As with scientific research, it’s hard to distinguish correlation from causation
Establishing links between cause and effect is a core function of scientific research, and one that we now know is far more difficult than previously thought. In the past decade, concerted efforts failed to replicate hundreds of studies published in prestigious journals, particularly in medicine and social sciences. This “replication crisis” has led to a flood of new research methods aimed at weeding out hidden factors that lead scientists to mistake mere correlations for causation, for actions that actually lead to change.
Businesses developing new strategies or just making everyday decisions have a similar problem, UCLA Anderson’s Olav Sorenson suggests in a phone interview. To manage effectively, corporate leaders must understand how their policies, orders and actions might change overall profits, customer service or another metric for the organization as a whole, amid a world of ever-changing variables — often unseen — that could also be driving change.
Often, managers lack some relevant piece of information required to accurately judge the ultimate outcome.
Consider a manager who sees the factory holding a large amount of one particular part, Sorenson says. In an attempt to run a leaner operation, she orders employees to reduce inventory. But the manager did not know that the employee in charge of that part intentionally maintained more inventory because the supplier regularly delivers late. Rather than improving the company’s bottom line, the inventory reduction leads to the loss of a key customer that didn’t get an order on time.
What Management Misses in Workplace Can Lead to Mistakes
A working paper by Sorenson and University of Toronto’s Michael D. Ryall suggests that common, real-world factors that management is unaware of — a particular employee’s penchant for altruism or their bad day in traffic, for example — can confound its ability to predict the end results of its decisions. This complexity leaves managers to wonder why things aren’t changing — or why they are — and what actually is causing the change.
The study model demonstrates this disruption to causal inference can occur even in best-case scenarios, in which employees obediently follow directives, and managers can see some immediate results, like a decline in storage costs that only later becomes a lost order. The factors that can ultimately affect outcomes are so pervasive, varied and, oftentimes, personal that managers can’t possibly know everything they should take into account.
How Organizations Can Mitigate Hidden Factors
Yet organizations effectively manage every day. How? Sorenson and Ryall note several strategies that help prevent hidden factors from wrecking well laid plans.
- Experimenting with a change on a small scale before enacting an edict, as opposed to plunging ahead, allows management to identify, and root out the cause of, unintended consequences, the study notes.
- Stronger protocols for keeping important information in the open, like suppliers’ delivery records, also facilitate accurate predictions about outcomes.
- A third solution, and the one Sorenson calls the most interesting, involves creating a corporate culture that allows the manager “to substitute this (hidden) information with things the manager is aware of,” he says.
For example, a workforce that makes joint commitments to certain values — perhaps “the customer is always right,” or “share your work issues” — is more predictable even if individual workers don’t personally subscribe to the value. Hiring people with similar values and attitudes also makes it easier for managers to infer private information that may be pertinent.
Sorenson points out that scrutiny over research findings, and a subsequent explosion of new methodologies, has led scientists in other fields “to think very carefully about whether or not they can say that two things are causally related or just correlated.”
Management research, on the other hand, typically starts with the assumption that leaders understand how their orders, and the subsequent efforts by those they manage, will affect organizational goals. The findings here suggest that the link between cause and effect is seldom so clear.