Office of Operations
21st Century Operations Using 21st Century Technologies
Table of Contents

Use of Decisionmaking and Information Management Systems in Mainstreaming TSMO

4. Considerations for Mainstreaming TSMO

Factors and Biases Affecting Decisionmaking

Several factors and common decisionmaking biases can influence individual decisionmaking and impact TSMO. These decisionmaking biases and other factors lead to the need for additional supportive software, such as DSSs and IMSs. These biases and factors affecting decisionmaking are not specific to personnel in transportation agencies and can occur in any setting. Managing these decisionmaking biases can lead to better decisions that apply TSMO more effectively, increasing its credibility and making it more likely to be mainstreamed.

Decisionmaking is a complicated process with a myriad of potential influences to consider. These influences can affect the decisions of individual staff, including those who are in leadership positions and drive overall organizational strategies. In an evaluation of different managerial decisionmaking processes, Omarli (2017) determined that factors affecting the administrative decisionmaking processes were personal, environmental, and psychological. The growing amount of data and fluid environment that TSMO decisionmakers are in can make TSMO decisionmaking prone to some of these factors.

There are four common biases that can affect decisionmaking in everyday tasks and are quite common in organizations:

  • Framing. A common decisionmaking bias involves people reacting differently to information depending on its phrasing, context, or “framing” (Tversky and Kahneman 1981). This bias can have profound impacts on organizational change efforts. A strategy to mitigate this bias is to change labeling/logos (which is why organizations will often spend time/resources on marketing refreshes), colors, or codes to indicate clearly that the context has changed (e.g., from normal operations to emergency operations or from one organizational structure to another). In addition, one should be aware of how information is presented and whether it may be framed in a negative or positive way, especially when making a business case to leadership to support TSMO efforts.
  • Confirmation bias. People often favor or seek out information that confirms a prior hypothesis or belief, leading to confirmation bias. (Wason 1968) This bias can affect leadership when there is the tendency to focus more on data that support an initial approach or only listen to opinions that support their plans. Instead, an alternative decisionmaking process or a properly deployed framework (with appropriate metrics and reporting support) could be to sample the full range of both negative and positive possibilities rather than just the positive ones.
  • Anchoring. Individuals have the tendency to rely on the first piece or limited pieces of information when planning or forming an estimate; this is known as anchoring (Ariely 2008,Tversky and Kahneman 1973). This bias often manifests itself in operational situations where the first incoming field reports (e.g., of evacuation times on a roadway) will drive estimates or the more salient images will affect planning. To mitigate this bias, one should be careful about weighting early or limited information and should generate alternative or counterfactual options. Another option is to constantly refine estimates as data become more reliable over time. (FHWA 2018) An ideal framework and reporting set of tools for management would iteratively adjust estimates as new data come in and present options across the full range to combat the tendency to overweight one part of the spectrum based on early estimates.
  • Groupthink. A bias that is particularly salient in more hierarchical and structured organizations is the concept of groupthink, as demonstrated by the famous Asch experiments (1951). It is defined as a desire for harmony, often at the expense of optimal solutions. Subordinates or peers may follow along with sub-optimal approaches so that the team or organization can “get along.” When leading a change effort or in a position of authority, one can combat this bias by: (1) encouraging objections consistently and publicly, (2) not indicating preference for a particular choice or approach until after the team has provided their opinions, (3) asking designated members to play “devil’s advocate,” and (4) regularly evaluating previous patterns to determine if there has been a standard approach that is regularly repeating (i.e., a “rut”). Transportation agencies are hierarchical and structured, so groupthink is a potential problem. If an agency wants to avoid some of the pitfalls of groupthink, then making it clear that alternative opinions and truth will be rewarded can help to increase the comfort level and improve information sharing, which is often restricted when groupthink is endemic. This can be accomplished by welcoming the identification of what is not working well so that it can be addressed.

