Content Analysis of Interactive Media Paul Skalski Cleveland State University Background Since the writing and publication of the Content Analysis Guidebook, there has been increased interest in interactive media content, particularly: Video games! And… Web 2.0 sites or User Generated Media (UGM). Sidebar: The Web 2.0/UGM Has exploded in popularity in the past 5 years. What are prominent examples? Facebook MySpace YouTube Wikipedia What do these have in common??? 4 of the 11 MOST visited websites! Four Considerations Key issues for the content analysis of interactive media include: 1. Creating content 2. Searching for content 3. Archiving content 4. Coding/analyzing content 1. Creating Content The fundamental difference between old media and newer interactive media. Users are in charge of much of what content looks like, with some restrictions. Web 2.0 users can create content within the templates provided by sites. Video game players have (some) control over what happens in a game, affecting the content. Specific Web 2.0 Content Issues User Generated Media (UGM) vs. User Collected Media (UCM)—the latter refers to activities such as posting videos from TV on YouTube. Also: The templates sites provide may change over time, necessitating FLUID codebooks to match fluid content. Specific V.G. Content Issues Smith (2006) identifies the following: 1. Player Skill 2. Time Frames Depending on skill, players may play in different ways, results in different content Whole games cannot be sampled like TV shows or movies. 3. Character Choice Players have increasing control over their characters. 2. Searching How do you select content for inclusion in a sample? With games, similar procedures can be used that have been used in TV and movie content analyses—e.g., selecting the most successful titles. With Web 2.0 sites, there is greater difficulty due to (potentially) millions of equal sampling units. 3. Archiving How do you store units for analysis? With games, typical procedure has been to record games as players play and store content on DVD (though DVR options now) With Web 2.0 pages, options include: Print screen Saving the file Creating PDFs Also software options for video/audio 4. Coding Analyzing the archived content. Includes: 1. Identifying units of analysis (e.g., individual user posts, game characters) 2. Creating a codebook 3. Creating coding sheets (may be electronic now) 4. Training, coding, intercoder reliability assessment, etc. Example: Shelton & Skalski (06) Content analysis of Facebook, created by Mark Zuckerberg at Harvard in 2004 as online college social network. Spread to other universities and now has 300 million unique users, including more than 90% of college students nationwide (plus just about everyone else now). Survey finding: More than 2/3 of users log in every day, for average of almost 20 minutes (Vance & Schmitt, 2006) What’s on Facebook? Users of Facebook create profiles that allow them to: Share personal information. Communicate through Wall posts and private messages. Create and join special interest groups. Add software applications (“killer app”) Post and view photos (number one photo site!) Sample Facebook Profile Controversy! The CONTENT of Facebook came under fire early, after searches by university officials, athletic offices, and employers: Campus police using site for investigations. Top LSU swimmers lost scholarships. Illinois University grad denied consulting job in Chicago based on interests. How prevalent is “controversial” content, as of 2006? Study Overview Although media coverage might suggest Facebook is filled with negative content, very little empirical evidence exists. Present study set out to examine the extent to which “pro-academic” and “antiacademic” content appear on Facebook, through method of content analysis. Research Questions RQ1: How prevalent is controversial content on Facebook? RQ2: How frequent is anti-academic behavior compared to pro-academic behavior? Sample Primary unit of analysis and sampling: The profile (and corresponding photos). QUESTION: What’s the best way to draw a random sample of Facebook profiles? ANSWER: The site has (had) a built-in random selector! Selected profiles and photo sets sampled and archived in PDF format. Measures All variables except sex and age coded as “present” or “absent.” Several basic profile content variables. Interests/Wall post content variables: Reference to partying Reference to alcohol Reference to drug use Profanity Measures Photo variables: Partying shown Alcohol shown Alcohol consumption shown Drugs shown Drug use shown Physically/sexually suggestive contact Nudity Nonverbal aggression Studying/reading Meeting with a group Sitting in class Training and reliability Five coders given detailed codebooks and coding sheets, which were refined during extensive training. Preliminary coding revealed need for two sets of coders: One profile (3), and one photos Cohen’s kappa on all but two variables was .80 or above (interests reference to partying = .66; drug use interest = .71). Selected results: Content by type (profile frequencies/percentages) Interests: Alcohol (23/11.1%) Partying (14/6.7%) Profanity (5/2.4%) Drug Use (4/1.9%) Wall Posts: Alcohol (76/36.5%) Partying (48/23.1%) Profanity 41/17.7%) Drug Use (3/1.4%) Selected results: Content by type (profile frequencies/percentages) Photos: Alcohol Shown (110/52.9%) Partying (95/45.7%) Sexually Suggestive Contact (51/24.5%) Alcohol Consumption Shown (28/13.5%) Nonverbal Aggression (9/4.3%) Drugs Shown (7/3.4%) Drug Use Shown (4/1.9%) Studying/Reading (2/1.0%) Sitting in Class (2/1.0%) Meeting with a Group (2/1.0%) Nudity (0/0%) Limitations and Future Directions Limitations: Sample only from University of Minnesota Private profiles much more common now Limited content categories and sources Photo sampling technique Future research: More multivariate analyses Linking content analysis and survey data The End Questions? Comments? Suggestions? For a copy of the paper and any of the coding materials, contact me, via email (p.skalski@csuohio.edu) or Facebook!