Scaling Cyberspace. The Idea & Our Stimuli Scaling websites: how do people perceive similarities between popular websites? What are the criteria? – Stimulus.

Scaling Cyberspace

The Idea & Our StimuliThe Idea & Our StimuliThe Idea & Our StimuliThe Idea & Our Stimuli

• Scaling websites: how do people perceive similarities between popular websites? What are the criteria?

– Stimulus set: 10 popular websites

HypothesesHypothesesHypothesesHypotheses

When rating popular websites on a similarity scale, people will…

…use credibility of the website as one of the bases for their judgments. (hypothesis 1)

…use purpose of the websites (ranging from ‘to inform only’ at one end to ‘to socialize only’ at the other) as one of the bases for their judgments.(hypothesis 2)

…consider how much of the information/material is user-contributed as a basis for their judgments (hypothesis 3).

Expected FindingsExpected Findings• A metric space with two or possibly three

dimensions (credibility, purpose of the website and user contribution)

• A tree structure may emerge where websites will

group on the basis on any one or more of the above dimensions

MeasurementMeasurementMeasurementMeasurement

• Subjects’ perceived similarity• Subjective dimensions:

–Fun factor–Familiarity–Credibility of the information–Purpose: from “to socialize only” to “inform only”–Friendliness of user interface–Variety of functions–Information Contributor: how much is provided by users

• Objective dimensions:–Traffic rank–Visiting Speed–The number of pages that link it

• Rating similarity of each pair• Subjects are given randomly ordered pairs 10c2 =45 pairs• Scale 1-21

• Rating similarity of each pair• Subjects are given randomly ordered pairs 10c2 =45 pairs• Scale 1-21

• Scale1 － 10

• Scale1 － 10

Alexa.comAlexa.com

Did not correlate with any variable we measured

? Less commercial purpose

Less user’s contribution

Less social function

Increasing credibility

Less fun/ more seriousness

Running MDSRunning MDS

Correlations

Running correlations between each dimension and the variables

fun cred var cont pur fam linked rank speedPearson Correlation

-0.722 0.898 0.158 -0.902 0.926 0.273 0.208 0.300 -0.162

Sig. (2-tailed) 0.018 0.000 0.663 0.000 0.000 0.446 0.564 0.399 0.655N 10 10 10 10 10 10 10 10 10Pearson Correlation

-0.204 0.037 -0.365 0.102 -0.137 -0.342 -0.323 0.089 -0.206

Sig. (2-tailed) 0.571 0.920 0.300 0.779 0.707 0.334 0.363 0.807 0.569N 10 10 10 10 10 10 10 10 10

Correlations MDS1

MDS2

Running INDSCALRunning INDSCAL Running INDSCALRunning INDSCAL

Stress = 0.2587

Running INDSCAL

Running AddTreeRunning AddTree Running AddTreeRunning AddTree

Superimposing Addtree onto MDS

Addtree stress = 0.0642

MDS stress = 0.10871