International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
A hybrid choice model of dairy novel production strategies
Raffaele Zanoli, Simona Naspetti

Last modified: 28 March 2017

Abstract


This paper deals with consumer acceptance of novel dairy production strategies. An hybrid choice model or integrated choice and latent variable model (ICLV), merging the choice model with the structural equation approach (SEM) for latent variables representing antecedents of choice, is used for the analysis. Consumer acceptance is modelled by an extended Theory of Planned Behaviour framework, where Choice is modelled as depending on Intention to purchase which is dependent on Attitude toward products applying a novel production strategy, Perceived behavioural Control, Subjective Norm and Moral Norm. Overall, six constructs were included in the: Perceived Risks (3 items: Bredahl, 2001; Tung et al., 2008), Perceived Benefits (3 items: Bredahl, 2001; Tung et al., 2008), Moral Norm (3 items: Bredahl, 2001; Dean et al., 2008), Subjective Norm (3 items: Dean et al., 2008; O’Connor et al., 2006; Olsen et al., 2008; Tung et al., 2008), Perceived Behavioural Control (3 items: Bredahl, 2001; Saba et al., 2003; Tung et al., 2008), and Attitude towards the production strategies (3 items: Bredahl, 2001; Cook and Fairweather, 2007; Davis et al., 1992; Tung et al., 2008). Purchase Intention was measured as a single item variable.

All multi-item constructs were measured using a 7-point Likert scale (from 1 = ‘‘strongly disagree’’ to 7 = ‘‘strongly agree’’). Only one item of the PBC construct was measured using a different 7-point Likert scale (from 1 = “no control’’ to 7 = ‘‘complete control).

For parsimony in administration of the questionnaire, many constructs were just identified. In any case, a confirmatory factor analysis (CFA) was conducted on multi-item scales (Attitude, Subjective Norm, Moral Norm, Perceived Behavioural Control, Perceived Benefits and Perceived Risks), for each innovation strategy across countries.  Given that the multi-item latent variables were measured by ordered categorical indicators, inspection of the data suggested an estimation method robust to departure from normality. Following  Finney and Di Stefano (in Hancock & Mueller, 2006), we used a Satorra-Bentler scaling of the variables with ML estimation.

Estimation of CFA models for each innovation and cross-country validation resulted in the exclusion of Perceived Risks and Perceived Behavioural Control - also exhibiting poor reliability i.e. Cronbach’s Alpha below 0.63 (King and He, 2006) -  from the final measurement model.

The choice experiment was designed as a labelled one. Each choice task contained a choice among 4 types of milk, each applying a novel production strategy:  milk from dairy farmers applying Agroforestry, milk from dairy farmers using Alternative Protein Sources, milk from dairy farmers applying Prolonged Maternal Feeding and the milk currently purchased. Attributes consisted of production method (3 levels, organic, low input, conventional) and price. The choice sets were built, for each respondent, using the “pivoting” strategy suggested by Hensher et al. (2005). The price levels presented to each respondent were not fixed in absolute sense across the entire sample, but varied according to the price reported for the currently purchased milk, in a previous question, by each respondent. The price shown to each respondent in each choice task varied according to a percentage change (+/- 10% or no change) from the reported price. The milk currently purchased was always offered as a choice with its originally reported price. Prices were in Euro (EUR) for AT, BE, FI, IT, in Danish Kroner (DKK) in DK, and in Pound Sterling (GBP) in UK.   Each respondent was presented with nine choice tasks. The allocation of price levels to alternatives was designed using a sequential D-efficient approach: starting from a 33 full factorial orthogonal design, we designed a pilot 3x3 fractional orthogonal design using Ngene. The final D-efficient design was based on estimating pilot results on 126 observations.

All novel strategy models were tested for cross-cultural validity (i.e. model measurement invariance across countries) and the constructs adapted according to results .

