Confirmation bias is probably the most pervasive fallacy. We are apt to agree and notice results that confirm what we already believe. As a result, we come away thinking that belief is further supported. ![[city-1265055_1280.jpg]] *Credit: https://pixabay.com/photos/big-city-crowded-taxi-yellow-1265055/* The Taxi Fallacy is a kind of confirmation and availability bias that suggests despite complete availability of data across an entire population, salient groupings increase the likelihood that we will incorrectly make claims of the groups disproportionately from the other members of their population. ## How the fallacy works The idea goes that you have a city where half of all drivers are reckless and irresponsible while the other half is cautious and responsible. The city starts a taxi service and randomly chooses among the population to assign as drivers. ![[Pie chart - Frame 1.jpg]] *Individuals are chosen from the general population at random to drive a taxi.* The fallacy asserts the public will incorrectly accuse the taxi drivers of being disproportionately dangerous compared to non-taxi drivers because of salience. That is, among the non-taxi drivers the color, make, model and any other salient attribute is mixed, while the appearance of the taxi drivers is homogenous: Yellow sedans with Taxi labelling. This is a type of confirmation and availability bias because we are apt to look for a pattern to explain the phenomenon. Once the pattern of the yellow taxi is conjectured, it will more easily reinforce itself each time a taxi driver exhibits reckless behavior. Meanwhile, any pattern conjectured from the non-taxi population (e.g. "red pickup trucks") will not reinforce nearly as often. Yet we know it is true that a taxi driver in this city is no more likely to drive recklessly than any other driver when chosen at random. Similarly, if in a violent mob, some individuals are chosen at random to wear bright lime green hats, the public will assume a disproportionate amount of violent behavior to those hat-wearers versus any other member of the mob wearing heterogenous clothing. ## Real life examples The argument that supplement consumption by vegans is proof that the vegan diet inherently lacks nutrients is a taxi fallacy. Any validity of the claim aside, the construction of the argument itself is still fallacious. The majority of customers in the supplement industry are non vegan by a large margin. Yet those non-vegan diets don't have the salience of a consistent identity label for all members. We don't see their presence as a unified group in the same way we don't see the heterogeneity of non-taxi vehicles as a group. As a result, rarely will anyone use supplement consumption as evidence to suggest non vegan diets inherently lack nutrients. ![[Pie chart.jpg]] *While not necessarily reflective of actual figures, this visually demonstrates the salience factor of a minority of a group in its population.* This may further be bolstered by the phenomenon of the spurious variable. Supplement intake isn't necessarily a measure of need, rather of interest. It can also be said that health-conscious individuals are more likely to seek supplements and that any one of a named diet is more likely to be health conscious than someone who does not identify with a named diet.