood function	-150.648
Restricted log likelihood	-168.410
Chi-squared	35.5240
Model prediction	0.71
* &** Denote significant at the 5 and 1% levels

The OWN variable, which refers to farm ownership, has the strongest explanatory power in usefulness perception among farmers who responded to the survey. This result suggests that farmers who own land/farm strongly feel that the various sources of information that they consult are useful. This result implies that a farmer who owns the land would be more committed and more interested to acquire risk management information compared to a farmer who does not own the land.

The estimated marginal effect for the OWN variable shows that farmers who own land are 16.92% more likely than farmers who rent to rate the risk management information they receive as very useful. One plausible explanation could be that ownership results in a better and long-term commitment that can be related to the perception of the usefulness of the various sources of information.

The second highest explanatory power comes from the MARKETING PLAN variable, which is also significant at p< 0.05. The estimated marginal effect suggests that farmers who have marketing plans are 10.94% more likely to rate the risk management information they receive to be very useful than farmers who do not have a marketing plan.

It is plausible that having a marketing plan results in knowing the specific needs in the farm operation, giving rise to the applicability and usefulness of the various information sources available. Over the past few years many top-notch producers have gone out of business. One contributing factor has been the inability to sell at profitable prices. Having a marketing plan can improve the odds of selling farm products at prices that ensure the survival of the farm business.

The result for AGE shows that age exerts downward pressure (negative influence) on USEFULNESS, which means that as people age, they are not as satisfied about the information they receive as are younger people. AGE is also the only variable that has a negative relationship to USEFULNESS. The estimated marginal effects suggest that moving from one age category to another will lower farmers' rating of the risk information sources by 0.0489.

In support of similar findings, Schnitkey et al. (1992) argued that age is related to farming experience, and that farmers with more experience should have less demand for external information. Following the experience argument, older farmers may find the cost of information gathering less desirable than younger farmers. Thus, it is expected that age will be negatively related to the usefulness of information received, particularly from media sources. To the contrary, Kool, Meulenberg, & Broens (1997) found that input suppliers were more likely to have established relationships with older producers. If farmers value the information provided by these relationships, age should be positively related to the usefulness of information received from personal information sources.

As one anonymous reviewer noted, the expected sign for the education variable is ambiguous. However, we hypothesized higher levels of education to be positively related to the usefulness of information received from all information sources. Higher levels of education should be consistent with increased ability to process information and/or individuals that seek out information. However, the estimated results suggest that education was less instrumental in determining attitudes toward any of the information sources. This result is consistent with the findings of Foltz et al. (1996) and Pompelli et al. (1997).

In general, it appears that within the population of small and limited resource farmers, education is a relatively less important indicator of preferences toward information sources. Similarly, other factors, including PRODUCTION, SALES, ETHNICITY, INSURANCE and FULL-TIME were less instrumental in influencing the way farmers rate/perceive the different sources of information.

Conclusions

The study employed survey data collected among small and limited resource farmers in north Alabama to determine the factors that influence farmers' perception of usefulness of sources information in managing agricultural risk. Each information source provided benefits to some farmers. Some sources, such as printed materials and face-to-face advice by other farmers, had broad, appreciative audiences with few distinguishing characteristics. Others, such as risk management experts and Internet-based information, are less well received in general, but are valued by certain groups of farmers.

To examine the effect of small and limited resource farmers' characteristics on information source rating, the study used a probit model. The data used to examine information preferences came from county survey of farmers. These farmers are among the small family farming operations in Alabama. There is little consistency with respect to the factors that influence the perceived usefulness of the sources.

Factors that appeared to be positively related to the perceived usefulness of information sources include farm ownership, farmers' age, and having a marketing plan. Factors such as ethnicity, education, sales, type of production, radio, television, marketing clubs, and having insurance were infrequently, if ever, related to the usefulness of information received from the sources. Although factors such as ethnicity, type of production, and education were generally unimportant in explaining information preferences in this sample, they could be important in the general farm population. Likewise, if a farmer's characteristic differs dramatically from that of sample farmers explored here, it would be unwise to assume that these factors were entirely unimportant.

There are several managerial implications of the study reported here. When selecting methods to deliver information to farmers, Extension and other agricultural educators must consider the type of information to be delivered, the capability of the information source for delivering the information, and their target clientele's preferences for receiving information from various sources.

When selecting information sources, Extension and other agricultural educators should recognize that there are differences with respect to the factors that influence attitudes toward each source. In other words, each source should be evaluated on a case-by-case basis. The factors that are important in explaining attitudes toward one information source may be very different than the factors that explain attitudes toward another information source. Finally, the fundamental limitations of this study pertain to survey data. These include, but are not limited to, coverage errors, non response and distortions of measurement errors, selection bias, omitted variable problems, and possible econometric problems previously highlighted in this article.

References

Amponsah, W. A. (1995). Computer adoption and use of information services by North Carolina commercial farmers. Journal of Agricultural and Applied Economics, 37(2), 565-576.

Bar-Shira, Z. (1992). Nonparametric test of the expected utility hypothesis. American Journal of Agricultural Economics, 74, 523-34.

Batte, M. T., Jones, E., & Schnitkey, G. D. (1990) Farm information use: An analysis of production and weather information for Midwestern cash grain farmers. Journal of Production Agriculture, 3(1), 76-83.

Bosch-Domenech, A., & Silvestre, J. (1999). Does risk aversion or attraction depend on income? An experiment. Economics Letters, 65, 265-73.

Butler, D. J. (2000). Do non-expected utility choice patterns spring from hazy preferences? An experimental study of choice errors. Journal of Economic Behavior and Organization, 41, 277-97.

