Previous PageTable Of Contents

The design, utility and adoption of decision support systems in the New Zealand pastoral industry

Neels Botha and Kris Atkins1

1 AgResearch Ltd, East Street, Private Bag 3123, Hamilton, New Zealand. www.socialsystems.co.nz Email: neels.botha@agresearch.co.nz

Abstract

The purpose of this paper is to: give an overview of the development process of Decision Support Systems (DSSs) and describe the role of end user involvement in that process; assess barriers to their adoption; and to estimate the adoption of DSSs in the New Zealand pastoral sector. There is no empirical data available on the adoption levels of DSSs in New Zealand. The paper briefly reports on an exploratory study about DSSs that were designed for the New Zealand pastoral industry. In this paper we regard DSSs as any tool or system designed to assist end users with decision-making. End users of these tools are farmers or rural change agents that interact with farmers. Data were collected through semi-structured interviews of several New Zealand farm systems scientists and systems modellers. As the research was exploratory in nature we use the views of interviewees on the adoption levels and use of DSSs as an indicator for adoption. We discuss whether DSSs are actually used for decision making or learning, and consider the development process of DSSs and end user involvement in that process. The paper also examines organisational barriers to the DSS development process. This is followed by a discussion of interviewees’ views of the adoption and use of DSSs in New Zealand’s pastoral industry. From the interviews and literature we conclude that: DSSs were used as learning tools rather than decision making tools; end-user involvement in DSS design is important to enhance usability and improve utility; factors related to the design of DSSs, the end users and the industry influence whether DSSs are adopted or rejected. More research is needed to verify the views of interviewees and to establish DSS adoption levels.

Three key learning: (1) DSS development should be a team effort with end user involvement from very early on in the tool development process; (2) all the parties involved in the development of DSSs should align their goals right from the beginning of the tool development process; (3) DSSs that help end users reach their goals, have a clear purpose, and are easy and cheap to use are likely to have high levels of adoption

Key words

Decision support systems; end-user involvement; adoption

Background

DSSs for farm management have been defined in different ways:

  • computer-based technologies that aid decision-making (Jakku et al, undated)
  • interactive computer programs that utilise analytic methods…for developing models to help decision makers formulate alternatives, analyze their impacts, and interpret and select appropriate options for implementation (Adelman, 1992: 2)
  • easy-to-use software on a computer readily accessible to a manager to provide interactive assistance in the manager’s decision process (McCown, 2002b: 19).

In this paper we use the term in a way similar to McCown (2002a) and McMaster (2002), i.e. it is easy-to-use software on a computer readily accessible to a manager to provide interactive assistance in the manager’s decision process.

In New Zealand, as in other parts of the world, there are concerns that there is little adoption and use of these tools in the agricultural sector, e.g. a recent project by Parminter, Botha and Smeaton (2004) that investigated the need for DSSs by dairy farmers in Australia. Some authors are quite explicit about the non-adoption of DSSs. In this regard Jørgensen (2001) says: “It seems to be a general experience that, even though large efforts are spent on development of DSS prototypes, most prototypes never come to a practical application. There may be several reasons for this, including the following:

  • The funding of research projects is focused towards development of prototypes, rather than operational systems.
  • The main purpose of the project was to achieve the knowledge and experience from developing the prototype.
  • The prototype requires input data or skills which can not be expected from the typical user
  • The prototype turned out to perform poorly.”

This concern is shared by McCown (2002a) and by McMaster et al (2002) who say: “Few farmers and ranchers adopt agricultural software such as decision support systems (DSS). While numerous decision aids are available, most are too difficult for producers to use, exclude components necessary for meaningful use on farms and ranches, and usually suffer from poor understanding by scientists of producer needs and how they process information.”

From a theoretical perspective DSSs may be of use when considering how to influence farmers’ behaviour. Firstly farmers can acquire (adopt) them and then use them to help them make decisions. Secondly, they can be useful in a one-on-one context between an extension agent and a farmer. When used on-farm in such a one-on-one situation, the focus of the interaction between them becomes the outputs generated by the DSS rather than the farmer’s or extension agent’s views and opinions. The DSS then makes recommendations or provides alternatives to the farmer rather than the extension agent. This provides a rich context for discussing “what the computer (tool) says” (Webby, personal communication 2005). It also helps the farmer achieve some objectivity about their situation.

