2. Literature Review

Organic food consumption has increased because of various food scares and consumers’ awareness of food safety (Azzurra et al., Reference Azzurra, Massimiliano and Angela2019). The food industry is tainted by its use of artificial toxic chemicals that affect human health adversely and cause unexpected deaths (Ashraf et al., Reference Ashraf, Joarder and Ratan2019; Rahman et al., Reference Rahman, Sultan, Rahman and Rashid2015). A significant reason for consumers’ organic food consumption is the issue of health, which influences their purchasing behavior (Carboni et al., Reference Carboni, Vassallo, Conforti and D’Amicis2000). Furthermore, organic food consumption is vital, as it increases sustainable diets and ensures food sustainability (Mørk et al., Reference Mørk, Bech-Larsen, Grunert and Tsalis2017). The European Union follows a strict policy for the organic sector, taking into consideration organic food and the eco-friendly behavior of consumers (Azzurra et al., Reference Azzurra, Massimiliano and Angela2019). Such practices increase social well-being and foster economic resilience (Schader et al., Reference Schader, Stolze and Niggli2014).

Consumers’ purchasing behavior goes through a psychological process which includes recognizing needs; searching for ways to meet these needs by collecting and understanding information; making and implementing plans; making decisions on product purchase; purchasing the product; and giving feedback post-purchase (Basha et al., Reference Basha, Mason, Shamsudin, Hussain and Salem2015). In the context of organically farmed food, consumers’ purchasing behavior has been investigated by several researchers. For example, Rana and Paul (Reference Rana and Paul2017) report that consumers’ attitude has a clear impact on their purchase intentions, including the cognitive approach, which indicates thinking, and the affective approach, which helps to form the feeling component (Aertsens et al., Reference Aertsens, Verbeke, Mondelaers and Van Huylenbroeck2009). In organic food purchases, the roles of affective and cognition attitudes are compensatory (Dean, Raats, and Shepherd, Reference Dean, Raats and Shepherd2008), while Michaelidou and Hassan (Reference Michaelidou and Hassan2008) claim that health consciousness, food safety concerns, and ethical self-identity influence consumers’ attitude toward organic food. Moreover, the qualities of organically farmed food, such as higher nutrition and better taste, and the avoidance of chemicals and pesticides in the production process, also influence consumers’ purchase intentions. For instance, in the United States, consumer demand for natural and organic foods has substantially increased due to renewed and increasing interest in nutrition, public health, animal welfare, and the impact of traditional agriculture on the environment. However, such health and environmental concerns have an insignificant effect on the consumption of organic food because of the limited knowledge about such food in emerging markets (Le-Anh and Tam Nguyen, Reference Le-Anh and Nguyen-To2020; Yilmaz and Ilter, Reference Yilmaz and Ilter2017).

Consumer’s knowledge about organically farmed food indicates their understanding of the level of the exploration and recognition of product characteristics (Muhummad et al., Reference Muhummad, Fathelrahman and Ullah2016). Therefore, product knowledge plays a vital role in forming attitudes toward organic food (Gan et al., Reference Gan, Chang, Tran and Cohen2016). Consumers who possess adequate information about such food and its production techniques are more prepared to choose organic products (Gracia and de Magistris, Reference Gracia and de Magistris2007; Saleki et al., Reference Saleki, Seyedsaleki and Rahimi2012). Such findings lead us to assume that low, medium, and high levels of product knowledge will influence consumers’ preference for conventional, safe, and organically farmed fish, respectively. Liu et al. (Reference Liu, Yan and Zhou2017) highlight that despite perceiving organic products to be beneficial for both health and the environment, consumers’ preferences for such products are subjective, as they have inadequate knowledge about the production standards of eco-labels. For instance, in Vietnam, consumers of organic food face problems due to their lack of related knowledge; many cannot even differentiate between organic and safe foods (Q & Me, 2018; Takayama, 2017). Such lack of knowledge creates a lack of trust (Vega-Zamora et al., Reference Vega-Zamora, Torres-Ruiz and Parras-Rosa2019) and confusion because of the multiplication of labels, leading consumers to perceive that organic products imply only high prices (Díaz et al., Reference Díaz, Pleite, Paz and García2012). Furthermore, limited knowledge about organic production techniques negatively affects organic purchases, particularly in developing countries (Gracia and Magistris, Reference Gracia and de Magistris2007; Yin et al., Reference Yin, Wu, Du and Chen2010).