Coble, K., Knight, O., Patrick, G. F., & Baquet, A. E. (1999). Crop producer risk management survey: A preliminary summary of selected data: A report from the understanding farmer risk management decision making and educational needs research project. Information Report 99-001, Mississippi State University.

Cubitt, R. P., & Starmer, C. (1998). Dynamic choice and the common ratio effect: An experimental investigation. Economic Journal, 108, 1362-81.

Dismukes, R., Harwood, J. L., & Bentley, S. E. (1997). Characteristics and risk management needs of limited resources and socially disadvantaged farmers. Agricultural Information Bulletin No 733, 104-109.

Elamin, E. M., & Rogers, L. F. (1992). Estimation and use of risk aversion coefficient for traditional dry land agriculture in western Sudan. Agricultural Economics, 7, 155-66.

Federation of Southern Cooperative (FSC). (2000). Risk management survey of African-American farmers: Preliminary findings. Federation of Southern Cooperatives/Land Assistance Fund (FSC/LAF). A Survey Funded by the Risk Management Agency of the United States Department of Agriculture.

Foltz, J., Lanclos, K., Guenthner, J., Makus, L., & Sanchez, W. (1996). The market for information and consultants in Idaho agriculture. Agribusiness, 12, 569 ­581.

Ford, S. A., & Babb, E. M. (1989). Farmer sources and uses of information. Agribusiness, 5, 465­ 476.

Gloy, B. A., Akridge, J. T., & Whipker, L. D. (2000). Sources of information for commercial farms: Usefulness of media and personal sources. International Food and Agribusiness Management Review, 3, 245-260.

Greene, W. (2000). LIMDEP, Version 7.0, Econometric software, New York: Plainview.

Harwood, J., Heifner, R., Coble, K., Perry, J., & Somwaru, A. (1999). Managing risk in farming: Concepts, research and analysis. Market and Trade Economics Division, ERS, USDA. Agricultural Economics Report No. 774.

Isengildina, O., & Hudson, D. M. (2001). Factors affecting hedging decisions using evidence from the cotton industry. Paper presented at the NCR-134 conference on applied commodity price analysis, forecasting, and market risk management, April 23-24, St-Louis, Missouri.

Jones, E., Battle, M. T., & Schnitkey, G. D. (1989). The impact of economic and socioeconomic factors on the demand for information: A case study of Ohio commercial farmers. Agribusiness, 5, 557-571.

Kahneman D., & Tversky, A. (1979) Prospect Theory: An analysis of decision under risk, Econometrica,47, 263 ­ 91.

Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1990). Experimental tests of the endowment effect and the Coase theorem. Journal of Political Economy, 98, 1325-47.

Kool, M., Meulenberg, M. T. G., & Broens, D. F. (1997). Extensiveness of farmers' buying process. Agribusiness, 13, 310 ­318.

Luce, R. D., & Raiffa, H. (1957). Games and decisions: Introduction and critical survey. New York: John Wiley.

Malhotra, N. K., Hall, H., Shaw, M. & Crisp, M. (1996). Marketing research: An applied orientation, Prentice Hall, Sydney, Australia.

Morrison, G. C. (2000). The endowment effect and expected utility. Scottish Journal of Political Economy, 47, 183-197.

Myerson, R. B. (1979). An axiomatic derivation of subjective probability, utility, and evaluation functions. Theory and Decision, 11, 339-52.

Nelson, G. A. (1997). Teaching agricultural producers to consider risk in decision-making. Faculty Paper 97-17. College Station, Texas: Texas A&M University

Ortmann, G. F., Patrick, G. F., Musser, W. N., & Doster, D. H. (1993). Use of private consultants and other sources of information by large cornbelt farmers. Agribusiness, 9, 391­ 402.

Paris, Q., & Caputo. M. R. (1993). Keeping the dream of rigorous hypothesis testing alive. American Journal of Agricultural Economics, 75, 25-41.

Perry, G. (1998). On training Ph.D.s in economics: What can economics programs learn from those in agricultural economics? American Journal of Agricultural Economics, 80, 608-15.

Pompelli, G., Morfaw, C., English, B. C., Bowling, R. G., Bullen, G. S., & Tegegne, F. (1997). Farm operators' preferences for soil conservation service information: results from three Tennessee watersheds. Journal of Production Agriculture, 10, 472­ 476.

Rapoport, A. (1966). Two-person game theory: The essential ideas. Ann Arbor: University of Michigan Press.

Roe, S. (1998). Wool producers use and knowledge of price risk management instruments. Honors Dissertation, Curtin University of Technology, Muresk Institute of Agriculture, Perth.

Schnitkey, G., Batte, M., Jones, E., & Botomogno, J. (1992). Information preferences of Ohio commercial farmers: Implications for extension. American Journal of Agricultural Economics, 74, 486 ­ 496.

Smith, J., & Mandac, A. (1995). Subjective versus objective yield distributions as measures of production risk. American Journal of Agricultural Economics, 77, 152-62.

Suvedi M., Campo, S., Lapinski, M. K. (1999). Michigan farmers' information seeking behaviors and perspectives on the delivery of information. Journal of Applied Communications, 83, 33-50.

Tiller, B. M. (2000). Price risk management tools and the Western Australian grain producer. Honors Dissertation, Curtin University of Technology, Muresk Institute of Agriculture, Perth.

Tversry, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty, 5, 297-323.

Von Neumann, J., & Morgenstern, O. (1944). Theory of games and economic behavior. New York: John Wiley.

Zuhair, S., Daniel, M. M., Taylor, B., & Kramer, R. A. (1992). Choice of utility function form: Its effect on classification of risk preferences and the prediction of farmer decisions. Agricultural Economics, 6, 333-44.