Our study was explorative in nature and investigated some of the factors influencing the uptake and use of DSSs from the viewpoint of individuals who were involved with designing and implementing a variety of these tools in the New Zealand pastoral sector. We drew on their rich experience and insights to better understand the development process, end user involvement, barriers to their adoption and to estimate the adoption of DSSs in the New Zealand pastoral sector.

Method

Using semi-structured interviews, we gathered data from seven individuals who have experience with DSSs in pastoral agriculture in New Zealand. The interviewees were key informants and their experience with DSSs ranged from seven to 30 years. They were all employed by a Crown Research Institute (CRI), a science research business owned by the New Zealand Government. The interviews were audio-taped, transcribed, edited and analysed. The findings are outlined in this paper.

Results

Two types of tool

We did not distinguish between different types of DSSs, but interviewees did and intuitively categorised them in different ways. How they classified the DSSs was inconsistent and they used different terms for the categories. For example one interviewee made the difference he saw between DSSs quite clear by saying: “But people don’t understand the difference between decision support and research. They don’t understand that one is to help people make decisions in their daily life whereas a research one is to ask questions and explore concepts”. Research tools are designed and used by researchers while decision support systems are developed for and used by other groups such as farmers. Because interviewees did not use the same categorisations when we interviewed them, we could not distinguish between DSSs for research and decision making in our analysis. We found that DSSs were not only developed for farmers, but also for other users like vets and farm consultants to help them when they interacted with farmers.

We conclude that there are basically two types of tools: research tools and decision support systems.

DSS – for decision support or learning?

In the past many of the DSSs were thought of as easy-to-use software on a computer readily accessible to a manager to provide interactive assistance in the manager’s decision process (McCown, 2002b: 19). However, three of the interviewees talked about DSSs as ‘learning tools’ and not as ‘decision support tools’, for example:

  • “I’d see a lot of these tools are learning tools rather than decision tools”.
  • “As a learning tool it was useful”.
  • “So it was a very useful analytical tool and again a learning exercise that involved groups of farmers as well”.
  • “I think a decision support tool teaches a farmer to anticipate an outcome from a certain set of inputs and after a while they can probably do it without using the model. And I don’t think that is necessarily a major problem”.

How a DSS is used indicates whether it is a ‘learning tool’ or a ‘decision support tool’. Our current position is that when the DSS is used as a ‘learning tool’, it is used simply to better understand and when it is used as a ‘decision support tool’ it is used to provide interactive assistance in a decision process. Hence, when farmers used DSSs as ‘learning tools’ they used it to gain a better understanding and not to make or support decisions. The concept of learning in this regard is important, and building on the conventional epistemological traditions of behaviourism, cognitivism, and constructivism one could further explore the concept ‘learning tool’ and how it is that farmers/landholders learn. Siemens (2004) says that “behaviourism, cognitivism, and constructivism are the three broad learning theories most often utilized in the creation of instructional environments, but that these theories were developed in a time when learning was not impacted through technology”. He is unclear about what this means, but he states that these broad learning theories are not relevant to analyse learning ‘that is impacted through technology’. This view applies to DSSs as learning tools, because our definition of DSSs is technology focussed i.e. “it is easy-to-use software on a computer…” He is developing an alternative learning theory called connectivism, which is “driven by the understanding that decisions are based on rapidly altering foundations. It is the integration of principles explored by chaos, network, and complexity and self-organization theories” (Siemens, 2004). He concludes that “connectivism provides insight into learning skills and tasks needed for learners to flourish in a digital era”. This is a new and developing theory and warrants further exploration and may bring new insights into developing and applying DSS as ‘learning tools’ as opposed to using the traditional epistemological traditions of behaviourism, cognitivism, and constructivism.

