Floating Eyeball

Factorials
A factorial has more than one independent variable.
2 Independent variables (gender, education level) with 2 levels each (male, female) and (elementary, secondary). This creates 4 conditions. (Grid)

Mixed Factorial
At least 1 between-subjects IV and 1 within-subjects IV
Mixed design must deal with equivalent groups problems (between) and sequence effects problems (within).

After you collect your data
Check to make sure your statistical assumptions were met
Do you have normal distribution? (bell curve, skewness and curtosis)
Do you have homogeneity of variance? (Have one group vary alot, have another group not)
Are standard deviations similar across conditions?
If assumptions weren’t met, you have to use non-parametric statistics.

P Values
Significance (sig if over threshold, not sig if under)
95% confidence for social science (p , .05%)
The probability that chance will account for the difference less than 5% of the time.
P values are not about effect size. It is simply a threshold question.

Effect Size
Magnitude of difference
P value is a threshold
Effect size measure how big the effect is
For ANOVAs use Partial nsquared (partial eta squared)
“Proportion of total variation attributable to the factor, partialling out (excluding) other factors from the total non-error variation.”

Randomization is a shorthand way to make groups equivalent.
The better the researcher understands the variables that “make a difference,”  randomization could be less valid rout to equivalence.

Stereoptype Validity Threat– Collect demographic data last.

Participant Effects
Demand characteristic means they act differently because they are in a study.
Hawthorne effect means they change behavior to be helpful to the researcher.

Experimenter Effects
Confirmation bias looks for confirmatory evidence.

Validity
Internal validity rules out other explanations of the results.
External validity relates to generalizability beyond the sample.
Ecological validity did the environment under which research is conducted generalize to other environments.

Lori’s domain is anything a user can see or touch.
“Positive information experience for customers.”
Drive faster adoption of IBM products…your job is not to be an artist, its to meet the deadlines and sell the products.
“Right content, right person, right time.”
“Proactive, embedded assistance in a progressive disclosure model. Providing less (at first) and then letting the user ask for more. If the user asks for help.”

Information developers (formerly known as technical communicators)  that work on the part of the product that communicates.

Writing for one context is “easy.” Writing simultaneously for audiences, formats, etc. is the new challenges.

Information model
Element model

Content Management Systems

Alternatives to the SEM…less radical alternatives than STOP and info mapping.

Modules, spreads, pouring of texts…

Weak modularity. Requires some contextual understanding.
Strong modularity. Standalone modularity.

Robert Horn looked at the paragraph.
Tracy addresses the whole document.

STOP: 2 page spread (text, graphics)
If the format has to be a 2 page spread what do you do when you have more info? One example is fold outs. The foldouts, when converted to pdf, were lost. Supposedly digital loses the constraints, but the foldouts weren’t scanned!

Graphics Theory Category

• Verbal Component
• Pictorial
• Schematic
• Composite

Components

• Signal icon
• Title (may be descriptor)
• Descriptor/Sub-descriptor
• Tag line
• Rating Element (comparative system)
• Scale
• Interpretive aid
• Source
• Other

Reader Reaction

• Pre-attentive
• Interpretive
• Response

Type of Label

• Endorsement
• Comparative. (Symbol, rating scale)
• Information Only
• Hybrid

Decision infographics…any recommender system will fit into the patterns. “Consumer” in the broadest sense of the term.

How do you design for “readers” as culture and consciousness shifts? Media space rapidly changing. Digital, multimodal.
Industrial age: producers + consumers. Now: everybody is a producer.

Books of the print era need to be fat enough so that the title appears on the side.