In addition to knowledge, consumers’ demographic variables also affect their preferences for organic food (Davies et al., Reference Davies, Titterington and Cochrane1995; Thompson, Reference Thompson1998). Females are more prepared to pay for organic apples, while males are more likely to pay for organic beef (Illichmann and Abdulai, Reference Illichmann and Abdulai2013). Moreover, consumers from highly educated groups (Annunziata et al., Reference Annunziata, Agovino and Mariani2019), especially women with more extensive education; those from higher-income households; and older consumers are all prepared to pay more for the welfare of farmed fish. They believe that eco-friendly fish are fresher and have a better life (Solgaard and Yang, Reference Solgaard and Yang2011). However, households with children are not willing to pay a price premium for organic products (Sriwaranun et al., Reference Sriwaranun, Gan, Lee and Cohen2015).

A number of previous studies have been conducted on consumers’ WTP for organic food. In one, Japanese consumers were shown to be willing to pay 8%–22% more for certified organic vegetables (Sakagami et al., Reference Sakagami, Sato and Ueta2006). In addition, Rodríguez et al. (Reference Rodríguez, Lacaze, Lupín and Canavari2009) found that consumers from Argentina were willing to pay for organic food, within a broad range of 6%–300%, while another study found that Dominican consumers were prepared to pay 17.5% more for organic products (Boys et al., Reference Boys, Willis and Carpio2014). Moreover, Canadian consumers were shown to be prepared to pay 10% extra for organic products (Vladicka and Cunningham, Reference Vladicka and Cunningham2002). As in the case of organic food in general, organic aquaculture is also presently focusing on consumers’ food consciousness. Organic aquaculture production considers animal welfare and public health issues in the production stages. Relevant information is provided by organic labels, increasing consumers’ awareness, preferences, and WTP (Mauracher et al., Reference Mauracher, Tempesta and Vecchiato2013), thus creating market demand and increasing producers’ revenues (Ankamah-Yeboah et al., Reference Ankamah-Yeboah, Jacobsen, Olsen, Nielsen and Nielsen2019). In Italy, organically farmed marine fish have great potential, as consumers are willing to pay 2.25 €/kg extra over average premium prices (Defrancesco, Reference Defrancesco2003). Disegna et al. (Reference Disegna, Mauracher, Procidano and Trevisan2009) report that in the case of organic trout, on average potential consumers are willing to pay 2.55 €/kg more, while Norwegian consumers are prepared to pay extra for organic and freedom-food salmon compared with the conventional version (Olesen et al., Reference Olesen, Alfnes, Røra and Kolstad2010).

In 2017, total global aquaculture production increased by 49% compared with 2016, with the total production of organic aquaculture almost 0.62 million metric tons, of which 86% was produced in Asia. As the single most valuable globally traded aquaculture product, farmed shrimp plays a vital role, being mainly produced in southeast Asia and Latin America. Recently, the world has focused on organic shrimp production due to consumers’ food awareness. Organic shrimp production began in Ecuador in May 2000 and was followed by other countries such as Indonesia, Peru, and Vietnam (Bergleiter, Reference Bergleiter2002). The world’s main shrimp producers are Thailand and China, whose economies rely mainly on the shrimp industry (Research and Marketing, 2019). As shrimp is an essential exportable product in Bangladesh, the country has practiced organic shrimp cultivation on a limited scale in order to meet consumer demand and increase its economic contribution. Black tiger shrimp (Penaeus monodon) production comprises 71.5% of the country’s total shrimp farming and represents more than 90% of the total export earnings from farmed shrimp (FAO, 2019). Shrimp farmers in the Bangladeshi economy mostly rely on wild shrimp stock because of the limited capacity of shrimp production (Alauddin and Hamid, Reference Alauddin and Hamid1999). In 2006, the production of wild marine shrimp was 3,200 tons, while the production of farmed Bagda and Golda shrimp was 38,000 and 12,000 tons, respectively (DoF, 2006).

