School librarians are increasingly interested in including digital resources, especially video and audio, in their recommendations to teachers. The 2,025 librarians reflected in the 2012 annual national SpeakUp! longitudinal survey report of educators, parents, and students administered by the independent educational consulting company Project Tomorrow stated selection priorities of

• content accuracy;
• ease of use by teacher and student;
• alignment with state and national curriculum standards;
• credibility of the organization producing materials; and
• flexibility of content for a variety of uses.³²

Indeed, fostering visual literacy, particularly in relationship to building interdisciplinary understanding is gaining profile as an important function of the school librarian.³³

Research on digital video in US schools is an emerging area of study in K–12 education, to date characterized by research limited to specific sites, funded by video-producing companies, or proprietary and released only in summary form.³⁴ For example, highlights from a study conducted by the Grunwald Associates research firm for the Public Broadcasting Service (PBS) in 2010 are only included in a PBS press release.³⁵ Highlights from this study indicated the benefits of the use of video derived from PBS television programming in the classroom but information about the study sample, questionnaire, or analysis process was not available. Earlier studies emphasized the power of video to facilitate science and mathematics learning.³⁶ However, many of these studies were conducted ten years or more years ago, and curriculum standards, educational technology, and content providers have changed dramatically in the ensuing years. Research has not kept pace with the advancement of streaming video adoption and use.

This paucity body of research on the use of educational digital streaming video is problematic considering the prevalence of teachers’ streaming media use and the instructional differentiation afforded by video technology. With the current growth of freely available web-based video services such as YouTube and TeacherTube (www.teachertube.com/) plus web 2.0 tools for content creation and manipulation, digital video use in the classroom is poised for rapid expansion in ways that may not yet be easily predicted.

Discovery Education Streaming (formerly United Streaming) is a leading subscription multimedia database used by more than 1 million educators and 30 million students in US schools.³⁷ Commissioned studies in Virginia and California have suggested positive relationships between the frequency of use of Discovery Education Streaming and grades 3–8 student achievement in state test results for mathematics and reading. Similar positive relationships have been observed with science learning in Virginia. In Florida, 2009 state test scores were 7.4 percent higher in schools that used Discovery Education Streaming.³⁸

Not all users felt that Discovery Education streaming benefits were easy to attain. Statewide users in Michigan appreciated the tool’s potential as a strong means to introduce new concepts or to allow students to work independently, especially in science. However, educators reported that they found video segments and videos to be outdated, often of poor quality, and the use of the system placed an undue bandwidth usage and network traffic burden on the district or school. Educators listed lack of time to find and preview videos, equipment to project videos to classes, and implementation support as barriers to unfettered use of Discovery Education Streaming. Despite the fact that school librarians are commonly left out of adoption decisions, many serve on the front lines of implementation, assisting teachers with bandwidth capacity management and hardware troubleshooting.³⁹

The use of nonprint and visual resources as learning tools has been underemphasized in science curricula and, even when included, unless the resources are accompanied with sufficient metadata to allow them to be adequately indexed by search engines and described in sufficient detail for users to assess their relevance, can be difficult to find and organize.⁴⁰ School librarians have a professional imperative to lead and facilitate the integration of multimedia, including subscription databases, into teaching and learning.⁴¹ Research led by Mardis suggests that school librarians and STEM teacher collaboration has the potential to enhance the science curriculum using digital video, audio, and applications to teachers but did not systematically integrate digital resources from subscription databases into their collections, thus complicating teachers’ and learners’ attempts to find and use them.⁴²

Data were collected on July 7 and 9, 2012, when the Discovery Education Streaming K–12 digital media library database contained approximately 148,000 multimedia assets accessible through browsing categories for Careers/Workplace Skills; English/Language Arts; Health; Mathematics; Research/Study Skills; Science; Social Studies; Teaching Practices; Visual and Performing Arts; and World Languages. From these categories, the researcher selected for analysis the following three categories: Science, Health, and Mathematics. Discovery Education Streaming developers assigned Science assets into seven subcategories: Earth/Space Science; History and Nature of Science; Inquiry; Life Science; Physical Science; Science and Technology; and Science in Personal and Social Perspectives. The researcher treated the Science and Technology subcategory as a proxy for a Technology category. Mathematics assets were assigned to subcategories of Algebra; Calculus; Data Analysis and Probability; Geometry; Measurement; Numbers and Operations; Problem Solving; and Trigonometry. Health assets were subcategorized in Alcohol and other Drugs; Growth and Development; Mental Health; Nutrition; Physical Activity; Safety; The Body; Tobacco; and Violence.

