Striking the Balance: Logic and Data in Decision-Making

The effective use of data enhances the precision and objectivity of decisions but developing logic remains to be crucial for building a solid decision-making foundation prior to using data. A well-rounded approach involves honing both logical reasoning skills and the ability to work with data, creating a synergy that strengthens analogic problem-solving. Undermining the sequence on developing logic prior to building data-skills, is a major talent derailer.

A deluge of interest on developing data skills, is like the detrimental impact of trying to learn statistics before mathematics. Yet, in the world of talent, we see limited applications of building capabilities on strengthening logic and an overwhelming emphasis on data-skills.

There are major reasons for prioritizing development of logic over development of data-skills.

McNamara effect refers to the fallacy where the leader considered body count as the measure of success in a war, ignoring other relevant insights like the shifting moods and feelings arising from the war, where people were largely ignored. When analyzing complex phenomena, logic surpasses usage of any metrics.

Logic helps us avoid spurious correlations. Since logic is a more intrinsic process of cognition, a bias in the logic results in selectively creating, reading and search of data to validate the bias in the logic. Valid cognition as a foundation of logic has early reference in Pali and Sanskrit texts with its emphasis on six ‘Pramans’ – through which logical skills can be developed through a combination of Inference, Comparison, Postulation, Non-apprehension, Verbal Testimony and Perception. An individual with strong discipline on Comparison and Inference, would therefore tend to compare multiple aspects of a data point, rather than be led by algorithmic nudges through data.

To start with, a simple validation exercise: A student was asked to use the number 2 two times to get the number 4. From a data perspective, the number 4 can be obtained through both multiplying and summing the 2’s, and that is ‘verified’ to be the right result. The definition of logic is about reasoning?conducted or assessed according to strict principles of validity, and therefore it is important to go beyond the verified data, to understand the validity of it – Did the number 4 arise out of summing or multiplying the 2 numbers.

The field of Human Resources is possibly the function with the largest ‘big data’ challenge (much to the disagreement of my friends from Retail, Marketing and Healthcare), and therefore a functional example from my favorite field of work, demonstrating a challenge with data.

1400 men and women employees were provided with 3 objectives, each having a different complexity level. 15%, 52% and 33% of the men could successfully achieve objectives 1, 2 and 3, respectively. A higher percentage of women (16%, 53% and 35% on objectives 1, 2 and 3 respectively) could successfully achieve the individual objectives, vis-à-vis the men. However, the year’s performance appraisal which had these three objectives, shows a higher percentage of men achievers overall.

?

?The above example is Simpson’s Paradox, illustrating how it is possible to draw two different (often opposing) conclusions based on the same data, depending on how we treat/process the data. Data cannot therefore help reach the right decision point, unless we go beyond it to look at the causal effects, which require application of logic.

A startup business introduces 10-minute learning app, claiming we can learn a substantial deal of any subject in 10 minutes, to be able to start a conversation. Despite knowing how impossible that is, we still register for that app, which takes massive volumes of our data and subsequently merchandizes it. Though the app has fundamentally no sustainable impact or utility, it receives massive funding not because of its utility (of which there is none), but because it is a tool to gather data, that is used for some other purpose and cross-selling. If we had applied logic that 10 minutes is a long time for Instagram, but not deep learning, it would have stopped us from registering for that app. The bigger issue is, what happens to employees in that startup, when investors, unsurprisingly, do not see the app’s long-term utility?

Focusing on data and incentivizing for it, has major repercussions. To reduce deaths arising from snakebites, the Government offered a financial incentive for every snakeskin brought to them to motivate snake-hunting. But instead, people started snake-farms to get that incentive. The incentive was removed on realizing the harm, resulting in the people releasing the snakes, further increasing the threat – an old example on how focus on data without logic preceding it, is a derailer.

There are simple methodologies to approach building logic first prior to focusing on data and its processing.

1. Leverage on fundamental logic-enhancing subjects early on (to start with, Arithmetic, rather than Statistics)
2. Create an environment and platforms that enhance and assess inductive and deductive logical reasoning.
3. Increase exposure to synoptic reading, expanding from exploratory and analytical reading.
4. Invest in activities that cover all the seven forms of Intelligence, going beyond just analytical intelligence.

Both logical reasoning and data skills contribute to effective decision-making, and finding the right equilibrium is crucial.

Open for views on this, and how to create analogical thinking, leveraging data intelligence – its impact on education, commerce and the way we approach a problem.