ISSS608 2016-17 T1 Assign2 Meenakshi

Abstract

The new open collaborative environment and open education content has had a great impact on student learning and teaching practices in the universities across globe. Wikipedia stands out to be the primary technology platform that is enabling learning and knowledge sharing. In this study we are trying to understand the teaching use of wikipedia from a dataSurvey of faculty members from two Spanish universities.

Problem and motivation

The objective is to analyse the multivariate survey data using appropriate visualisation tools and techniques.-
The results would enable us understand the main attributes that influence the teaching uses of wikipedia.
This will give us insights on how the open learning platform is being received and how are the practices going to be in the academic world in the future.

Tools Used

Tableau version 10.0
JMP

Approach

Following the recommended Visual Analytics Application Design Process, we approach the problem is 3 steps

Identify theme of interest

The focus is on Visualizing factors that influence the use of Wikipedia as a platform for teaching and learning in the academic world. What do faculty members consider important in order to use and promote wikipedia.

Define questions for investigation

The wiki4HE survey data, contains information about the faculty members basic profile. Survey questions are defined based on these factors:
From the survey question under each of these categories we can understand the meaning of these attributes

1. Perceived Usefulness and Ease of Use
   
2. Perceived Enjoyment
   
3. Quality
   
4. Visibility
   
5. SocialImage
   
6. Sharing Attitude
   
7. Use Behavior
   
8. Profile 2.0
   
9. Job relevance
   
10. Use Behavior
    
11. Behavioral Intention
    
12. Incentives
    
13. Experience
    

The associations to be studied

1. Who are the poeple taking the survey? Some demogrpahics on the survey participants
2. Do faculty members from various departments differ in their opinion on the perceived usefulness and Quality.
3. Is Quality the key factor considered by most faculty members. Does it influence the perceived usefulness and ease of use
   
4. Is the Use behavior and Behavioral intention defined by quality
   
5. Does the faculty Social profile influence the sharing attitude
   
6. How important is the Job relevance for the use of wikipedia. Does it also affect the sharing attitude
   
7. Does the Social Image- colleague's perception, influence the perceived usefulness and behavior intention
   
8. How does visibility affect the use behavior and behavioral intention
   

As we progress with data explorations, some questions may have to be re defined.

Analysis

Data Preparation and cleaning

• The JMP tool is used for this task. The wiki4HE csv file is loaded in JMP to check the data types and get a preview/summary of the data.
  

• 

• The data types for some of the variables needs to be redefined. Such as - Changing Gender and Phd to categorical data.
  
• The raw data has some missing values and '?' mark fields in the table. Referring to the data dictionary, appropriate values are recoded using JMP. Through this process we can make sense of some of the '?' fields. Recoding also aids interpreting data during analysis.The steps are summarized in table below-
  

• 

• 

• Add a ID column to represent individual survey respondenst with a unique ID wiki4HE_Clean xls file.
• Add a category column to the individual questions in a new sheet in the the wiki4HE_Clean xls file.

• 

• The data table in JMP is now ready to check for missing values. Check the distribution of all parameters to understand more.
  

It is seen that the response is missing for some of the questions, however we consider the %Total of responses for each scale during our analysis. Hence these missing values can be ignored in the table.

• 

• 

• 

• The cleaned data from JMP is saved as csv/xls file types to be ready to use for the visual analysis.
• Visual analysis - Import the cleaned data to tableau to analyse survey data.
  

The Qns category sheet and the wiki4HE_Clean xls are joined by QuestionID and imported to Tableau. We need the Question Ids to be rows instead of columns to visualise the Likert scale data. The Pivot function in tableau achieves this.

• 

Techniques

In this assignment we intend to visualize the survey data, where Likert scale questions are used. A quantitative value is assigned to the opinions expressed by respondents on a 5 point scale from Strongly agree = 5 to neutral = 3 to strongly disagree = 1.
We need to choose a technique which would best express the overall response across various categories and be able to read the association between parameters. Parallel co-ordinates would be more useful when we have high dimensional data.

• Stacked bars with some measure value indicator would help us compare response between parameters and hence some inference on the association can be made
• The Divergent bar charts, can best express the sentiment of the survey respondents as a measure towards positive and negative sentiment.

Hence above two techniques are used in my approach.

Visual analysis journey and Results

The Published tableau workbook is in the link below.
Wikipedia dashboard
DivergentBar Charts
Updated - Visualize Wikipedia survey data
This is an interactive dashboard with filters and useful to explore the answers we are looking for.

Who are the survey respondents

Most faculty members are Adjunct faculty and are not registered users in wikipedia. The survey had almost equal number of PhD and non-PhD holders

Quality

• Quality of content influences the perceived usefulness and enjoyment more than its influence on ease of use. We see that Quality is an important factor determining the usefulness.
• Social image - It is interesting to note that colleague's perception of quality is influencing how quality is being perceived by faculty.

Sharing attitude

• In the initial investigation, the assumption that social profile influences the sharing attitude is not necessarily true.However the job relevance and the university considering to promote the use of open collaborative environment seems to be influencing the sharing attitude.

This can be visualized as shown

• 

Since different color bars have different widths for profile and sharing attitude, we infer that they are not well associated or influencing one another.

Use behavior and intention

• The usage of wikipedia by faculty in the future, does not seem to be affected by their sharing attitude. But more influenced by perceived usefulness, Social Image and Quality.
• The current usage of wikipedia for teaching and student learning is not significantly influenced by quality as expected in the initial investigation.
  

The "association of parameters" interface in the dashboard is useful to learn the relationship between other factors. The influence of certain factors on the use of wikipedia and the possible answers for faculty practices can be visualized with this dash board.

Visualize sentiment

With the trellis plot in JMP we can see the responses in all, However it is not very useful for drawing clear inferences on the sentiment unlike divergent bar charts.

• 

  JMP trellis plot

As mentioned the Divergent bar charts are a good option here. For instance, we see that most respondents' opinions lie between neutral to agree when it comes to quality of content.

• 

  Quality sentiment

Majority faculty agrees that it is important to share content on open platforms and that students be familiar with such platforms.

• 

  Sharing attitude

We can see a positive attitude when it comes to future use of wikipedia in the academic world.

• 

  Future intention to use.

References

Data Source
www.datarevelations.com

Citation

Meseguer, A., Aibar, E., Lladós, J., Minguillón, J., Lerga, M. (2015). “Factors that influence the teaching use of Wikipedia in Higher Education”. JASIST, Journal of the Association for Information Science and Technology. ISSN: 2330-1635. doi: 10.1002/asi.23488.