Differential Item Functioning (DIF) Analysis

Evaluating the degree to which measure meaningfulness is generalized across subgroups within a population is important. Studies that focus on validity at the item level within an instrument are investigating differential item functioning (DIF) (Myers, Wolfe, Feltz, & Penfield, 2006). Notably, DIF exists when an item functions varies with respondents from different groups.

Table 5.7 presents the DIF results for private and public transport commuters.

The DIF measure columns represent each item’s difficulty logits separately for those using public and private transportation. The DIF contrast column represents the difference between the DIF measures. A t-test was applied to examine differences between the measures of the two groups. The p-values indicate that policy beliefs associated with the items differ significantly between the two groups when marked with an asterisk (*) if p < 0.05. In this case, the DIF contrasts of Items 4, 11, and 12 are negative, which indicates these items cause more difficulties for people

commuting by public transport than private transportation. In other words, people who commuted by private transport are more confident than those using public transport commuters that providing instant traffic information to reduce driving time, subsidizing the public to modify cars to use LPG, and implementing electronic toll collection (ETC) would be effective approaches for achieving sustainable transportation,. In addition, the positive DIF contrast values for Items 8, 9, and 15 imply more difficulties for people who commuted using private transport than those using public transport. That is, people who commuted using public transport are more confident than private transport commuters that congestion road pricing on CBD, increasing gasoline prices to reduce car use, and increasing parking fees to reduce car use are better approaches to achieving sustainable transportation.

Table 5.7 The differences between various commute modes for each item and measuring policy beliefs (sorted by the DIF contrast)

Item 4 Provide instant traffic information to reduce driving

time and, thus, reduce oil consumption and improve air quality

-0.21 0.19 -0.40 0.0015*

11 Subsidize the public to modify cars by using LPG 0.34 0.66 -0.32 0.0052*

12 Implement electronic toll collection (ETC) on highways to alleviate congestion and, thus, reduce emissions

0.02 0.26 -0.24 0.0490*

5 Develop new energy sources (electrical cars, fuel-batteries, etc.) to substitute for the use of fossil fuel

-1.17 -1.02 -0.16 0.3309

14 Ramp metering and HOV lane for freeways 0.39 0.54 -0.16 0.1673 10 Encourage public and private firms to use public

transport to slow the use of cars and motorcycles

0.28 0.42 -0.14 0.2328

2 Build public transport centers to facilitate transfer between different public transportation modes

-0.88 -0.83 -0.05 0.7513

7 Build a friendly walking environment to reduce the use of private modes

-0.82 -0.79 -0.03 0.8383

13 Total vehicle volume control by limiting authorization of licenses (set quotas) to reduce auto growth

0.26 0.21 0.05 0.6746

6 Establish bicycle lanes to promote the use of bicycles -0.67 -0.75 0.08 0.6003 3 Integrate schedules and tickets for public transportation

(e.g., EasyCard, One Day Pass, etc.) to make transfer easier

-0.46 -0.57 0.11 0.4589

1 Construct rail transport systems (e.g., MRT, HSR, Train, etc.) to promote public transportation

-0.93 -1.14 0.21 0.1931

15 Congestion Road Pricing on CBD 1.08 0.85 0.23 0.0283*

8 Increase gasoline prices to reduce car use 1.44 1.14 0.29 0.0053*

9 Increase parking fees to reduce car use 1.23 0.90 0.33 0.0021*

*p < 0.05

Table 5.8 presents the DIF results for people who owned or did not own a passenger car. Again, the p-values reveal whether policy beliefs associated with the items differed significantly between the two groups (i.e., * = p < 0.05). If the DIF measures of the two groups are both negative, it indicates the respondents of both groups generally have weak beliefs that the policy will result in sustainable

transportation, and both consider them to be difficult policies. Therefore, the items the not lend themselves to discussion in spite of the fact the DIF contrasts are significantly different. The results of the analysis examining subgroups that used private versus public transportation were the same as for those in subgroups that owned or did not own a passenger car.

Table 5.8 Differences between owning passenger car or not in terms of each item for measuring policy beliefs (ordered by DIF contrast)