Decision Support System development process

There are different models of and approaches to the process of developing DSSs. For example Gachet and Haettenschwiler (2003) who proposed a bipartite approach in which the software engineering part is separated from the knowledge engineering part. Another example is Turban (1995) who described a development process consisting of 11 phases for DSS constructed by end users. Nasirin S, Winter N and Coppock (undated) indicated that end user involvement is important in the development process, especially during implementation. Because of its importance there had been a great amount of studies assessing user involvement in DSS project implementation (Kivijarvi and Zmud, 1993). We asked interviewees about end-user involvement and we constructed a basic development process for DSSs shown in Figure 1, based on the views of the interviewees. End users could be involved in any or all of the development stages and the question for us was whether end users were actually involved and to what extent. This is discussed later in the paper.

Figure 1. The process of developing a Decision Support System.

We did not ask interviewees specifically about the development process of DSSs, but during our analysis it became clear that they discussed important considerations about DSS development. These are shown in Table 1.

Table 1. Important considerations for the development of DSSs.

Consideration

Interviewees’ description

Purpose

“Defining what actually is the model (to) be used for. Depending on the purpose you will have different tools”

Focus on users’ needs

“You need to ask farmers what sort of tools they need to make their own business run better more effectively whatever”
“We tend to worry about perfection whereas what the user wants is thereabouts”

Funder’s needs

“The funder involved – what do they want? Things can sometimes be very, very tight”

Goal achievement

For a modeller it is important to be “very clear what the priorities are” and to “get delivery of what you set out to deliver”

Document properly
for future design
and continuation

“Take the principles and the algorithms out of various bits of code and rewrite then in one language”.
“We were constantly frustrated by the lack of documentation in the program”.

Cost

“Any software development always takes at least 50% longer than the original budget”

Strong leadership and team

“There’s got to be a champion for any model to work; or a group of people who are really keen and really devoted to it”

Table 1 shows that up to four parties could be involved in the design of DSSs for New Zealand pastoral farmers: modellers, scientists, the funder and farmers/landholders. Sometimes the scientists did the modelling and other times they worked with modellers. They are all equally important and between them they have to be clear about the purpose of the DSS. They have to align their needs and be transparent about their goals before the development process begins. If this does not occur it leads to frustration and decreases the likelihood of success i.e. the adoption of the tools by a large group of farmers/landholders. Based on the interviews, we also identified several organisational factors that hinder DSS development (Table 2).

Table 2. Organisational barriers to the development of DSSs.

Organisational barriers

Co-location, financial and administrative

“We have geographical boundaries. We have financial boundaries and administrative boundaries and often that impedes the progress we make in modelling teams”

Poor co-ordination and model integration

“Often the issues we have there is that we have a whole lot of different models (DSTs) which are built on a little bit different principles - and (they therefore) give slightly different answers. So we don’t have an integrated approach to actually having a suite of models which ‘click together’ and contribute to one sort of shared purpose”

Poor leadership

“I think we’ve failed. In my view we’ve had a lack of leadership, of vision and of cohesive action. It’s been splintered and we still have that problem”

Not following best practice

“…setting up what the goals of the whole thing are. And then coming up with specification, formulation, testing (testing is really important), reformulation, retesting. And in that case the goals and formulation need to involve end users in terms of getting your specs. And then the testing must involve end users and then the reformulation is retested. And we just don’t go through those steps”

Table 2 shows that development teams which were not co-located presented problems. It was difficult for development teams to operate properly if team members were not based at the same site. Team members had to travel to meet and it was difficult to get team cohesion. There were also financial and administrative issues that hindered development, but their specifics are unknown. In this case the organisation developed several DSSs over the years. Each of the tools was different in that they used different equations and were based on different principles. These tools could be used together to provide insights and answers, but this did not happen because of poor co-ordination and lack of model integration. Table 2 shows that poor leadership was identified as a problem and that best practice was not followed. These factors hindered the development of DSSs and could have impacted on their adoption too. It seems that a lot of untapped potential went to waste.

We conclude that all the parties involved in the development of DSSs should align their goals right from the beginning of the development process and the way organisations operate can hinder the development of DSTs.

