Spuriousness is the incorrect inference of a causal relationship between two variables where the relationship is in reality only accidental.
Researchers attempt to identify or eliminate spuriousness by the use of random assignment in an experimental design or through the use of control (extraneous) variables in the manipulation of data during analysis.
A more serious concern is whether the observed relationship is spurious. Despite some statistical attempts to control for clear sources of spuriousness, there are potentially an infinite number of variables that are unaccounted for. For example, suppose that power reputation is an attribution based on the fact that certain people are closer to the action in the organization. Suppose that being closer to the action also gives people certain advantages in knowing the social network. Then one could argue that the observed relationship between reputational power and network knowledge is spurious. In part, one could also argue that being "closer to the action" is already controlled for by controlling for centrality in the network; but then again, it may not control for all of it. I have controlled for what I argued to be the most reasonable sources of spuriousness. But, clearly, one cannot conclude that all sources of spuriousness have been eliminated. - Extract from: Assessing the political landscape: structure, cognition, and power in organizations - Administrative Science Quarterly, June, 1990 by David Krackhardt
Conceptual Meaning and Spuriousness in Ratio
Correlations: The Case of Restaurant Tipping