Chapter #7 Modeling Interaction Flashcards
Descriptive Model
reduction or partition of a problem space
Examples: Politics, GroupWare, Keyboards, Two Handed input, Circumplex model of affect, Graphical input
CSCW
Also know as Groupware
Name (computer supported cooperative work)
Def: People working collaboratively with computer technology
Quadrant Model of Groupware
A descriptive model
2x2 space
Same time, Different time
Same place | | |
Different place | | |
Critiquing the Model, Groupware
Lots of new ways to collaborate through technology didn’t exist when this model came out, hence some new forms of collaboration don’t fit in just one category.
KAM
Key Action Model
A descriptive model
Symbol Keys: Produce graphical symbols, ex A
Executive Keys: Does an execution, ex Esc
Modifier Keys: Modifies the affects of other keys, Ex Crtl
Critiquing the Model, KAM
Some keys seem like they belong in multiple categories.
Also the right hand is noted to be super busy with lots of executive keys
Study of hand usage is called
laterality or bimanual control
Guiard’s Model of Bimanual Control
Descriptive model
Non-preferred hand: does corese movements, sets frame and leads the preferred hand.
Preferred hand: does fine movements, works within the frame and follows the non-preferred hand.
Critiquing the Model, Guiard’s Model
Developed in phycology and did not do testing with computers
What’s the argument for where scrolling should be?
Non preferred hand since the preferred hand, normally the right, tends to be overloaded with the mouse and right side of the keyboard
CMA
Circumplex Model of Affect
2D descriptive model of human effect or emotion
Horizontal axis: pleasure vs displeasure
Vertical axis: arousal vs sleep
high arousal displeasure + pleasure
sleep
Applications of CMA
How expressive lighting in a robot can express emotion
emotion through shape-changing interfaces
emotional state in play environments
etc
SAM
Self Assessment Manikin
Follows the axis from CMA, with 2x9
such that it has a manakin conveying each emotional level
Three State Model of Graphical Input
Developed by Buxton
0: out of range (ie mouse off desk)
1: tracking; mouse moving along desk
2: dragging; mouse moving while holding down click
Ex: new touchpads and that invention of the tactile touchpad mouse
Who also had a three state model but for the light pen?
Newman
Predictive model
Uses predictive values to predict the outcome of dependent variables
Exact definition: Predicts the outcome on a criterion variable (aka
dependent variable or human response) based on the value
of one or more predictor variables
uses numbers and is an equation
Linear Prediction Equation
basically y = mx+b
expresses a linear relationship between a predictor variable (x) and a criterion variable (y)
used in linear regression
Fitts law, 3 applications
Predictive equation for determining alternatives
Throughput as a dependent variable
Check if conforms to Fitts law
Fitts Law, Task Paradigms
Serial Task: Ex hit one thing on one side then the other, and go back and forth
Discrete Task: Two things to hit, one on each side, a light will light up on one side at a time and you have to hit that side
Fitts Index of Difficulty
ID log_2(A/W + 1), unit is bits
Fitts’ Index of Performance
TP = IDe/MT
Throughput and the Speed-Accuracy Tradeoff
TP = (log_2(Ae/4.1333*SDx + 1)) / MT
Choice Reaction Time
n stimuli with n ways to respond to each
Ex fingers to each light
Equation: RT = a + blog_2(n+1)
KLM
Keystroke Level Model
predicts error free task completion times
t_execute = t_k + t_p + t_h + t_ d + t_m + t_r
K keystroking P pointing H homing
– D drawing M mental prep R system response
Expert Behaviour in KLM
where someone who is experienced doesn’t have to think before doing the action. Ex knowing what buttons to press to quick write beep
Involves the M and Mp for mental processing between actions like k, p, h,d
Two approaches for KLM modeling with M_P
All-In: Include Mp at every reasonable juncture (novice user)
All-Out: remove all Mp (expert)
Mental Operator for Visual Search, Mv
Mv, mental time it takes to visual search.
Primarily in things where you enter a letter and it gives word predictions. Hence taking time to read suggestions is Mv. But overall it should save time
Skill Acquisition
We start as novices with poor performance, over time with practice we gain skills and preform better and potentially become experts.
Dependent Variable: proficiency
Independent: Amount of Practice
Power Law of Learning
Relationship between practice and proficiency is non-linear
Tn = T1 * n^a , Time specific one, curves left top to down so like slope
Sn = S1 * n^a , Speed specific one, curves left boom to top so like x
Log-Log Model
In power law of learning
In a graph, whether speed or time variation, the x and y are put into log scales then the relationship between data is linear
More than one predictor
Essentially allows multiple predictor values in an equation
ex y = b1x1 + b2x2 + ….
Also known as multiple regression
Stepwise Linear Regression
Variables are added one at a time to determine which has the highest r (variability), then tested in that manner
MCM
Model Continuum Model
Graph that shows the differences between descriptive and predictive models
Analogy -> categories / Design spaces -> Stats -> Equations