Engineering assets, which do not have their own dedicated browsing category or subcategory in Discovery Education Streaming, were retrieved by the researcher through a keyword search of the entire database for the word of “engineering.” Engineering assets were not analyzed in subcategories because Discovery Education Streaming does not organize assets on this topic into browsing categories. Engineering assets (N = 600) were located through a keyword search of the term “engineering in the entire Discovery Education Streaming database.”

To mimic the typical Discovery Education Streaming user experience, the researcher used the subscriber interface to collect data on July 7 (Health), July 8 (Mathematics and Engineering), and July 9 (Science), 2012. Due to additions to and withdrawals from the resource collection, database asset counts change daily. Therefore, to aid analysis precision and study replication, data for each category were collected on a single day. The researcher reviewed the Discovery Education Streaming database and recorded asset counts by subject, asset type, and grade level. Local content was excluded from asset counts and the range of possible media types is illustrated in table 2 below.

For each subcategory, the researcher narrowed the view display to the five copyright date ranges available: 1988 or older, 1989–93, 1994–98, 1999–2003, or 2004 or newer. Within each copyright date range, the researcher narrowed the display four times to display the results from only one grade level band, i.e., to K–2, 3–5, 6–8, and 9–12, at a time. For each data collection point, the researcher marked the results and exported the results as comma separated value (CSV) files.

Data were initially analyzed using the Statistical Package for the Social Sciences (SPSS). Case summary reports reflecting descriptive and frequency statistics were generated for each of the four subject category areas. The case reports were also exported to Excel to create the tables and charts displayed in the Results section of this paper. Within the case summary reports, Science and Health results were analyzed by grade level, copyright date, and subcategory. Technology is included in the Science subcategory of “Science and Technology.” Mathematics assets were analyzed by grade level and subcategory. Because the Engineering category was not subcategorized, no results could be obtained from browsing. All Engineering results were obtained via keyword search of the Discover Streaming database.

On the date of data collection (July 9, 2012), Discovery Education Streaming contained 71,702 science assets assigned to eight subcategories. First, all assets in the Science category were analyzed. Table 3 illustrates the distribution of these resources across media type, grade level, and subcategory. As table 3 shows, assets are not distributed equally across subjects and media types.

The majority of Discovery Education Streaming assets are assigned types of video segment (n = 40,148 or 55 percent), image (n =25,996 or 36 percent), and/or full video (n =5,184 or 7 percent). The remaining media types represented about 2 percent of assets. Records for the Science category’s assets were examined by grade level, and subcategory. Table 2 illustrates the results. The distribution of assets across grade bands was 25,032 (34 percent) assigned to grades 6–8; 23,521 (32 percent) to grades 9–12; 15,002 (20 percent) to grades 3–5; and 8,147 (11 percent) to grades K–2. It should be noted that the remaining 3 percent difference possibly reflects rounding and asset assignment to multiple grade bands. Figure 1 further illustrates the distribution of science assets across grade levels and categories.

As figure 1 indicates, most asset records were assigned to the Life Sciences subcategory with 29,204 (40 percent) records, followed by Earth/Space Sciences with 13,977 (19 percent), Physical Science with 7,582 (10 percent) and History and Nature of Science with 7,549 (10 percent). The least number of records were assigned to the Science and Technology subcategory with 5,268 (7 percent); Science in Personal and Social Perspectives with 4,540 (6 percent); and Inquiry with 3,582 (4 percent). Total percentages reflect rounding and asset assignment to multiple subcategories.