Jørgensen (2001) said that ‘most (DSS) prototypes never come to a practical application’. We tested the applicability of this statement to New Zealand’s pastoral industry by asking interviewees about the adoption of DSSs they were and are involved with. We discuss our findings in the next section.

End user involvement and consultation

Our definition of ‘end user’ was ‘the individuals using the decision support system’. Some of the interviewees were slightly confused because the boundaries of ‘using the tool’ were blurred. For example in some cases interviewees’ involvement started with describing the problem or having the idea (Figure 1) and continued right through the process up to the ‘release’ stage. They were actually designers and end users of the same tool at the same time and hence the question confused them. Also, for one interviewee ‘end user’ meant the funder and he therefore said end users were involved in DSSs “just about every time”. We clarified this confusion by explaining that we were interested in finding out about all end users whether it was themselves or others. In this paper ‘end user’ means the individuals using a decision support system, i.e. farmers/landholders and excludes researchers, scientists or modellers who were involved with developing it.

We could identify only two cases out of nine DSSs where farmers/landholders were actually involved right from the beginning of the development process. In all other cases they were engaged fairly late in the development process or right at the end when they were expected to adopt the tool. In this regard Rogers (1983: 144) said that “the most crucial decision in the entire innovation-development process is the decision to begin diffusing the innovation to potential adopters”. One interviewee made the following comment about scientists and farmers/landholders: “I’m not convinced that that’s ever worked through to the point where scientists actually go and spec it (the DSS) with the end users”. He went on to explain his view by saying that “many scientists feel they know what farmers need”. He argued that scientists sometimes get it right and sometimes wrong but that “the loose intention often is that a model will be used by end users”.

According to the same interviewee, the beginning of a DSS often is when “scientists think it would be a good idea to build a model”. He said that scientists would talk about ‘we’ when they talk about the development process, but that ‘we’ is “…often the group of people who constructed (the model) plus a few ‘tame’ farmers. ‘We’ is a very exclusive group”. He went on to say that “when we do consult end users we often consult top-end, end users”. When interacting with farmers/landholders, scientists tend to concentrate their efforts on a particular group of people. Rogers (1983: 323) called this the ‘homophily principle’. Homophily is the degree to which pairs of individuals who interact are similar in certain attributes. Scientists tend to interact with farmers/landholders who have attributes similar to them and that can become “a very exclusive group” as the interviewee remarked. Not all the interviewees made strong statements like these, and only one of them said that farmers/landholders were engaged “from the word go”.

We conclude that almost all the DSSs in our case were initiated by scientists and when farmers/landholders were consulted they represented a small and selected group.

Adoption of DSSs

Between the interviewees nine different DSSs were talked about. Only three of them were reported as being adopted by farmers/landholders. One of the interviewees said the following:

  • “But there are a few out of - there’s probably 4 maybe 5 (adopters) who would still be users of it. I’ve (scientist) used it extensively every since it was developed for designing farm systems. I still use it in that way. And it’s been used a lot in research helping us look at interactions and what ifs”
  • “So that was very successful because they were part of the development of it right from the word go. And I know people still use it”.

Three of the interviewees are involved with a research tool that is currently being developed into a decision support system for environmental consultants. One Regional Council is planning to use it for regulatory purposes and others have shown some interest. Regional Councils are local government bodies who manage environmental issues and resource management in New Zealand. This particular research tool has been in development for almost 15 years and the interviewees reported on its adoption as follows:

  • “That’s been a slow process. Like the first model came out in 1992 - so what are we up to now 2005 - so it’s been a 15 year growth period? It’s gone from no use at all - it was just a tool that was sitting out there - to slowly become an industry standard now”
  • “It’s getting a lot of profile and at the same time it’s used more and more by fertiliser reps in particular to do nutrient budgets for individual farms. The uptake is not huge, but there is definitely some uptake”
  • “We know something like about 600 to 700 CDs have gone out of it, and (internet) downloads are somewhere roundabout 1,500. We don’t know how many of those are different versions and multiple downloads”.