Item 1 Construct rail transport systems (e.g., MRT, HSR, Train,

etc.) to promote public transportation

-1.39 -0.94 -0.44 0.0415*

15 Congestion Road Pricing on CBD 0.69 1.04 -0.35 0.0141*

9 Increase parking fees to reduce car use 0.81 1.15 -0.34 0.0153*

8 Increase gasoline prices to reduce car use 1.06 1.36 -0.31 0.0243*

3 Integrate schedules and tickets for public transportation (e.g., EasyCard, One Day Pass, etc.) to make transfer easier

-0.60 -0.49 -0.11 0.5549

13 Total vehicle volume control by limiting authorization of licenses (set quotas) to reduce auto growth

0.19 0.26 -0.06 0.6868

2 Build public transport centers to facilitate transfer between different public transportation modes

-0.88 -0.86 -0.02 0.9173

14 Ramp metering and HOV lane for freeways 0.47 0.45 0.02 0.8669 5 Develop new energy sources (electrical cars,

fuel-batteries, etc.) to substitute for the use of fossil fuel

-1.08 -1.11 0.03 0.8835

7 Build a friendly walking environment to reduce the use of private modes

-0.69 -0.83 0.15 0.4436

10 Encourage public and private firms to use public transport to slow the use of cars and motorcycles

0.54 0.29 0.25 0.0822

12 Implement electronic toll collection (ETC) on highways to alleviate congestion and, thus, reduce emission

0.36 0.06 0.30 0.0466*

4 Provide instant traffic information to reduce driving time and, thus, reduce oil consumption and improve air quality

0.23 -0.10 0.33 0.0344*

11 Subsidize public to modify car by using LPG 0.74 0.42 0.33 0.0219*

6 Establish bicycle lanes to promote the use of bicycle -0.40 -0.77 0.37 0.0431*

*p<0.05

5.4 Discussion

From this study, it was found that the public are more inclined to support the idea

of providing a more efficient and friendly public transportation service to attract their patronage than limiting private car use by increasing usage costs. In other words, they are more likely to support strategies that do not infringe on their preferences and freedoms. More specifically, people enjoy the freedom and flexibility their cars provide, and they would not want to be forced to change their transportation mode simply because of the increased costs of car use. Thus, people would support strategies that provide the options for people to make a choice to use public transportation or develop new energy sources (electric cars, fuel-batteries, etc.) to substitute for fossil fuel rather than strategies forcing people to use public transportation.

Developing new energy sources to substitute for fossil fuel can benefit public transportation from energy intensity improvement, but it takes a long time and can be difficult and expensive. Huby and Burkitt (2000) show that improving public transportation alone is not likely to make a significant impact on car use and should not be the only means to solve sustainable transportation problems. Therefore, developing an economic means (such as increasing usage costs) to limit private car use is indispensable to sustainable transportation in the short-term. Even if people have lower confidence in strategies related to increasing gasoline prices or parking fees, it does not mean those cost-increasing strategies are not worth pursuing. Policies to achieve sustainable transportation should be guided by strategies that alternatively use rewards and penalties. Through this approach, people can quickly learn which behaviors are and are not encouraged.

Respondents’ commuting modes were also collected by this survey. The result shows that the policy preference rankings of the public transport commuters are the same as those using private transport modes. In terms of policy difficulties, private

transport commuters indicated that constructing rail transport systems is the best policy for achieving sustainable transportation. It seems reasonable to conclude that private transport commuters were more comfortable with higher personal energy use.

The private transport commuters agreed with the idea that developing new energy sources is beneficial to facilitating sustainable transportation; however, slow action cannot save a critical situation; they also think constructing rail transportation systems will advance sustainable transportation.

It is important to note that the bi associated with providing instant traffic information for private transport commuters was positive, but negative for public transport commuters. That means most private transport commuters believe providing instant traffic information is relatively useless for enhancing sustainable transportation.

On the other hand, respondents willing to act to reduce private transport usage, as compared to all respondents, were more confident that the idea of constructing rail transportation systems would be of more benefit to facilitate sustainable transportation. People who would like to mitigate the use of their private transport would be attracted to using public transport to commute. The analysis also revealed that the policy beliefs of people willing to act to mitigate private transport using the economic instruments were stronger than all respondents’ in achieving the goal of sustainable transport.

DIF is the resulting condition when an item functions differently for respondents from different groups. In this study, people who commuted by private transport were more confident than public transport commuters that providing instant traffic information to reduce driving time, subsidizing the public to modify cars to use LPG,

and implementing electronic toll collection (ETC) would achieve the goal of sustainable transportation. In addition, people who commute by public transport were more confident than private transport commuters that congestion road pricing on CBD, increasing gasoline prices to reduce car use, and increasing parking fees to reduce car would be better approaches for achieving sustainable transportation. The findings for subgroups of people who owned or did not own a passenger car were the same as for the subgroups using private or public transport.

C HAPTER 6 A NALYSIS OF THE B ELIEFS IN S USTAINABLE T RANSPORTATION P OLICIES FROM S ENIOR O FFICIALS

AND THE P UBLIC

In this work, policy beliefs regarding sustainable transportation from the senior officials and the public were investigated and the conclusions are provided in previous chapters. It should be noted that the differences in the two questionnaires included the dates of the survey, policy items, and policy beliefs in the contributions of sustainable transportation policies. However, in order to explore differences in policy beliefs between the senior officials and the public, this chapter analyzes the differences between the two groups (the public vs. senior officials) and provides some noteworthy outcomes.