3. Organic Agriculture and Fish in Bangladesh

Organic farming in Bangladesh was introduced in the late 1970s with the support of various NGOs such as PROSHIKA and UBINIG in order to provide seasonal vegetables in a suitable, equitable, and productive way following the principles of biodiversity (Iqbal, Reference Iqbal2015; Proshika, 2004). Presently in Bangladesh, consumers’ preferences are shifting from traditional to organic food because of its unique characteristics, such as safety, concern for the environment, nutrition, and sensory attributes (Mukul et al., Reference Mukul, Afrin and Hassan2013). Consumers like organic food because of its better taste, health benefits, and its attractiveness as a fashionable product. Organic products have great potential as exports as well as in local markets in Bangladesh. Both young and older people consume organic products, with more men preferring organic food; as the fish buyers in the family, they are more willing to pay a price premium for organic foods than women (Ahmed and Rahman, Reference Ahmed and Rahman2015; Iqbal, Reference Iqbal2015). Furthermore, consumers’ level of education, income, consciousness, and household size, together with the price, and, for example, the breed of fish, in relation to food safety and quality assurance, affect the WTP for organic food (Sarma and Raha, Reference Sarma and Raha2016). Prince and Krairit (Reference Prince and Krairit2017) report that consumers with children, older people, and men who have the regular habit of buying organic meat are encouraged to buy such products in Bangladesh. They add that organic food attributes such as health benefits, verbal recommendations, purchase convenience, and availability significantly affect consumers’ intention to buy organic meat. The lack of awareness and knowledge about such products and their price premium significantly and negatively influence people’s intention to buy organic foods (Ahmed and Rahman, Reference Ahmed and Rahman2015; Iqbal, Reference Iqbal2015). Many consumers claim that organic products are in insufficient supply in Bangladeshi local markets, and that they are limited to particular shops, and continue to be poorly certified. Therefore, consumers have low trust in organic food producers and sales personnel (Ahmed and Rahman, Reference Ahmed and Rahman2015; Sumi and Kabir, Reference Sumi and Kabir2018).

In Bangladesh, the shrimp sector contributes greatly to income from the foreign exchange market, consequently expanding food production, and improving the livelihoods and income of farming households and associated groups (Ahmed, Reference Ahmed2013; Ahmed et al., Reference Ahmed, Thompson and Glaser2018; Islam, Reference Islam2008). This sector is the second-largest export industry, generating US$380 million annually, which is 5.6% of the total value of exports (DoF, 2006), with 1.2 million people directly, and 4.8 million households indirectly linked to the sector (USAID, 2006). Although the advent and expansion of the many types of aquaculture of fish and shrimp are not sustainable (Hossain and Hasan, Reference Hossain and Hasan2017), and outbreaks of disease are the main hindrance for shrimp culture (Alam et al., Reference Alam, Pokrant, Yakupitiyage and Phillips2007), shrimp farming in Bangladesh enjoys high demand in the global market, especially in Europe. However, in recent years, international demand for Bangladeshi shrimp has decreased from US$417 million in 2017–2018 to US$365 million in 2018–2019 (Rahman, Reference Rahman2019). Such a decline is the result of failing to meet the international demand for world-class certification of products and the competition generated from the introduction of the farming of “Litopenaeus vannamei” (white leg shrimp) (Rahman, Reference Rahman2019). Moreover, the shrimp industry in Bangladesh has faced substantial economic losses because of infections from viral diseases such as the White Spot Syndrome Virus (Alam et al., Reference Alam, Pokrant, Yakupitiyage and Phillips2007; Mazid and Banu, Reference Mazid and Banu2002). In addition, the industry has faced low yields, lack of adequate technology, price fluctuations in international markets, bans imposed by the European Union, and lack of government stimulus (Alam et al., Reference Alam, Pokrant, Yakupitiyage and Phillips2007; Chowdhury et al., Reference Chowdhury, Shivakoti and Salequzzaman2006; Paul and Vogl, Reference Paul and Vogl2011).

International markets demand high standards and quality shrimp products, a fact that should focus producers’ attention on strictly following standardized policy and shrimp production regulations. As a result, alternative organic shrimp production was introduced in Bangladesh. In this regard, the Swiss Import Promotion Program regulated by the Swiss government started the Organic Shrimp Project in 2005 to help small and medium enterprises in developing and emerging economies through consulting, training, marketing support, and the facilitation of access to trade fairs (Paul and Vogl, Reference Paul and Vogl2012). In Bangladesh, the agro-climatic conditions, biophysical resources, abundant ponds, and available low-lying rice fields with sufficient wild post larvae in coastal areas are very favorable for farming large freshwater shrimp (Ahmed et al., Reference Ahmed, Demaine and Muir2008; Ahmed et al., Reference Ahmed, Allison and Muir2010). Therefore, organic shrimp culture in Bangladesh has great potential due to the increased demand from consumers for organic products in the international market (Ahmed et al., Reference Ahmed, Thompson and Glaser2018; Aschemann-Witzel and Zielke, Reference Aschemann-Witzel and Zielke2017; Ruangpan, Reference Ruangpan2007). However, unstable monthly incomes and different sociopolitical forces hinder the purchase of such food products in Bangladesh (Ashraf et al., Reference Ashraf, Joarder and Ratan2019).