All the interviewees commented in general about the adoption of DSSs because we asked them about it. None said that DSSs were adopted highly and used regularly, and three made statements about the low levels of adoption of DSSs:

  • “In fact virtually none of them (DSSs) are used widely in agriculture - some of them have been very potent in terms of learning”
  • “The fact is that there aren’t a lot of biophysical models used in agriculture in New Zealand by farmers and consultants”
  • “There’s very few of the tools that we’ve developed over the last 30 years being used much at all”
  • “If we are talking about the utility of these types of models they are amazingly low - very, very low”
  • “You’d be struggling if there’s a potential set of users - if 5% would sustain use”
  • “Models as a whole in agriculture I think had low uptake. I think where we’ve seen high uptake has been in intensive animal industries”.

New Zealand has about 50,000 farmers (Lambert, personal communication 2005). Scientists interacted with a very small number of them and even from that small group very few have been reported to be using the DSS. We conclude that the level of adoption of DSSs in the New Zealand pastoral industry is very low, and continued use is probably restricted to the original farmer/landholder groups. Empirical data is required to confirm the conclusion. Interviewees talked to us about the reasons for the poor adoption of DSSs and this is discussed in the next section.

Reasons for the poor adoption levels

Based on the interviews, we identified two themes: DSS design; and farmers/landholders and their industry. These are shown in tables 3 and 4.

(a) DSS design

Table 3 shows that there are four main aspects about DSS design that hinder their adoption: the DSS does not help the farmers/landholders to achieve their goals; the purpose of the DSS is unclear; the DSS is too difficult to use because it is too complex and/or requires capabilities the farmers/landholders don’t have; and development costs are high. From table 3 it is clear that a DSS must assist farmers/landholders to achieve their goals if it is to be adopted.

Table 3. DSS design and adoption.

Aspect

Interviewee comment

The DSS l is not aligned with farmers/landholders goals so it does not help them to achieve their goals

“We consistently struggle to formulate the models which meet the purposes of end users – and which in terms of being user friendly having the right level of inputs for the user”
“Many scientists feel they know what farmers need which sometimes is different from what farmers want and sometimes it’s the same”
“If people were clear about who the end user was it would be very helpful. Often I don’t think they can understand who the user is”
“We have to be really careful in the formulating, specking, testing to make sure that this is what people want and that it’s going to give them what they want”
“We have to find out – we don’t know, we have to find out – what the requirements of any particular project are which involves a model. And we don’t do that well”
“We continue to target top end technocrats that love scientists but there’s probably only 2 o 3 per cent of farmers or consultants in that range. So we have a real problem with how we target the population that are going to be our end users.”
“Unless they (DSTs) are adding a lot of value they won’t be (used)”

DSS purpose/goal is unclear

“It’s quite obvious why there’s a range of them (DSTs) because they’re created for different reasons. Although they are not clearly documented in my view”
“They often think that something which is suitable for researchers will also be suitable for farmers and that’s patently not going to work”
“Often people don’t understand that the difference between generically decision support tools and research tools”

Complexity and overestimation of user capability

“Now of course we [messed] it up earlier on because we did make the models too complex”
“We over-design – they’re (our DSSs are) too sophisticated”
““As soon as there’s any level of complexity there’s a barrier to use”
“I just don’t think ordinary people can do what experts do just by watching them”
“We also tend to look upon models as entities which run themselves. In fact the people who run them are as important as the model”

High costs

“They are expensive. They are far more expensive than most people imagine”

One of the interviewees talked about reconciling farmers/landholders needs and requirements with DSS design and purpose. He said: “There is a real conflict between what an end user wants and what the owners (of the DSS) are happy delivering”. This problem typically arises when the farmers/landholders are not involved from early on in the development of the DSS. In this particular case scientists developed the DSS over a number of years without engaging farmers/landholders. Hence the interviewee talked about the ‘owners’ of the DSS and the farmers/landholders as if they were two separate entities each with their own needs and goals. It is also important to find out if and how the DSS could help end users to achieve their goals. These goals should be identified and described before the DSS’s specifications are considered (see Figure 1). This is also the appropriate time to engage with funders, because DSSs can be expensive (Table 3) and funders may want to ‘own’ the DSS. Ownership and IP issues are increasingly becoming an issue (Table 4). Early engagement provides sufficient time for modellers, scientists, funders and end users to align their goals. If these goals are comparable it increases the chance of success.

