Microsoft access as a modelling tool for cognitive work analysis Introduction A number of computer based modelling tools have been proposed and some developed for components of cognitive work analysis or for the framework in its entirety. Inevitably, those tools have limitations. They lack the flexibility that is desired for a framework such as cognitive work analysis which has not yet achieved a stable level of maturity. Additionally, the tools are custom built by single research groups for this specific purpose and lack the organisational support that is desirable for a software tool that could continue to find use through many generations. In this note, I illustrate the use of Microsoft Access as a modelling tool for cognitive work analysis. It is a widely used software application which has been supported by Microsoft through many generations. It has widespread use in diverse domains and is likely to continue as a key component of the Microsoft Office Suite. It is a relational database application and its relational properties in particular allow it to take account of the relationships within and between the multiple stages of cognitive work analysis. The source data for this illustration are drawn from my recent tutorial on cognitive work analysis (Lintern, 2013). Cognitive work analysis in brief Cognitive work analysis is a multi stage framework in which each stage deals with one or more sets of cognitively relevant capabilities and constraints. I will not describe the framework in any more detail here but if you are not familiar with cognitive work analysis, I have a number of resources on my site, www.cognitivesystemsdesign.net, (ranging from a couple of pages up to book length) that explain it. Access is a relational database In common with Excel, Access is a database application. In contrast to Excel, it allows an analyst to build relationships between tables. If you do not understand what that means, or if you are otherwise unfamiliar with Access, you will probably need to consult an Access tutorial. Here I will assume you know the basics. The ability to model relationships is the most powerful reason that Access is a suitable tool for modelling the results of cognitive work analysis. This particular feature allows changes you make in any part of the analysis to flow automatically through to other linked tables. For example, if you change a term or find a spelling error or need to add elements to any of your analytic products, these changes will update automatically in any of the linked tables. I typically use Visio, another Microsoft application, to represent the products of a cognitive work analysis. Inevitably, I find things I want to adjust but any single change can require adjustments in several of the analytic products.
This is a particularly labour intensive and time consuming business. In Access, you make the change once and it updates automatically through the analysis. However, Microsoft Access does not offer the types of visual layouts supported by Visio and other dedicated tools. I find the visual layout to be important early in an analysis as I am exploring the work domain and the patterns of work but less so as I proceed. Typically, I start with Visio and then revert to tables as I become more familiar with the work domain and the patterns of work. Access modelling The analytic products of cognitive work analysis are represented as tables within Access. Figure 1 offers an overview of the tables I have developed for this tutorial example. The links shown in Figure 1 are relationships as modelled in Access. I will illustrate how to build relationships in Access by reference to how I modelled means ends link within the abstraction decomposition space. I will then outline the overall structure of the complete analysis. Means ends links within the abstraction decomposition space I develop individual tables for each of the levels of the abstraction decomposition space and then specify relationships between tables that represent adjacent levels (Figure 1). Note, for example, the relationship between system purpose and domain values. Figure 1: Screen capture of the Microsoft Access relationships view showing the tables and relationship links for the abstraction decomposition space and the work task docket
There is one potentially confusing issue with relationships. The means ends links between levels in an abstraction decomposition space constitute many to many mappings (a single element at one level can support many elements at the next level up and a single element at the upper level can be supported by many elements at the next level down). Access copes with many to many mappings by use of a junction table (a table that sits between the two other tables). I believe the use of a junction table complicates the representation of the abstraction decomposition space without offering any added value. I have chosen, in this modelling exercise, to avoid use of junction tables by treating the relationships as one to many. Given that choice, I needed to decide whether my one to many mappings go from means to ends or from ends to means. That is, should my model show many means mapping into a single end or many ends being supported by a single means? Arbitrarily, I chose the former. For example, my system purpose table lists two purposes and shows which values support them. In the Access relationships view, this is indicated by the line connecting the two tables. You will note the symbols on the line, the number one at the values end and the infinity symbol at the