The CE is founded on neoclassical micro-economic consumer theory. It assumes that consumers are highly rational decision makers who seek to optimize their utility by selecting the best option among the available alternatives, subject to constraints such as time and income (Ben-Akiva & Lerman 1985). So, a consumer makes a decision by evaluating the costs and benefits of the given alternatives, and selects among them which yields the highest utility to him/her (benefits-cost). In the CE, the choice is discrete in nature (Hanemann 1984), that is, only one alternative from the given choice set can be selected. Therefore, it is also known as discrete choice experiment (DCE).

The alternative j will be chosen over some option g, if expected utility for individual, i , (Uij) exceeds the expected utility (Uig) for all alternatives. This implies that the probability of selecting an option is likely to increase with utility from the option and probability (P) that individual, I, will choose option j over other options g in a complete choice set R, is given by:

P (jC) = P{ (Uij >Uig, s.t.  g Є R, and j ≠g)} (2)

Usually, in non-market valuation studies using CE model, respondents are given a choice card containing three alternatives including status-quo (current situation) to select. These alternatives in the choice set are outcomes of policies or programs related to the problem being investigated and distinguished by different levels. The levels could be either qualitative or numerical. The effects of the alternatives are described based on the Lancaster’s characteristics theory of value (Lancaster 1966). The theory states that an individual derives utility from the attributes possessed by the goods rather than the goods per se. The choices; therefore, are considered as a function of the attributes that the given alternatives hold. Including the outcomes of policies or programs related to mitigate the damaged caused by invasive plant species, Mikania in particular, a monetary value is included as one of the attributes as the policy cost to the respondent.

Since, the alternatives are evaluated indirectly through their characteristics; it may be possible that consumers may have preference ordering over the attributes included (Lancaster 1966). It is widely accepted that an increase in a level of the preferred attributes or positive externalities such as; biodiversity, forest products and income generating activities, will result, ceteris paribus, in high utility from the goods or services to respondents. On the other hand, the utility from the goods or services decreases with increasing the level of the unwanted attributes or negative externalities such as abundance of invasive plants, pollution and in most cases monetary value. Thus, an individual makes trade-offs between attributes (based on their preference) in the different alternatives in a choice set, implicitly (Alpizar et al. 2001). For example, a change in cost of the alternative can lead a discrete switch from one alternative to another alternative. The CE; therefore, links the Lancaster’s theory with the model for consumer demand for discrete choices (Hanemann 1984).

The SP survey assumes that the utility derives is not restrictive to the attributes presented in alternatives. There are other unobservable components which create inconsistency including socio-economic characteristics of the respondents, attributes not included in the experiment, measurement error and heterogeneity of preferences (Hanemann & Kanninen 1999). For example, for example an array of policy traits were used in the previous CE surveys to estimate the value of non-market forest benefits (e.g. Rolfe et al. 2000, Lehtonen et al. 2003, Xuehong et al. 2007). All these attributes and their interactions cannot be included in a single CE because it increases the size of the experimental design. These unobserved components are unknown to the researcher but respondents consider while making a choice (Haab & McConnell 2002).

Choices made in CEs; therefore, are analysed by using random utility theory (RUT) by including an error term in the utility function to reflect the unobservable factors in the individuals’ utility function (Hensher et al. 2005). The random utility framework links the deterministic model with a statistical model of human behaviour to address those inconsistencies (McFadden 1974, Manski 1977). The indirect utility function is decomposed into two components; deterministic component (V) and a stochastic term (ε). The former component is related to the attributes included in the choice sets.

A random model with the utility function that individual (i), is associated with alternative (j) models the consumer choice behaviour as;

Uijt = Vijt + ijt (3)

where, individual i (i= 1, ……N) obtains utility U from choosing alternative j (j= A, B, C) in each of the choice sets t (t= 1, ….n) presented to them.