End user surveys may be a useful instrument to get a better understanding of farmers/landholders’ needs. If they are not used in the right manner or at the right stage of decision tool development they will create problems. One interviewee talked about that when he said: “The other problem we’ve got is - an end user survey is very good at telling you what their current software is not doing. But is usually pretty crappy at telling you what your future development should look like”. To help align developers’ goals with farmers/landholders’ needs, end user surveys should be designed to help identify farmers/landholders’ needs rather than doing trouble-shooting. Software development can then be done with those needs in mind. End user surveys or personal involvement will help clarify the purpose of the tool early on in the tool development process. Table 3 shows the importance of purpose clarity. Adoption levels will be disappointing when scientists and/or modellers design DSSs and then try to find a purpose or market for them. If farmers/landholders’ needs are similar to the purpose of the tool, one of the main barriers to adoption will have been overcome. For example if the farmers/landholders’ need is “we want to be able to assess optimum pasture cover” and the DSS’s purpose is “to assess optimum pasture cover” farmers/landholders’ need is similar to the purpose of the tool.

Some DSSs are costly to develop. This may cause funding problems and create long time lags which may impact on their adoption because they arrive late to the market. Other DSSs may be costly to run, for example if they need lots of difficult to acquire data inputs. If using DSSs do not add value they will probably be poorly adopted.

(b) Farmers/landholders and their industry

DSS complexity and farmers/landholders capability are factors equally relevant to the design of DSSs (Tables 3 and 4). DSSs that are difficult to use because they are complex or complicated are likely to be disregarded and rejected because they take a lot of energy and effort to use (Tables 3 and 4). They can be difficult to use because of either design or lack of user capability or both. Designers should understand farmers/landholders’ capabilities and design DSS accordingly.

Table 4. Adoption of DSSs and factors related to farmers/landholders and their industry.

Adopters/users

Complexity, energy, learning, value

“As soon as there’s any level of complexity there’s a barrier to use because if people only use them occasionally there’s a really rapid learning curve you have to go through every time. And that adds time and takes mental energy and people struggle with that”
“Farmers think they (DSSs) are too hard. Farmers don’t recognise the value to using them and they are pressed for time. They spend 10 or 12 hours a day working outside they are not computer literate”

Industry

Resource allocation: time, energy

“Our (pastoral) industry is low cost because we produce mainly commodity products from our livestock industries which means that we have we try to minimise inputs and maximise outputs. And minimising inputs includes time and mental energy - people just get stretched a long way. So to hope that people would use decision support systems in their everyday farming is difficult”

Funding and Intellectual Property

“There were some complexities around the way it was funded and who owned the model all that sort of thing. There are IP issues”

Size of the market

“The New Zealand market is too small to even worry about trying to promote it”

Perception of DSSs

“They are often relatively difficult to use, or perceived to be difficult to use (by the industry)”

Moreover, if DSSs are used irregularly it could take a lot of re-learning each time they are used (Table 4). With use over time farmers/landholders will get more confident in using DSSs but most would probably not bother to learn how to use them if they are complicated. The programming and mathematical equations that underpin and drive DSSs are sometimes quite complicated and complex, but their user interface should be very user friendly. If the farmers/landholders’ capabilities in terms of using DSSs are understood, it is easier for the designer to create a user interface with an acceptable level of simplicity to use. DSSs can also be difficult to use because they require detailed and sophisticated input data to run. The adoption level of DSSs with this requirement will probably be low, because they tend to be very cumbersome to use.

Table 4 shows that the nature of the pastoral industry has been flagged as a reason for the poor adoption levels of DSSs. New Zealand pastoral farmers are pressed for time and generally spend long hours working outside. Time and energy are part of their input costs and need to be accounted for in DSS development and design. It has become common practice for funders to retain the Intellectual Property associated with the product they fund. This has been raised as an issue that can impact on the adoption of DSSs. One interviewee said that DSSs generally are viewed as difficult to use by the industry and that the size of the market in New Zealand is too small to even promote them. We conclude that DSSs that help farmers/landholders reach their goals, have a clear purpose, and are easy and cheap to run are likely to have higher levels of adoption.