purpose end. Following the Access convention, this is to be read as each value appears only once in the values table but many times in the purposes table. Relationships involving the work task docket I create a work situations table, a work tasks table, a cognitive processes and cognitive states table, a cognitive strategies table, and a cognitive strategies and cognitive modes table (Figure 1). The work situations table lists the potential work situations. The work tasks table lists the work tasks. A one to many relationship with the domain functions table is used to identify the domain functions associated with particular work tasks. A one to many relationship with the work situations table is used to identify the work situations associated with particular work tasks. The cognitive processes and cognitive states table contains a work tasks column, which is populated via a relationships link to the work tasks table. There is also a column for each cognitive state and each cognitive process associated with the decision ladder. Check boxes are used to indicate the cognitive states and cognitive processes activated for each work task. I have so far marked the cognitive states and cognitive processes associated only with the plan and replan work tasks. The cognitive strategies table lists all potential cognitive strategies and also the reasons a worker might use a particular strategy. The cognitive strategies and cognitive modes table has a work tasks column which is populated via a relationships link to the work tasks table. It has a strategies column, which is populated via relationship link to the cognitive strategies table. This column associates work tasks with strategies. Because some work tasks involve the use of multiple strategies (actually or potentially), specific work tasks may appear in the work tasks column more than once. This table also has separate columns for the three cognitive modes (skills, rules, knowledge). Check boxes are used to indicate the cognitive modes activated for each cognitive strategy.
Relationships involving the social transactions docket I create a social transactions table, a transactions reach table, an agents table, a transactions demands table, two transactions dimensions tables (spatial and temporal), and a transaction resources (or design ideas) table (Figure 2). Figure 2: Screen capture of the Microsoft Access relationships view showing the tables and relationship links for the social transactions docket The social transactions table has columns for transaction reach (populated via a relationship link to the transactions reach table), for interacting agents (populated via a relationship link to the agents table), for transaction demands (populated via a relationship link to the transactions demands table), for spatial transaction dimensions (populated via a relationship link to the spatial transactions dimensions table), for temporal transaction dimensions (populated via a relationship link to the temporal transactions dimensions table), and for resources or design ideas (populated via a relationship link to the resources or design ideas table. The transactions reach table has a column for transaction reach. The agents table has a column for agents. The transactions demands table has a column for transaction demands.
Each of the two transactions dimensions tables (spatial and temporal) has a column for the relevant transaction dimension. The transaction resources (or design ideas) table has a column for transaction resources or design ideas. Relationships involving the staffing docket I create a staffing table, a work modules table, and a staff roles table (Figure 3). Figure 3: Screen capture of the Microsoft Access relationships view showing the tables and relationship links for the staffing docket The staffing table has columns for platoon subgroup (populated via a relationship link to the work modules table), and for role and rank (both populated via relationship links to the staff roles table). Additional columns have been created for other criteria but have not yet been filled in. The work modules table has a column for platoon subgroup. The staff roles table has columns for role designation and for rank. Summary This completes my brief description of how I used Microsoft Access to model the results of my cognitive work analysis described in Lintern (2013). I intend the Microsoft Access model that this note describes as an illustration of the way Access can be used rather than as a description of the way it should be used. I have explored several variations on ways of populating tables and developing relationships. I have not included descriptions of those in this note because I wanted to present a relatively simple tutorial that would suggest the possibilities rather than to provide guidance on all requirements.
The use of Microsoft Access as a modelling tool for cognitive work analysis has three primary benefits. It comfortably accommodates the results from analysis of an extensive system, one so large that it would tax the capabilities of tools that offered visual representations of the analytic products. Additionally, changes in or additions to one table flow automatically through to other linked tables. Finally, Microsoft Access is a well established software application with solid organisational support and is likely to continue to be maintained as a viable product through many generations. Reference Lintern, Gavan (2013). Joker One: A Tutorial in Cognitive Work Analysis. Melbourne, Australia: Cognitive Systems Design.