4. Data and Methods

4.1. Participants

The study intends to develop an integrated picture of consumers’ preferences for farmed shrimp. To obtain their opinions, three divisions (counties) were purposively selected, Chattogram, Dhaka, and Rangpur. Geographically, Chattogram and Rangpur are located in the southern and northern parts of the country, respectively, whereas Dhaka is located in the center (Figure 1). In terms of shrimp production, Chattogram is the highest-level area, Dhaka the middle-level, and Rangpur the lowest-level (Appendix A). To recruit respondents, consumers’ living standards and fish consumption frequency were also considered. Dhaka and Chattogram are in the second- and third-lowest poverty line positions, while Rangpur is at the lowest level (Appendix A).

Figure 1. Black shading indicates the study area.

In addition, in the cities of Chattogram and Dhaka have the highest per capita fish consumption amongst cities in the country, while Rangpur has the lowest consumption (Needham and Funge-Smith, Reference Needham and Funge-Smith2014). These varying criteria helped to choose the representative sample and were suitable for our attempt to explore the growing safety consciousness and diverse fish choices in an emerging market such as Bangladesh. The reason for selecting shrimp was that this species is cultured in inland freshwater, coastal brackish water, and marine saline water (on a limited scale). For the data collection, an experimental design method was followed, in which households were selected randomly by conducting a direct interview method. A questionnaire was sent to the respondents, who were asked to complete it, as well as take part in a face to face interview. Only people older than 21 were asked to participate in the survey, since those under 21 are generally not responsible for fish purchases in the family. Each survey took on average 15 minutes.

Before producing the final survey, the questionnaire was translated from English into the native language Bengali with the help of professional language editors. A pretest survey with 48 respondents was then conducted in two cities (Dhaka, with 25 participants, and Chattogram, with 23) to ensure that they understood the content of all the questions, and that no semantic problems or linguistic complexities existed. As no significant difficulties were found, it was decided to keep the original format for the final version. The Dean Committee, University of Chittagong, Bangladesh, approved the ethical standard of the experiment research content. The survey was conducted from August 21 to October 17, 2019. A total of 704 households took part in the survey; however, 44 responses were excluded due to their unsuitability (incomplete data). Therefore, 660 samples were finally used in the study. In a within-subject design study, the experiment provided a dataset of n = 660 × 9 = 5,940.

4.2. Questionnaire and Measurements

The first section of the questionnaire consisted of consumers’ product knowledge regarding farmed fish, and their attitude toward it. The second section concerned choice of fish based on ranking, focusing on fish attributes. Nine alternatives were presented in a table, and respondents were asked to rank these according to their preferences. The final section was a demographic survey concerning fish-eating behavior. To ascertain the consumers’ product knowledge, six questions were posed based on the revolutionary theory of product knowledge taxonomy expounded by Russo and Johnson (Reference Russo and Johnson1980), using seven-point Likert scale items, ranging from 1 (strongly disagree) to 7 (strongly agree) (Appendix B). To construct general attitudes toward farmed shrimp, five seven-point bipolar scales, from bad to good, negative to positive, unfavorable to favorable, dull to exciting, and terrible to high, were employed to describe respondents’ feelings about farmed shrimp in general. The study used exploratory factor analysis (EFA) to find the best number of dimensions and their common associations based on responses to particular issues (consumers’ knowledge and their attitude toward farmed fish), in order to form a pattern matrix (Hair et al., Reference Hair, Black, Babin and Anderson2010). The EFA considered three statements related to the construct of product knowledge, with the remainder concerning attitude (Table 1). The mean values of the extracted factors from each variable were then measured for use as independent variables. Respondents who gave scores of 5 or below were regarded as agreeing less or having lower perceived values (low knowledge or negative attitude). In contrast, those who gave scores 5 were deemed to be neutral (medium knowledge or neutral attitude). Finally, scores of 6 and above indicated that participants were in strong agreement or had high perceived values (high knowledge or positive attitude).

Table 1. Outcome of explorative factor analysis

Note: Extraction method: Principal Component Analysis.