Engineering assets (N = 600) were examined by media type and grade level. Discovery Education Streaming does not provide browsing subcategories in the Engineering category, so results were obtained via keyword search. Table 4 illustrates the asset distribution across media type, grade level, and copyright date.

As table 4 shows, the most predominant media types were Video Segment (n = 317 or 52 percent) and Full Video (n = 214 or 35 percent). The remaining assets were comprised of Images (n = 49 or 8 percent), Encyclopedia Entries (n = 14 or 2 percent). Activity (n = 2 or 1.5 percent) and Content Collection (n = 2 or 1.5 percent) account for the remaining 3 percent of media types. Most engineering resources were for grades 6–8 (n = 388 or 64 percent) and grades 9–12 (n = 168 or 28 percent). The remaining forty-four assets (8 percent) were for students in grades K–5. The majority of the assets had copyright dates 1999–2003 (n = 340 or 56 percent) and 2004 or newer (n = 260 or 43 percent).

Discovery Education Streaming included 13,743 assets in the Mathematics category on the date of data collection (July 8, 2012). Table 5 illustrates the distribution of assets in the Mathematics browsing category across grade level, media type, and subcategory.

As table 5 shows, assets are not distributed equally across media types. The majority of Discovery Education Streaming assets are assigned types of “segment” (an edited portion of a video) (n = 10,906 or 79 percent), full video (n = 2,112 or 15 percent), and song (recorded music and lyrics) (n = 491 or 3 percent). The remaining 234 (3 percent) assets represented other media types. Media types are listed and defined in table 2.

Records for Mathematics assets were then examined by grade level, and subcategory. The distribution of assets across grade bands was 4,547 (33 percent) to grades 9–12; 3,812 (27 percent) assigned to grades 6–8; 2,695 (19 percent) to grades K–2; and 2,689 (19 percent) to grades 3–5. It should be noted that the remaining 2 percent difference perhaps reflects rounding and asset assignment to multiple grade bands (see figure 2).

Assets were assigned to eight subcategories. Most asset records were assigned to Numbers and Operations with 4,031 (29 percent), followed by Problem Solving with 2,915 (21 percent) records Algebra with 1,875 (13 percent), Geometry with 1773 (12 percent), Data Analysis and Probability with 1,429 (10 percent), and Measurement with 1,217 (8 percent). The fewest records were assigned Calculus and Trigonometry accounted for 2 percent. Total percentages reflect rounding and asset assignment to multiple subcategories.

On the date of data collection (July 8, 2012), Discovery Education Streaming included 14,603 assets in the Health category. Table 6 illustrates the distribution of assets in the Health browsing category across grade level, media type, subcategory, and copyright date.

Table 6 shows that the number of assets varies across media types, grade levels, and subcategories. The majority of Discovery Education Streaming Health assets were assigned types of video segment (n = 7786 or 53 percent), image (n = 4,507 or 30 percent), and/or full video (n = 2,048 or 14 percent). The remaining media types represented about 3 percent of assets.

Records for Health assets were then examined by grade level, and subcategory. Table 6 illustrates the results. The distribution of assets across grade bands was 5,001 (34 percent) assigned to grades 9–12; 4,770 (32 percent) to grades 6–8; 2,646 (18 percent) to grades K–2; and 2,186 (14 percent) to grades 3–5. It should be noted that the remaining 2 percent difference possibly reflects rounding and asset assignment to multiple grade bands. Figure 4 further illustrates the distribution of science assets across grade levels and categories.

Figure 3 shows that most asset records were assigned Nutrition with 3,408 (23 percent); Growth and Development with 3,406 (23 percent); The Body (1,592 or 10 percent); Physical Activity with 1,490 (10 percent); and Mental Health with 1,430 (9 percent) followed by Safety with 1,390 (9 percent); and Alcohol with 1,236 (8 percent). The fewest records were assigned Violence (486 or 3 percent) and Tobacco (165 or 1 percent). Total percentages reflect rounding and asset assignment to multiple subcategories.

Assets were analyzed in aggregate for an overall impression of media type, grade level, subcategory, and copyright date distribution. Figure 4 reflects media types across all categories and grade levels.

