4. Materials and methods

The study is confined to Madurai semi-urban areas. Madurai district (region) is existing 15 semi-urban areas (see Table 1). The sample units were selected adopting the stratified random sampling method. A total of 150 schedule 10 household respondents from each semi-urban area. The design of the survey followed recommendations from the NOAA panel on contingent valuation (see [52, 53]) and consist of two sections. Questions in the survey's first section asked about respondent's socio economic conditions in the household's survey section two questioned respondents about their willingness to pay. The hypothetical improved condition, and how each consumer would pay for the improved waste management services in Madurai (Figure 1).

The contingent valuation employed a single-bounded dichotomous choice format by open-ended questions in the WTP section. The survey was conducted March–April 2012. The survey was given to 150 randomly selected in Madurai semiurban areas data covered socioeconomic characteristics of the household, including gender, age, marital status, education, household income, family size, employment and WTP for environmental improvement and better solid waste management. Table 2 describes the variables.

## 4.1 Willingness to pay for improved waste management services in the study area

The Logit regression model had been used for studying about the probability of occurrence of an event by fitting a logit function. It is a generalized linear model used for binomial regression. The logit model was adopted since the Ordinary Least Square (OLS) producer was not appropriate particularly when the dependent variable is dichotomous. The problem with the OLS estimate however is the non- fulfillment of O = (Yi/X) since E (Yi/X) in the liner probability model measures the conditional probability of the event Y occurring given X1 and must necessarily lie between 0 and 1 [54]. Like many other forms of regression analysis,

#### Figure 1. Madurai semi-urban areas.


#### Table 2. Description of the variables.

it makes use of several predictor variables that might be either numerical or categorical. This Study had applied the logit regression of willingness to pay for improved environmental quality, to determine the willingness of the respondents to bear the costs of improving the environmental quality in the study area. The Logit Model had been used to analyze the respondents' willingness to pay for an improved waste management service and the factors influencing their willingness to pay.

Household Willingness to Pay for Improved Solid Waste Management Services: Using Contingent… DOI: http://dx.doi.org/10.5772/intechopen.83598

## 4.2 Willingness to pay for improved waste management services

To obtain the willingness to pay by the households for an improvement in their solid waste management, the responses of the households for willingness to pay was regressed on the socio economic characteristics. The coefficient estimates obtained for the WTP of the respondents (sex, age, education, family size, monthly size, monthly income, present cleaning status and maximum amount), the logit regression Model [55] was specified as

$$\mathbf{Y} = \frac{1}{\mathbf{1} + \exp^{-(\|\mathbf{b\_0} + \mathbf{b\_1}\mathbf{x}\|)}}$$

where

Y = Response of households', sex, age, education, family size, monthly size, monthly income, present cleaning status and maximum amount of willing to pay for respondents to the willingness to pay question which was either

'1' if Yes or '0' if No.

β0 = is the intercept which is constant

β1 = is the coefficient of the price that the household are willing to pay

for waste management services.

X = is a set of independent variable
