One does not simply crowdsource the Semantic Web: 10 years with people, URIs,...
Simple Program for Enhancing Quality in Discussion Boards
1. SPEQ-DB: Simple Program for Enhancing Quality in
Discussion Boards
Sara Colon Cerezo***, Nicole Dubin*, Rafael Hernandez**, Dr. Brian Thoms*
*California State University, Channel Islands, **Oxnard College, ***Santa Barbara City College
PROJECT BACKGROUND
Under the field of human computer interaction, the subfield of captology guides
how technology can influence behavior [1]. Through an analysis of previous
online conversations, this research improves upon the design of an existing online
discussion board to incorporate a simple algorithm to facilitate on-topic and
readable discussion posts.
In total, 1,629 conversations were mined for readability and keyword density.
Readability was accessed using Readability Metrics, an open-source application
programming interface for managing texts and their readability scores [2]. Key-
word density was calculated as a ratio of total keywords found over total words
posted minus all stop-words.
Our analysis found that while readability increased from originations to responses,
there was a 10% decrease in response readability (Figure1) and a 13%
decrease in response keyword density (Figure2). Additionally, there was the
tendency for users to move away from topics as discussions aged, which is evident
by the downward slope of Figure2.
SYSTEM ARCHITECTURE
USER INTERFACE DESIGN
SYSTEM CODE
RESEARCH METHODOLOGY
[1] Fogg, B., & Nass, C. (1997). “Silicon sycophants: The effects of computers
that flatter,” Int’l Journal of Human Computer Studies, 46(5).
[2] Ipeirotis, P. (2012). “Readability Metrics API,” Mashape. Accessed June 19,
2016 from https://market.mashape.com/ipeirotis/readability-metrics.
[3] Hansen, D., Shneiderman, B. and Smith, M. (2011). Analyzing Social Media
Networks with NodeXL: Insights from a Connected World, Burlington: Morgan
Kaufmann.
[4] Simon, H. (1996). The Sciences of the Artificial Third Edition, Cambridge,
MA : MIT Press.
[5] Thoms, B., Eryilmaz, E. (2014). “How Media Choice Affects Learner
Interactions in Distance Learning Classes,” Computers & Education, v75, pp.
112-126
ACKNOWLEDGMENTS
• Funding provided by the Title V, US Department of Education Grant
• Project ACCESO Summer Research Institute
REFERENCES
In design science research, researchers are concerned with the way things
ought to be in order to attain goals and they construct artifacts as a way of
achieving these goals [4]. Building atop [5], this research asks the following
research questions:
q R1: To what extent will S.P.E.Q. DB enhance the quality of both origination
and response posts in online conversations?
q R2: To what extent will S.P.E.Q. DB increase levels of network density
within the online community?
Figure 4 represents S.P.E.Q. DB. The primary goals of the new design are to
influence higher quality interactions and facilitate a more cohesive social network
by providing a responsive way for users to judge the quality of their posts. Users
can view the quality report by clicking the Analyze Button below the textbox.
Design improvements include:
Ø Group QI: Calculated using the average for all individual posts for that
discussion.
Ø Individual QI: Calculated using the new QI formula.
Ø Quality Gauge: Compares the individual QI against the group’s average QI.
Ø User Pins: Allows users to keep track of posts they are interested in.
SNA graphs were constructed in NodeXL, which is an open source extension for
MS Excel that provides a range of basic network analytics and visualization
features [3]. Each node represents a user, and each edge (i.e. line between two
nodes) is an interaction between users (i.e. responding to a post).
Summary: Most users who fall within an acceptable level of quality tend to be more
central to the network, while users with lower quality scores tend to be situated on
the outskirts of the social network. This finding suggests that as the quality of user
posts increases, the number of responses that user receives also increases, thus
increasing the density of the social network.
q Blue Disks represent users whose QI
falls more than one standard
deviation above the mean QI.
q Green Squares represent users
whose QI falls within one standard
deviation of the mean QI.
q Red Triangles represent users
whose QI falls more than one
standard deviation below the mean
QI.
q Node sizes represent the sum of
responses a user received per post.
q Labels indicate the NodeID and QI.
Identifying 1) topic focus and 2) readability as important factors that influence the
flow of online conversations, a simple formula for determining the quality index (QI)
was constructed resulting in even higher disparities between originations and
responses (Figure3).
Figure 4. User Interface
S.P.E.Q. Discussion Board
Step 1: Calculates readability and key word density for users posts.
Step 3: Calculates post level QI.
Step 2: Calculates thread level readability and key word density values.