Conclusions

Most of the DSSs were actually used for learning, rather than decision support. Almost all the DSSs were initiated by scientists and when farmers/landholders were consulted they were from small and selected groups of pastoral farmers. It is important that all the parties involved in the development of DSSs align their goals right from the beginning of the tool development process. In addition the way in which organisations operate can hinder the development of DSSs. Finally, the estimated level of adoption of DSSs in the New Zealand pastoral industry is very low, and continued use is probably restricted to the original farmers/landholders groups. DSSs that help farmers/landholders reach their goals, have a clear purpose, and are easy and cheap to use are likely to have higher levels of adoption. Three key learnings are: DSS development should be a team effort with end user involvement from very early on in the tool development process; all the parties involved in the development of DSSs should align their goals right from the beginning of the tool development process; DSSs that help end users reach their goals, have a clear purpose, and are easy and cheap to use are likely to have high levels of adoption

References

Adelman L (1992). Evaluating decision support and expert systems. New York: Wiley.

Gachet A and Haettenschwiler P (2003). Developing Intelligent Decision Support Systems: A Bipartite Approach. In: Palade V, Howlett RJ. Lakhmi J (eds) Knowledge-Based Intelligent Information and Engineering Systems. Proceedings of the 7th KES Conference, Springer-Verlag (LNAI 2774), Berlin Heidelberg, Germany: pp. 87-93.

Jakku E, Thorburn P and Gambley C (undated). Sociological Concepts for Understanding Agricultural Decision Support Systems. Tropical Landscapes Program, CSIRO Sustainable Ecosystems. http://216.239.57.104/search?q=cache:N3GOFJBpcFgJ:www.tasa.org.au/conferencepapers04/docs/RURAL/JAKKU_THORBURN_GAMBLEY.pdf+Decision+Support+Systems+agriculture&hl=en

Jørgensen E (2001). Available on Website: http://www.jbs.agrsci.dk/~ejo/models.html

Kivijarvi, H and Zmud RW (1993). DSS implementation activities, problem domain characteristics and DSS success. European Journal of Information Systems, 2(3), 159-168.

McCown R (2002a). Changing systems for supporting farmers’ decisions: problems, paradigms, and prospects, Agricultural Systems 74 (1): 179-220.

McCown R (2002b). Locating agricultural decision support systems in the troubled past and socio-technical complexity of ‘models for management’, AgriculturalSystems 74 (1): 11-25

McMaster GS, Ascough Ii JC, Shaffer MJ, Byrne PF, Haley SD, Nielsen DC, Andales AA, Dunn GH, Weltz MA, Ahuja LR (2002). Parameterizing Gpfarm: An Agricultural Decision Support System For Integrating Science, Economics, Resource Use, And Environmental Impacts. International Society Of Ecological Modeling Annual Meetings. In: A.E. Rizzoli And A.J. Jakeman (EDS.), Integrated Assessment And Decision Support Proceedings Of The 1st Biennial Meeting Of The Iemss. June 24-27, 2002, Lugano, Switzerland. Vol. 1, Pg 72-77.

Nasirin S, Winter N and Coppock P (undated). Factors influencing user involvement in DSS project implementation: Some lessons from the United Kingdom health sector. Paper available on: http://is.lse.ac.uk/asp/aspecis/20050008.pdf

Parminter TG, Botha CAJ and Smeaton D, (2004). Needs Analysis for Dairy Industry Grazing Management Research and Extension: Final Report. AgResearch client report prepared for Dairy Australia

Rogers EM (1983). The diffusion of innovations. 3rd ed. The Free Press: New York.

Siemens G (2004). Connectivism: A Learning Theory for the Digital Age. Paper available on http://www.elearnspace.org/Articles/connectivism.htm

Turban E (1995). Decision support and expert systems: Management support systems. (Third ed.) NY.: Macmillan Publishing Company.

Previous PageTop Of Page