The extent to which deregulation increases the competitiveness of retail electricity markets depends largely on consumer switching activity. The USA, UK, Norway, Sweden, and Australia have all implemented electricity market reforms but consumers have often been reluctant to switch suppliers. In New Zealand, most consumers have not switched suppliers despite potential annual power bill savings of $150. Campaigns promoting consumer switching rely on price differences and ignore the value of non-price attributes, including whether the electricity is generated from renewable energy sources. This thesis improves our understanding of consumer switching and the demand for green electricity by analysing consumer preferences for the attributes of electricity services, estimating willingness to pay (WTP) for non-price attributes, and explaining consumer switching in terms of eight attributes:- power bill, call waiting time, fixed rate contract, discount, loyalty rewards, renewables, ownership of supplier, and supplier type. The analysis is based on a panel choice dataset generated using a choice experiment which was administered in 2014 to an online panel of 224 electricity bill payers in New Zealand. The multinomial logit, random parameter logit and latent class models are used to analyse the choice data with psychological constructs included to explain heterogeneity of preferences. The effect of attribute non-attendance (AN-A) and hypothetical bias on WTP estimates is investigated. We also explore whether using shorter versions of the New Ecological Paradigm (NEP) Scale to measure environmental attitudes (EA) affects estimates of WTP for green electricity. The results indicate that non-price attributes of electricity services are significant determinants of consumer switching. Three latent classes with distinct preferences for the attributes are identified. Class 1 (40%) is mainly concerned about the power bill, and would switch supplier to save at least NZ$125 per year in power bills, ceteris paribus. This value mainly captures the status quo effect. Class 2 (46%) exhibits no status quo effect and values all the attributes offered including renewables, and particularly dislikes entrants from other sectors which have to charge at least NZ$135 less per year compared to a traditional retailer for a 50% chance of attracting customers. Class 3 (14%) consists of captive and loyal respondents who would not switch supplier for any realistic bill savings. We find that failing to account for attribute non-attendance results in WTP estimates that are significantly lower for attributes that are not normally included in standard electricity plans. Also, respondents who claim to have ignored some attributes may not have done so; instead they assigned lower weights to these attributes. Respondents with low certainty scores are less sensitive to the power bill and are predicted to have significantly higher WTP. We find that using shorter versions of the NEP Scale to measure EA, increases bias in WTP for green electricity. We conclude that price differences in retail markets reflect, in part, consumer preferences for non-price attributes, and that providing consumers with information on the levels of non-price attributes could influence switching rates, the uptake of green electricity and potentially the level of competition in the retail electricity sector.