Observations in Using Information Management Systems and Decision Support Systems for Mainstreaming TSMO

As noted in the introduction, there are no known examples of DSSs and IMSs specifically for mainstreaming TSMO; however, as discussed in the previous sections, many of these systems were integrated into larger systems and enabled data sharing, which facilitates mainstreaming TSMO. In addition, IMSs and DSSs possibly could be tailored for mainstreaming TSMO purposes (e.g., using big data to augment business intelligence decisions that rely on an enterprise-level DSS to make a business case for TSMO).

There are several general considerations and lessons learned for IMSs and DSSs that can support mainstreaming TSMO:

  • User needs. It is vital for a DSS or IMS to support the needs of the users and accommodate their knowledge, skills, abilities, and goals of activities. Any software used to mainstream TSMO efforts should be developed with the user at the center. With DSS, one should also be mindful of the points along the decisionmaking pathway where biases and errors tend to be most prevalent in most users. This will become critical as DSS software is brought on that can help agencies mainstream TSMO, especially when that means interacting with staff who do not fully understand or appreciate TSMO.
  • Data fusion, integration, interoperability, and quality. Several examples noted the integration of disparate databases or connecting systems, both intra- and inter-agency (the latter is of particular importance within the ICM context). Agencies can consider ways to standardize data variables (and any data collection instruments) and structures across units to the extent possible as well as develop plans for integrating databases, performing data fusion (which entails a further step of replacing or reducing the data), and ensuring interoperability. This effort will allow for lower costs for developing and integrating both an IMS and a DSS. It will also facilitate planning, operating, management, staffing, and other investment decisions that take into account data related to operations as well as other data types. The completeness, accuracy, reliability, and fidelity of the data are important for supporting whatever decisions are made (as well as the accuracy of those options developed through analysis or provided to decisionmakers).
  • Software flexibility and usability. To facilitate use, software needs to be useable and appealing, as well as allow for multiple users with different goals to interact with it to meet their needs. Allowing users the ability to tailor or modify reports to suit each user’s (or unit’s) needs is also important. For example, Pennsylvania DOT mentioned a district executive scorecard that has different performance measures that are easily pulled based on the needs of executives and other users, with tailorable output. (Pennsylvania DOT 2018) This flexibility will allow for more widespread use, adoption, and integration into the normal DOT workflows, which will go a long way in mainstreaming TSMO.
  • Maintenance and evolution. The costs of maintaining a DSS/IMS or updating/upgrading both systems as data and information change are often overlooked in planning. Ideally, both the DSS and IMS would be capable of supporting the integration of modular components that can be updated easily.
  • Monitoring and evaluating performance. Another often-overlooked aspect of DSS and IMS is the ability to monitor their performance and evaluate that performance against a benchmark to assess the value added. That information would be helpful in making the business case for mainstreaming TSMO in any organization.
  • Planning. As noted by several DOTs, the software development and integration process was also part of the IT planning process (including budgeting) and policies. Early involvement of TSMO will allow it to be integrated into the core components and plans of any DSS and IMS (see the Wisconsin DOT example in table 4). Systems or tools, such as Ohio DOT’s TOAST, allow TSMO to be more readily and consistently considered in the transportation planning process.
  • Partnerships. Several examples noted in the previous sections highlighted the importance of forming partnerships with local universities, vendors, and other non‑transportation agencies that may be of benefit. Efforts to mainstream TSMO will also likely benefit from leveraging these partnerships in the context of IMS and DSS.
  • Culture. Software (e.g., for DSS and IMS) is not developed or integrated in a vacuum. The organizational culture and structure provide a base for how that software is to be used. Software can have TSMO’s needs as a central component, indicating TSMO is central within an agency. Conversely, software that is developed with an eye toward integrating TSMO throughout the agency may help to drive a change in culture that supports mainstreaming TSMO.

These lessons learned and discussion points are meant to provide several topics for consideration in an agency’s efforts using DSS and IMS to support efforts to mainstream TSMO.