Saturday, October 31, 2009

Constant Improvement - Analytical Reasoning

As part of refining the Zen of 180 system into a full-fledged online LSAT diagnostic tool, I've been trying to apply Teach For America's philosophy of constant improvement to the Zen task standards, especially in regard to the reading comprehension and analytical reasoning--known colloquially as logic games--sections.

Last week we talked about how Zen of 180 considers passage structure more important than content; today we'll cover a more comprehensive system for discussing logic games.
  1. Most LSAT test-prep companies separate logic games into three broad categories: sequencing, grouping, and matching/assignment.
Granted, some tutors try to provide some more context with subgroups, such as linear sequencing versus "tree sequencing." However, these distinctions are so broad as to not carry meaning.

Imagine if the only word we had for "dog" was still dog, but that term also meant every four-legged animal with fur that was shorter than an adult human. Calling an animal a "dog" might mean that it is man's best friend, or that it's a dangerous beast with razor fangs and monstrous appetite, say a tiger.

Obviously, your approach to each animal would be completely different: pet versus hide. Yet, LSAT test-prep companies don't make the distinction between a grouping game that requires a completely different approach from another "grouping" game.

Because the whole concept of Zen task standards is performance driven, we're hoping to categorize games by their diagrams. We hope to codify the tasks that test-takers should be performing during the logic games section: creating a specific diagram for a subset of rule-interactions and then manipulating information within that diagram.
  1. Most LSAT test-prep companies acknowledge the complexity of logic games by designated them as hybrids of the sequencing, grouping, and assignment types.
The problem is that almost every single modern LSAT logic game is hybridized: what good is a system of classification where the basic unit is never seen in isolation? Calling a given game a hybrid sequencing/grouping has very little bearing on the test-taker; what matters is how to set up the diagram and how many points are devoted to a given task.

Thus, the classification system for analytical reasoning will probably look like this, using the example of June 2007, section 1, questions 18-23:

Matrix Assignment

Rules 2, Conditional 4

As in, the LSAT is asking you to draw a matrix diagram to model the rules, and 2 questions deal with elements of the game, while 4 hinge on conditionals presented by individual questions. Thus, the test-taker has a solid understanding of how to represent the information and which questions apply to which task.