POLS 3650 Lecture Notes - Lecture 19: Frequency Distribution, Null Hypothesis, Statistical Significance

• School: University of Guelph
• Department: Political Science
• Course: POLS 3650
• Professor: Edward Koning

Document Summary

Relational prediction: what is our best guess of the value on the dv if we knew the value on the iv. The better the relational prediction compared to the marginal prediction, the more strongly the variables are associated. The main appeal of pre statistics is that they are easy to interpret: range from 0 to 1, values show proportionate reduction in error. Lambda is a pre statistic and is much easier to interpret. Follows the standard pre formula: = (cid:3045)(cid:3032)(cid:3031)(cid:3048)(cid:3030)(cid:3047)(cid:3042)(cid:3041) (cid:3041) (cid:3032)(cid:3045)(cid:3045)(cid:3042)(cid:3045) (cid:3033) (cid:3047) (cid:3032) (cid:3049)(cid:3028)(cid:3039)(cid:3048)(cid:3032) (cid:3042)(cid:3041) (cid:3046) (cid:3038)(cid:3041)(cid:3042)(cid:3050)(cid:3041) (cid:3041)(cid:3048)(cid:3040)(cid:3029)(cid:3032)(cid:3045) (cid:3042)(cid:3033) (cid:3042)(cid:3045)(cid:3034)(cid:3041)(cid:3028)(cid:3039) (cid:3032)(cid:3045)(cid:3045)(cid:3042)(cid:3045)(cid:3046) Our marginal prediction is the modal value on the dv. Our relational prediction is the modal value on the dv for each separate value on the iv. Lambda would show us how much better we would get at predicting someone"s voting behaviour if we knew their religion. Example: one with just a frequency distribution and one with a cross-stat.