How can you improve your critical thinking skills when working with data?

For me, Data tells the truest story yet people have the power to design their own story or narrative based on the same dataset. That's the beauty of data, people and the problem. one must want a certain outcome to define a problem when working with data. for example, while dealing with pass rates of students at one of my previous roles, I felt I knew the problem, students are not passing. But that's too generic and data was not giving me anything. But when I determined the outome .e.g. increase the pass rate by % and design intervention considering factors, I started to look at the data in much more efficient manner and started segregating data to get better result.

Understanding and defining the problem is like the first step in figuring things out. Gathering information and data, is the preferred and common method for the majority of people, and it is undoubtedly an effective and ideal method! But what about the history of the problem! Looking at the history of the problem is a means of gaining deeper insights into its evolution and understanding the context in which it has developed. Examining the chronological aspects allows individuals to identify patterns, root causes, and potential contributing factors that may have led to the current situation. This historical perspective provides a comprehensive view, shedding light on the nuances and complexities that might not be immediately apparent.

In my experience in working with biomedical data, objectivity is of primary importance to produce reliable findings and progress the field. Reducing bias and ulterior motives is critical to preserve the integrity of research and these can creep into our work easier than you realize. Also succinctly framing the problem/question and formulating the testable plan to execute that goal is of vital importance to avoid distractions and work efficiently when managing large sets of data.

You need to be really careful of bias while analyzing data and be skeptical. If you have a set hypothesis/ outcome, be extra cautious because you will be narrating the story you WANT to tell by the data you choose and analyze; however, the story you want to tell might not be the narrative that serves the purpose. one must be able to go back to full circle and use comparative models to justify the analysis..

When analyzing data, I find it’s crucial to first understand the business or research question at hand. Each dataset tells a story, and as an analyst, it’s my job to piece that narrative together. I often start by examining the data from different angles, looking for patterns or anomalies. It’s not just about running statistical tests, but about understanding why a particular number is the way it is. I try to present my insights in a way that is accessible to all stakeholders, often using visuals or analogies. Lastly, reflecting on the data means questioning my own biases and considering the ethical implications of data use. This comprehensive approach has helped me deliver actionable insights that drive strategic decisions.

Evaluating data is a multi-faceted process that hinges on robust statistical validation. For instance, consider the concept of 'representativeness', which is pivotal to the validity of conclusions. In practical terms, this means ensuring that your sample accurately reflects the population you're studying. If you're analyzing customer data to inform product development, you might apply stratified sampling to account for different customer segments, thus preventing bias towards a particular group. This approach enhances the credibility of your findings, as it mitigates the risk of drawing conclusions from a skewed dataset that does not encapsulate the diversity of the entire customer base.

To add to this answer, It is easier for the people who create the reports, However, one needs to think about other's perspective. Not everyone is a data person, and neither is critical thinker. You might think you are being clear and concise; may be to the audience you are not making sense due to technical jargons. Best to choose simplicity over overly decorated information.

(Plan, Do, Check & Act) - PDCA cycle is a proven method of problem solution, application of PDCA produces better results, the best part of this methodology is that continuous improvement is "built-in".

Actually answers covered almost of it but yet, in my experience, the main challenge will be always cognitive biases of all their types. To avoid them cultivate self-awareness, question assumptions, seek diverse perspectives, and encourage open-mindedness. Use data and evidence to inform decisions, implement decision-making frameworks, and invite constructive criticism. Take time to reflect, consider alternative viewpoints, and be mindful of emotional influences. Always review and update processes on regular basis to minimize bias and foster your critical thinking