In the second stage, three fish attributes (farm type, price, and farming system) and three levels (low, average, and high) in terms of the three attributes were considered in order to design product alternatives. A total of 3³ (27) hypothetical products could be generated by combining the attributes and levels. In the study, three types of production field, marine, coastal, and inland, were considered. To select the fish attributes and to account for their optimal levels, a focus group discussion was held. For practical analysis, an orthogonal fractional factorial design was used, which considered only the main effects of the attributes. This decision was based on the evidence that the main effects explained the variance in the choice model at a level of between 70% and 90% (Dawes and Corrigan, Reference Dawes and Corrigan1974). The program employed (SPSS, version 26) helped reduce the minimum number of choice sets from 27 to 9. Following the study of Balcombe et al. (Reference Balcombe, Fraser and Falco2010), the participants were instructed to think about the choice scenarios as if they were real. Textual and visual information (see Figure 2) were given to the participants for them to have an idea of the fish type and aqua farming systems used in the choice experiment.

Figure 2. Relevant information (textual and visual) about the fish type and the farmed fish production methods.

Sources of images: Apart from the image related to coastal aquaculture (own source), all other images were taken from freepik: https://www.freepik.com/.

In the third section, the respondents’ demographics and socioeconomic characteristics in relation to fish consumption frequency and the markets where they shopped for fresh fish were explored. Finally, the question of whether they could contribute to saving natural fish stocks from depletion through their personal choice of fish was approached in a binary setting. To rank the choice of product alternatives, nine fish products were presented in a table (see Figure 3). To analyze the ranked choice, the study employed an ROL model. As the most preferred choice was independent of the remaining choices ranked in the model, the independence of irrelevant alternatives property was applicable (Beggs et al., Reference Beggs, Cardell and Hausman1981). The ROL model was employed to estimate consumers’ fish attributes and the interactions between them, the socioeconomic variables and their WTP in the selected areas of Chattogram, Dhaka, and Rangpur. Interaction terms between the fish attributes and socioeconomic variables were also included in this section because those comprising mobility and self-care were the most salient (Nicolet et al., Reference Nicolet, Hoogenboom, Pratchett and Willis2018).

Figure 3. Farmed Shrimp Choices with Focus on Attributes.

In the experiment, the choice of farmed shrimp was observed only amongst people over 21 years old and responsible for buying fish for the family. However, such a sampling selection could be the cause of sampling bias. Theoretically, the ROL model (equation 5) used in the study could fit, even though the same choices are not made by everyone, and consumers’ socioeconomic variables change over time. This unobserved characteristic of the observed sample and nonrandom sample bias can be adjusted with endogenous covariates in an ordinary sample-selection model (Schwiebert, Reference Schwiebert2015). In this study, the selection model fits because the preferences of the respondents can differ completely from those who are not responsible for buying fish for the family and who are below 21. Therefore, the differences are for unseen reasons, and it is not clear if these unobservable characteristics may lead to biased results. To control this bias, whether the sample selection mattered was checked by using an extended ordered probit regression (EOPR) in STATA (the Statistical Software, version 16), which outlined whether there was a change in the unobserved characteristics of the respondents’ fish choice. In this case, “unobserved” refers to unused factors (gender, the log of income, education, consumption frequency, market type, and awareness of farmed fish) in the estimated model, which may contribute to the analysis of the consumers’ choice concerning “choice of farmed fish to save wild fish stock,” which is known as an unobserved preference component or responsibility level. These unused factors were observed in the EOPR model as explanatory variables to analyze choice behavior. The endogeneity effect of the probability of the purchase of fish on fish choice was also checked with the EOPR model. The first and second panels in the results show the choice equation and the preference to save wild fish (selection) equation. The results demonstrate an insignificant positive correlation (ρ) between the residuals of overall farmed shrimp choice and the preference to save wild fish stock, at 0.074; S.E. = 0.242; z-score = 0.31; p = 0.758, indicating that selection was not an issue (StataCorp, 2019). Moreover, the almost zero positive correlation between farmed shrimp choice and the preference to buy farmed fish in order to save wild fish stock indicates that farmed shrimp choice is not endogenous with the personal preference for fish to save wild stock, and is not significant (ρ = 0.005; S.E. = 0.074; z-score = −0.07; p = 0.942). Such findings indicate that those who are more likely to choose farmed shrimp are more likely to prefer (albeit insignificantly) to save wild fish stock through their personal fish choice (Drukker, Reference Drukker2017).