We compare results from hybrid choice models estimated both by a sequential approach (Yáñez et al., 2010) and a simultaneous approach (Ashok et al., 2002; Temme et al., 2008) by using commonly available commercial software: Nlogit, Biogeme, and Mplus software. The approach is a kind of characteristic Latent Variable Model (LVM) for Discrete Choice, in which factors are applied as covariates, that are characteristics of the decision makers (Rungie, 2013).

The two approaches result in substantially similar results. The results allow us to establish that, although consumers had positive attitudes for all novel production strategies (and especially Prolonged Maternal Feeding), this attitude did not significantly influence choice. On average, the average premium price consumers are willing to pay more is never exceeding 60% of current milk price in AT, BE, IT and UK, while Danish and Finnish consumers are not willing to pay a cent more (actually they would like to get a discount for purchasing milk produced according to this strategy). Although in the right tail of the WTP distribution a niche of consumers that are willing to pay substantially more exists, it appears that none of the three novel proposed strategies are actually viable for producers unless specific subsidies ate paid to counteract the higher production costs.

 

REFERENCES

Ashok, K., Dillon, W.R., Yuan, S., 2002. Extending Discrete Choice Models to Incorporate Attitudinal and Other Latent Variables. J. Mark. Res. 39, 31–46.

Bredahl, L., 2001. Determinants of consumer attitudes and purchase intentions with regard to genetically modified foods - results of a cross-national survey. J. Consum. Policy 24, 23–61. doi:10.1023/A:1010950406128

Cook, A.J., Fairweather, J.R., 2007. Intentions of New Zealanders to purchase lamb or beef made using nanotechnology. Br. Food J. doi:10.1108/00070700710780670

Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1992. Extrinsic and Intrinsic Motivation to Use Computers in the Workplace. J. Appl. Soc. Psychol. 22, 1111–1132. doi:10.1111/j.1559-1816.1992.tb00945.x

Dean, M., Raats, M.M., Shepherd, R., 2008. Moral concerns and consumer choice of fresh and processed organic foods. J. Appl. Soc. Psychol. 38, 2088–2107. doi:10.1111/j.1559-1816.2008.00382.x

Hancock, G.R., Mueller, O.R., 2006. Structural Equation Modeling: A Second Course.

King, W.R., He, J., 2006. A meta-analysis of the technology acceptance model 43, 740–755. doi:10.1016/j.im.2006.05.003

O’Connor, E., Cowan, C., Williams, G., O’Connell, J., Boland, M.P., 2006. Irish consumer acceptance of a hypothetical second-generation GM yogurt product. Food Qual. Prefer. 17, 400–411. doi:10.1016/j.foodqual.2005.05.003

Ottar, S., Heide, M., Calvo, D., Toften, K., 2008. Explaining intention to consume a new fish product : A cross-generational and cross-cultural comparison 19, 618–627. doi:10.1016/j.foodqual.2008.04.007

Rungie, C., 2013. A Review of Latent Variable Models for Discrete Choice, in: International Choice Modelling Conference 2013.

Saba, A., Messina, F., 2003. Attitudes towards organic foods and risk/benefit perception associated with pesticides. Food Qual. Prefer. 14, 637–645. doi:10.1016/S0950-3293(02)00188-X

Temme, D., Paulssen, M., Dannewald, T., 2008. Incorporating latent variables into discrete choice models-a simultaneous estimation approach using SEM software. BuR -bus. Res. 1, 220–37.

Tung, F.C., Chang, S.C., Chou, C.M., 2008. An extension of trust and TAM model with IDT in the adoption of the electronic logistics information system in HIS in the medical industry. Int. J. Med. Inform. 77, 324–335. doi:10.1016/j.ijmedinf.2007.06.006

Yáñez, M.F., Raveau, S., Ortúzar, J. de D., 2010. Inclusion of latent variables in Mixed Logit models: Modelling and forecasting. Transp. Res. Part A Policy Pract. 44, 744–753. doi:10.1016/j.tra.2010.07.007

 


Conference registration is required in order to view papers.