Join Doug Rose for an in-depth discussion in this video Run question meetings, part of Learning Data Science: Ask Great Questions.
- I once worked for an organization that was just starting out with data science. The team was in a meeting to predict how many people would participate in medical studies. The data analyst showed charts of their past studies. He said they could create a data model to predict who might participate in future studies. A week later, the team got back together. The data analyst showed historical data that suggested that each study would have a certain number of participants. Again, the team listened and wrote down the predictions. After the meeting, I asked the team why they didn't have any questions.
They said that the data analyst knew the data better than anyone. They didn't feel comfortable questioning the results. This is a common way that many organizations view questions. They see questions as a form of judgement. It's not seen as a way to learn, instead it's seen as a softer way to tell you that you're wrong. Questioning and learning are the key differences between data science and just looking at the data. Remember that data science is about using the scientific method to gain insights. Asking good questions is the core of the scientific method.
One of the best ways to make sure that your team is asking good questions is to set up a good environment to exchange ideas. If you're the research lead, it's your responsibility to be a questioning leader. Set a focus on learning. Everyone on the team should strive for deep listening. This is a much more focused way to listen to each other's ideas. They should push back against these ideas without feeling like they're being judged. A good way to set-up this environment is to have question meetings. In these meetings you encourage the participants to ask questions before making statements.
This is sometimes called a question-first approach. These meetings are about creating the maximum number of questions. They're focused on everyone asking their questions and listening. If you're the research lead, discourage anyone from bringing their smartphones or laptops. You want everyone focused on listening to each other. Imagine if the meeting about medical studies was run with a focus on good questions. The research lead would start up the meeting by asking, does everyone know why we're having this meeting? Then wait for a response.
A good question-leader is not afraid of short periods of silence. Don't try to answer your own questions. Give everyone in the room some time to think about their answer. Once you're satisfied that everyone understands the meaning, then they should present the challenge. The research lead can start the meeting by summing-up the challenge with something like, we waste a lot of money because we don't know who will show up for our medical studies. Then leave the solution to the team. If you're the research lead, you might want to leave the question open-ended. You could say, we need to do a better job predicting who might show up.
Then sit down and see if anyone starts asking questions. If after a few minutes no one says anything, then you can ask another question. Maybe something like, does everyone understand why this is a challenge? What you're hoping to get from the team is something like, why do some studies fill up more than others? You want to avoid quick statements that shut down the conversation. You wouldn't want someone to say, let's research how other companies fill up their studies. This will keep people from coming up with their best ideas. Remember, it's the discussion that will give your team the greatest value.
You want the team to feel comfortable exploring the data. If you're the research lead, don't get too discouraged if the team has a hard time asking great questions. Most organizations still feel that it's better to have a group of people who make statements. This clarity is still seen as more valuable than not knowing. It might take a few meetings before people are comfortable asking good questions. Once you have a few question meetings, you might find that a group of people who ask tough questions will often gain more insights than people who make quick statements.
- Harnessing the power of questions
- Testing your reasoning
- Identifying question types
- Organizing questions
- Rooting out assumptions
- Finding errors
- Highlighting missing data
- Overcoming question bias