This preview shows pages 1–9. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Models of Choice Reaction Time & Motor Skills Brian Bailey Announcements Homework1 assigned Reserve last 10 min to form teams Group information forms due next week Models in HCI Simplified abstractions of reality Explain or predict observed behavior accepted model is best explanation to date may be modified or rejected difficult or impossible to fully verify Inputs, outputs, and assumptions Hicks Law  Choice Reaction Time Models reaction time under uncertainty Decision time T increases with uncertainty about the judgment or decision to be made T = a + k*H, where H is the entropy of the decision; a and k are constants. H = log 2 (N + 1), if probabilities are equal H= = + 1 2 ) 1 / 1 ( log i i i p p Menu Design Use a broad shallow or a narrow deep structure? Derive from Hicks Law T ~= a + k*log 2 N N = total items to choose from a and k are coefficients of the regression line b = menu branching factor (items per menu) If we assume that a=0, k=1, then T ~= (menus to choose from) * (time per menu) T ~= (log b N) * (log 2 b) T ~= log 2 N Is This True? Suppose 64 menu items (N = 64) Case b = 4: T = (log 4 64)*(log 2 4) = 6 Case b = 8: T = (log 8 64)*(log 2 8) = 6 Make either fewer choices in more submenus or more choices in fewer submenus But constants not typically a=0 and k=1, thus T ~= (log b N) * (a + k*log 2 b) T ~= a*log b N + k*log b N*log 2 b T ~= a*log b N + k*log 2 N T ~= k*log 2 N + a*log b N Use Broad Shallow...
View Full
Document
 Fall '08
 Karahalios,K

Click to edit the document details