PRE-TEST

A total of six powertrains were included in the market survey: petrol, diesel, hybrid electric, plug-in hybrid electric, pure electric and hydrogen. All powertrains included in the survey are currently available in the market except for hydrogen. Although hydrogen vehicles are not yet available to purchase in Australia, it has high possibility of entering the market in the future. The inclusion of such a powertrain in the stated choice experiment allows the author to explore consumer preferences around this yet-to-be- realized powertrain, which is one of the greatest advantage of stated choice experiment. By adding this powertrain in the survey, the benchmark of how consumers view a brand- new powertrain can be set. The comparison between consumers’ preferences and opinions of this powertrain and other already-available powertrains can be made.

It is worth mentioning that not all powertrains that had existed in the market are incorporated in the survey. In particular, pure LPG and LPG dual fuel vehicles are not selected. Although these two powertrains entered the market relatively early, LPG fuelled vehicles now have very little share of the new passenger vehicle market. There are no LPG dual fuel vehicles selling in the market and only two pure LPG vehicle models provided by Ford. These powertrains have lost the majority of their popularity in the Australian vehicle market since its prime time in the 1990s. Asking consumers preferences towards powertrains that have already been phased out of the market is not necessary and cannot provide much useful information towards the development of future vehicle market landscape. Therefore, these two powertrains are not included in the market survey, nor will be included in the system dynamics model simulation in the following chapters.

In order to enhance the reliability of the market survey, demographic quotas were added to survey panel. A panel that represents the Australian vehicle consumers in terms of basic demographic characteristics such as age, gender, household income, and education attainment level is the data collection goal of this survey. Screening questions are added prior to survey questions to make sure the panel is representative of Australian vehicle consumers. In Section 5.3, the demographics of the survey panel will be compared with the national data for validation of the survey.

Market survey in this thesis is divided into three main parts: the first part is to gauge consumers’ attitudes and knowledge towards AFVs; the second part is to carry out the

stated choice experiment to reveal consumers’ preferences in vehicle purchase decisions; the third part is to collect generic information about respondents’ vehicle ownership and more demographic information for further analyses. In the following sections, design details of these three parts of the market survey will be presented.

5.2.1 Part One: Attitude questions for consumer familiarity and affinity

Australian consumers’ attitudes towards AFVs are measured from four aspects: familiarity and experience level, knowledge and biases, vehicle attributes ranking, and willingness to consider purchasing AFVs.

In order to gauge Australian consumers’ familiarity and knowledge level of AFVs, the survey asked respondents to indicate their experience level with different alternative powertrains out of five familiarity levels, from “never heard of it” to “I have owned/ am currently owning”. Better familiarity and deeper understanding on AFV technologies increase the possibility of alternative powertrains getting selected to consumers’ consideration sets. Survey responses of this question provide a general picture of how familiar Australian vehicle consumers are with different powertrains. This piece of information is also valuable to later dynamic model simulation and calibration.

Biases and knowledge about AFVs were measured by asking respondents to rate every powertrain they know in terms of fuel efficiency, CO2 emissions, reliability and driving

range. Fuel efficiency and CO2 emissions were selected for the reason that these two

vehicle attributes are often regarded as the most beneficial features of alternative fuel powertrains. By asking respondents to rate fuel efficiency and CO2 emissions for

alternative powertrains alongside traditional powertrains, information about whether respondents are fully aware of the benefits of AFVs were retrieved. Reliability was selected because it is regarded as the most important attribute for vehicles in the work of Caulfield et al. (2010). Although reliability is not a vehicle attribute that directly associates with vehicle powertrains, whether or not consumers perceive AFV reliability as the same as traditional powertrains is still an important piece of information to collect. Finally, driving range was selected as the last attribute for the bias and knowledge section. Limited driving range is a major concern for AFVs, especially for powertrains that rely solely on alternative fuel that are not prevalent in refuelling network. Information about whether consumers over or underestimate the challenge they will face if they switch to alternative fuel powertrains were gathered from this question.

In attitudinal questions, consumer preferences were also investigated. Respondents were asked to rate their preferences over twelve specifically chosen vehicle attributes. These vehicle attributes included attributes that were incorporated in the stated choice experiment questions as well as additional attributes that were not included in choice scenarios. The additional attributes are regarded as significant to consumers in vehicle choices through the literature but are not suitable to be included into the stated choice experiment since these attributes normally do not show substantial differences between traditional and alternative powertrains. These questions aimed at gathering information about what vehicle attributes were the most important in consumers’ minds.

In the end, in attitudinal questions, consumers’ willingness to consider AFVs were investigated. Respondents were required to first state if they had considered alternative fuel powertrains in their latest vehicle purchase. They were then asked if they would consider an alternative fuel equivalent to their latest vehicle purchase if the vehicle brand and model were kept the same. The second question was designed to measure consumers’ willingness of AFV adoption without the interference of limited alternative fuel model variety.