5. Econometric Model

The study followed the conjoint valuation model, which indicates consumers’ decision to maximize their utility. In this case, the respondents were asked to rank the items by order of their preferences, with the most preferred alternative indicating high utility. Assuming that U _ij is the rank given to alternative j by respondent i, if there are J alternatives, then U _ij may consider integer values from 1 through J, where 1 is the “best” and J is the “worst” in terms of ranking (Allison and Christakis, Reference Allison and Christakis1994). A model for such data can be generated from a random utility model, the same model that is employed to explain the standard multinomial logit model (Allison and Christakis, Reference Allison and Christakis1994). With J possible alternatives, the utility given by alternative j for individual i is defined in a linear function as

(1) $${U_{ij}} = \;{V_{ij}} + \;{E_{ij\;}},$$

where each U _ij is the sum of a systematic component V _ij and a random component E _ij. Each E _ij is independent and equally distributed with an extreme value or double exponential distribution. Each V _ij can be assumed to be a numerical quantity indicating the degree to which respondent i prefers alternative j over other alternatives that reflect utility. To estimate the utility weights and identical prices, the log-likelihood function was used, which is the ROL (Hausman and Ruud, Reference Hausman and Ruud1987). The utility index of the mth alternative was designated by

(2) $${V_m} = \;x_m'\;\beta + \;{u_{m,}}\;\;\;\;m = 1, \ldots .,\;$$

where x _m is a K-vector of alternative attributes; ß is a K-vector of utility weights; and u _m is a randomly distributed error term (with an extreme value distribution). The logit probability that alternative j is preferred to alternatives 1,…, j-1 is

(3) $${F_j}[{x_{1, \ldots \ldots ,}}\,{x_i^\prime};\beta ] = \raise0.7ex{\exp (x_j^i|\beta )}\Big/ \lower0.7ex{[\sum\nolimits_{i = 1}^j {\exp \,(x_i^\prime\,\beta )} ]},j \gt 1.$$

Denoting the index of the alternative ranked mth by r _m , the probability of observing the rank ordering $r \equiv \left( {\;{r_{1,\; \ldots ..,\;\;}}\;{r_M}\;} \right)$ is

(4) $$\Pr (r,x;\beta ) = \prod _{m = 2}^M{F_m}[{x_{rM, \ldots \ldots ,}}{x_{rM - m + 1}};\beta ],$$

where x = [x _m ; m=1, …, M], which has the suitable form of the product of M − 1 usual logit likelihood functions. Beggs, Cardell and Hausman (Reference Beggs, Cardell and Hausman1981) developed and applied this model in a field experiment on choice, which was the result of the independence from irrelevant alternatives (IIA) property of the logit specification. In selecting alternatives, mutual exclusiveness was followed, explaining that all other alternatives were rejected by selecting one alternative (Train, Reference Train2009). If a sample considers N observations of x, the attributes of the M alternatives, and the rank ordering of alternatives is r, then the log-likelihood index for the sample of n observations is

(5) $\eqalign{ L\left( {\beta } \right) = \mathop \sum \limits_{n = 1}^N {\rm{log}}\left[ {{\rm{Pr}}\left( {{r_{n,}}{x_n};\beta } \right)} \right] \cr \qquad\hskip-3pt= \mathop \sum \limits_{n = 1}^N \mathop \sum \limits_{m = 1}^{M - 1} {\rm{log}}\left[ {{F_{M - m + 1}}\left[ {{x_{{r_M}, \ldots \ldots ,}}{x_{{r_m}}};\beta } \right]} \right]. \cr} $

This log-likelihood function is the sum of the ordinary logit log-likelihood functions that can be used to estimate the ROL (Hausman and Rudd, Reference Hausman and Ruud1987), meaning that the parameters estimated in the model are applicable to the probability of the observed ranking. Positive parameters indicate that the predictor variable is likely to increase the probability of ranking the associated product attribute. Alternatively, negative parameters show that the explanatory value tends to decrease the ranking probability (Zhang and Khachatryan, Reference Zhang and Khachatryan2019). The parameter estimation commands of “cmrologit” of the STATA programming fit a choice model for rank-ordered alternatives that assumes IIA is true and allows tied ranks.

In addition, “individual choices are correlated with individual-specific explanatory variables, which take the same value across the choice categories” (Franses and Paap, Reference Franses and Paap2001). Marginal values based on estimated parameters reflect the WTP based on product attributes. According to Train (Reference Train2009), the estimate can be calculated as the negative ratio of the coefficient of an attribute variable ( ${\beta _{attribute}}$ ) to the price coefficient ( ${\beta _{price}}$ ); the formula is as follows:

(6) $$WT{P_{attribute}} = \; - {{{\beta _{attribute}}} \over {{\beta _{price}}}}$$