As the figure shows, video segments (n = 59,157 or 59 percent) and images or (n = 30,608 or 30 percent), full videos (n = 9,558 or 9 percent), and songs (1,003 or 1 percent) comprised the majority of Discovery Education Streaming STEM assets. The final 1 percent of assets was comprised of the remaining media types listed in table 2.

Figure 5 illustrates a comparison of copyright dates for all subject categories, which reflect that the majority (75,636 or 83 percent) of assets are older than seven years. As demonstrated by figure 5, science is the category with the most assets (71,702 or 71 percent) and most of the science assets (54,440 or 79 percent) are older than seven years. Health assets are the second largest STEM category with nearly 14 percent (14,603) and 73 percent (10,646) of those assets seven years or older. Math assets comprised almost another 14 percent (13,743) and 73 percent of them were older than seven years. The remaining 1 percent of assets related to engineering topics and 57 percent (340) had copyright dates older than 2003.

Research has demonstrated that teachers benefit from support in identifying high quality instructional materials but that school librarians are often frustrated in their attempts to support STEM learning and promote digital materials into school library collections. This analysis revealed that Discovery Education Streaming could be a potential source of multimedia content for school library collections because it contains more than 100,000 assets in science, technology, engineering, mathematics, and health. Science was the largest category, followed by health; technology and engineering topics represented the smallest number of STEM-related assets. Topically, Discovery Education Streaming appeared to be a good source of content for life sciences, algebra, numbers and operations, and nutrition. Discovery Education streaming also appeared to be a better source for assets for grades 3 and higher, with a concentration on upper elementary and middle grades. While assets in the Health category were plentiful, they were lacking in the Tobacco and Violence subcategories, and these topics are important aspects of learning about healthy lifestyles. Mathematics also appeared to be short on support for important advanced topics such as Trigonometry and Calculus. Science was astonishingly low on assets in the Inquiry category—an important topic to which school librarians can definitely contribute content and process.

The modest asset counts in the Science and Technology and Engineering subcategories are a concern because proponents of STEM education reform advocate increasing the visibility of technology and engineering in the standard K–12 curriculum. Technology and engineering relate to the ways that humans modify the natural environment and therefore it is essential that STEM learning be “expanded to include all kinds of devices, instruments, and tools that can be applied in both domains of science and engineering.”⁴⁸

As its name suggests, Discovery Education Streaming contains a significant amount of streaming video segments and full videos. It also contains a large number of images. Given that content providers range from broadcast sources including the Discovery Channel and PBS and governmental organizations such as NASA and Smithsonian, it is likely that the content of those visual assets reflects accurate science and good quality production. However, the age of the assets in Discovery Education Streaming exceeded seven years, a collection development threshold for STEM resources. In particular, science, engineering, technology and health are highly dynamic fields with rapid advances in knowledge and application occurring constantly. Older or dated resources in these areas run the risk of containing inaccurate or incomplete information and potentially misleading learners.

This study has limitations, including the fact that Discovery Education Streaming asset counts change daily. While efforts were made to analyze all assets in a particular category, the constantly changing asset counts may undermine efforts to replicate the study using the most current database contents. Additionally, neither engineering nor technology were topics represented in Discovery Education Streaming with browsing categories, and using keyword searches of the database to identify assets related to engineering topics may have resulted in the exclusion of relevant assets. Other concerns relate to whether the subject areas with lower numbers reflect of a lack of available K–12 resources in general or whether they are disproportionately underrepresented in the database. It is possible that there are there competing databases that differ substantially from Discovery Education Streaming in the scope of their content. However, considering Discovery Education Streaming’s domination of the market, the researcher chose not to focus on possible resources, but on the most readily available resources to teachers, librarians, and students in US schools.

Ultimately, asset quantity differs from asset quality, and to gauge the true value of Discovery Education Streaming multimedia assets to a school library collection, one may need to do more than count them. This study was an attempt to gain an idea of the extent of the Discovery Education streaming’s collection and a starting point for further research. For example, researchers may wish to use a similar descriptive approach to characterize the contents of other resource databases or delve further into which Discovery Education Streaming resources are actually use by educators and learners.