5.2.2 Part Two: Stated choice experiment for consumer preferences

The second part of the market survey was dedicated to the stated choice experiment. The stated choice experiment can provide quantitative data for how consumers make their decisions within their evoked set. The coefficients associated with each vehicle attributes reveal how performances of different vehicle attributes influence consumer choices. Before respondents proceeded to the stated choice questions, they were provided with a description page introducing each powertrain as well as all vehicle attributes that were included in the experiment. The provided description covered the basic information about the powertrains and vehicle attributes so that respondents clearly understood their tasks and the choice scenarios presented. Providing such information, although in an impartial way, might affect respondents’ perceptions of the powertrains negatively or positively, especially in cases where the powertrain is completely unfamiliar to the respondents. However, since the description only provided the very basic information of the powertrains, the experiment could still capture consumers’ attitudes and opinions towards alternative powertrains. The slight potential perception sways due to provided

information were negligible compare to the overall consumer attitudes and opinions that the experiment captured.

The experimental design was created using statistical software SAS following the d-efficiency design method (Kuhfeld, 2010). Every respondent answered a total of 16 choice scenarios questions (an example of one choice scenario was shown in Figure 5-2), with each question providing 6 vehicles with different powertrains to choose from. The experiment was designed to be a branded choice experiment with each powertrain representing an alternative in the choice questions. The six powertrains that were included in the experiment are petrol, diesel, HEV, PHEV, EV and hydrogen vehicles. This wide range of powertrain types were selected to cover not only powertrains that are already exist in the market but also ones that are likely to be later introduced into Australia.

Figure 5-2 Example of choice scenario question in the survey

There were five vehicle attributes used in the stated choice question design: vehicle purchase price, annual fuel cost, car style, fuel availability and driving range. The first four attributes were generic attributes that applied to all alternatives. While the last attribute, driving range, was an alternative-specific attribute that only applied to pure electric vehicles and hydrogen vehicles. The detail of the attributes and their levels can be seen in Table 5-1.

Table 5-1 Stated choice experiment attributes and levels

Attributes Alternative Number of levels Levels

Purchase Price

(AUD) All six 8

20k, 33k, 46k, 59k, 72k, 85k, 98k, 111k

Annual Fuel Cost (AUD)

Petrol 4 800, 1600, 2400, 3200

Diesel, HEV 4 600, 1200, 1800, 2400

PHEV, EV 4 500, 700, 900, 1100

Hydrogen 4 1200, 1600, 2000, 2400

Vehicle Body Style All six 4

Sports/Coupe, Hatch/Wagon, Sedan, SUV/People mover Fuel Availability (%) EV, hydrogen 4 15%, 40%, 65% and 90% of current petrol service stations Petrol, Diesel, HEV

and PHEV - 100% fuel availability

Driving Range (km) EV 4 150, 300, 450, 600

Hydrogen 4 300, 400, 500, 600

There were two monetary attributes in the experiment, purchase price and annual fuel cost. A wide range of price levels was selected from 20k to 111k in order to cover the different prices for all body types and powertrains. No price increments were added to alternative powertrains to find out consumers’ preferences towards powertrain despite any price differences. In this experiment, fuel efficiency was treated as a monetary attribute instead of a technical attribute. The reason of such arrangement was to reduce respondents’ confusion about the different fuel efficiency units for alternative powertrains. Annual fuel cost is much easier to understand and relate to for respondents who are not tech-savvy (Massiani, 2014). It can also include the effect of fuel price so that the cognitive differences on powertrain prices between respondents are eliminated. Vehicle body type describes the size and style of the vehicle. The size and style of the vehicle is particularly important to consumers however not often included in stated choice studies (Massiani, 2014). Here in this study, four vehicle body styles depicting both the style and size of the vehicle were included in the experiment so that aesthetics and functionality of the vehicle can be reflected in the choice question to some extent. In

addition, visual aids (silhouettes of different vehicle body types) were provided in each choice scenarios to help respondents to distinguish different car sizes easily (Figure 5-2). Attribute fuel availability was designed to investigate consumers’ opinion about powertrains that require alternative refuelling, i.e. pure EV and hydrogen vehicles. Comparing alternative fuel availability to the established petrol service station network is the most intuitive way for respondents to gauge the fuel availability level of new alternative powertrains. Thus, percentage of the current petrol service station number were used in the survey to indicate fuel availability for our respondents. By doing this, respondents were automatically offered the information that the fuel availabilities of powertrains that can be refuelled at the petrol station (i.e. petrol, diesel, HEV and PHEV) are 100%. In the later model regression, fuel availability was used as a generic attribute simply because the respondents were provided with the information.

Driving range is an alternative-specific attribute for electric vehicles and hydrogen vehicles. Together with fuel availability, driving range has been identified to be a crucial vehicle attribute relating to consumers’ fuel range anxiety (Needell et al., 2016, Franke and Krems, 2013). Four levels of driving ranges were selected for EV and Hydrogen vehicles. Different values were chosen to mimic the technology limitation these powertrains were facing.

5.2.3 Part Three: Demographic information from respondents

The final part of the market survey comprises questions that aims to get demographic information about respondents. More general questions about respondents’ household size, the model and year of their latest purchased vehicle, their weekly commuting distance, and their vehicle ownership were asked. Respondents’ age, gender, education and household income information were gathered to assist the analyses of survey results of part one and two. This demographic information adds additional dimension of the survey results analyses and allows the author to further understand consumer choices influenced by demographics.