The Exploratory Phase

Almost every study should begin with an exploratory phase in which the researcher gets used to the research site, and develops the theories that will be tested later.

There are several objectives of this phase:

Understanding the lingo. It is especially important for those team-members who are not employees of the company to pick up the jargon and the site-specific references that litter. This will be useful not only so you can understand what people are saying, but also so that you can use language they understand when creating questionnaires and other data collection instruments.

Understanding the context. The specific problem or question that you are investigating does not occur in a vacuum. To fully understand why things happen in a place the way they do, you need to understand the larger context in which they are embedded.

As you know, the organization has a culture. There are norms, there are cognitive schemas, there are beliefs. These cultural and cognitive elements serve to determine how people will frame or assign meaning to events, and how they will construct responsens to those events.

Culture is layered like an onion. At the core, are fundamental beliefs, values that are unquestioned and maybe even unknown. On the outer surface are today's opinions about how we should do this or that, what the market will do -- these are constantly changing. In the middle are cultural elements that are well-connected to core values and therefore have some permanence. But changes here would not threaten the existence of the organization.

The organization also has a structure. It has a certain size. It has a certain allocation of people to locations. It has a set of core technologies that it uses to produce its products or services. It has a division of tasks. It has a division of labor. It has a degree of differentiation. It has mechanisms of integration, coordination and control. It has rules, procedures, career paths, job slots, reporting relationships, departments, teams, groups, divisions.

The organization has politics. There are personalities, ambitions, interests, power-bases. There are diverse stakeholders pushing and pulling in different directions. For some of the external stakeholders, the organization is just a pawn or battleground as they deal with their own politics in a much larger system.

Developing an initial diagnosis. The ultimate objective of exploratory work is to develop a theory (typically, several theories) of how things work. You want to sift through the thousands of variables and relationships among variables that might impinge on the situation and boil them down to those few that you believe to have the most effect on what you're studying.

It is important to focus on the important things and throw away the rest in order to bring order and clarity to bear. A model of the real world is useful only if (a) it captures the key processes, and (b) it is significantly simpler than the real world. If it is not simpler than the real world, you have not achieved any explanation and there has been no value added -- the model has no power.

Once you have boiled things down to a few basic ideas, you can decide what data you really need to collect to answer remaining questions. In other words, you can choose the variables that you will need to measure in order to test your theories and really nail down your understanding of what's happening.

A note on a paradox. The main purpose of exploratory work is to collect some initial data in order to develop a theoretical framework which then guides further data collection. But without a theoretical framework to start with, how do you know what data to collect in the first place? If you think about it, you will realize that the only way to do it is to have a theoretical framework. A key function of the theoretical framework is to tell you what is relevant and what may be ignored. Without one, you will pay attention to the color shoes people wear, and the number of flowers in the wallpaper on the wall.

It seems circular, but what it really is is a spiral. We all carry with us generic theories of how things work which we have inherited from our culture and learned from experience and are implicit in language and there are some really basic theories, like the notion of cause and effect, that may be hardwired into mammalian brains.

These generic theories, which we are usually quite unaware of, guide our initial exploration into a problem. For example, we all have beliefs about humanity in general. For example, some people basically trust other people. They think that even in the worst human beings, there is a core of goodness that has been buried by a rough life, getting in the wrong crowd and so on. Others think that human beings are fundamentally self-serving, utility-maximizing animals. If they give to charity, it is so that people will think highly of them. If they never break a law, it is because the threat of jail is a deterrent. If they never lie, it is only because they are afraid of being caught. These fundamental notions of human nature affect how we understand human behavior in the workplace. They tend to lead us down some paths rather than others.

Another generic theory is that people are rational. They make choices among alternatives based on the perceived features of those alternatives and the utility of those features to the individual. For example, if there are three products on the shelf and each has its benefits, the person calculates the sum of benefits weighted by the value to obtain an overall utility, and then buys the one with the highest utility.

One Consumer's Ratings of Products on Attributes on 5-point Scale

The important thing is to realize is that the theories we carry around with us unconsciously circumscribe what we think of as relevant data, which in turn determines in part what our new theories will be, and so on.

When you do your exploratory work, one useful heuristic to help guide you is to think in terms of systems. For example, I helped on a project last year in which a company was experiencing some clashes between management and employees of a department. This was a department of phone representatives where what they do is answer the phone and deal with customer questions and problems about their accounts. This had been a happy group before, but now there was some burnout and resentment toward management. One of the complaints the phone reps had was that certain training classes were often canceled by management. Their grade promotions and consequent salary increases were based on their obtaining certain skills which were taught in a series of classes run by the training department. Based on exploratory data analysis, I constructed the following diagram:

In other words, when classes were held frequently, employees where happy, but if there was a lot of customer demand, the classes would take people off the phones, which would increase customer holding times which affected the managers' pay, so they would cancel classes.

A proposed solution was to create an automated, self-paced training module.