Report on the Impact of School Libraries on Student Achievement

5.1 Academic attainment

Haycock (1995a, 1995b) presents a useful overview of the evidence that links school libraries and student achievement. Although the focus of the current review is on research completed since 1990, it is useful to note the key research findings in a range of pre-1990 studies which show that:

• in schools with good libraries and the services of a school librarian students perform significantly better on tests for basic research skills;
• students perform significantly better in reading comprehension and in their ability to express effectively ideas in relation to their reading;
• more reading occurs when there is a school library;
• the guidance of a librarian appears to exert significant influence on student achievement in information-gathering;
• in schools with good libraries and full-time librarians, students perform better at higher levels in reading comprehension, and in knowledge and use of reference materials than students in schools with minimal or no library service;
• student achievement in reading, study skills and use of newspapers was significantly greater at seventh grade level in schools with professional librarians than in schools without them.

Since the introduction of the Information Power guidelines in 1988, Haycock suggests, research has mostly focused on the instructional role of the librarian.

The most influential body of research into the impact of school libraries on student achievement is that of Keith Curry Lance and his colleagues in the United States. To date studies have been conducted in eight states and several researchers have replicated his model in other states. A distinguishing feature of the Lance model is that it controls for social and community variables. Although the studies confirm that students' socioeconomic status (represented by the percentage of students eligible for National School Lunch Programs and the percentage of minority students) is the most powerful predictor of student performance on state reading tests, they also show that after accounting for the impact on academic achievement of socioeconomic conditions, 'library media predictors almost always outperformed other school characteristics, such as teacher-pupil ratio and per pupil expenditure' (Lance 2001a, p. 6.) This body of research is important not only for what it reveals about the positive impact of school libraries on student achievement but also for its replicability, at least in the United States.

The first Colorado study by Lance and associates was carried out in 1991-92 using 1988-89 data collected by the Colorado Department of Education (Lance 1993). The study focused on 221 public schools in Colorado and student performance on the Iowa Test of Basic Skills (ITBS) (for elementary and junior school students) or the Test of Achievement and Proficiency (TAP) (for secondary students). Designed to gain better empirical evidence of the nature of the relationship between school libraries and achievement in Colorado's public schools, the study focused on three issues in particular: the relationship between expenditure of school libraries and test performance; the characteristics of the school library programs that explained this relationship; and the contribution of library specialists to student test performance.

The study identified several aspects of school library service as direct and/or indirect predictors of academic achievement while controlling for a range of other school and community variables. In particular, it was found that the size of a school library's staff and collection is the best school predictor of academic achievement, represented in this study by reading scores alone. According to the study, students who score higher on norm-referenced tests tend to come from schools with more library staff, more books, more periodicals and more videos (Lance et al. 1993). The school librarian was found in the first Colorado study to contribute to student academic achievement by shaping the school library collection and collaborating with classroom teachers.

Despite its original contribution to the research on school libraries and student achievement, there were some limitations in the original study, which Lance sought to address in subsequent studies. The greatest limitation Lance identified was the use of norm-referenced test scores to 'operationalise academic achievement'. At the time of the first Colorado study authentic assessment techniques had only just begun to replace norm-referenced tests. Other limitations identified were the lack of available data for many potential library, school and community predictors, such as data related to alternative teaching styles, prevalence of disciplinary problems, and student turnover rate, the way in which information skills are taught, how teachers and library specialists plan co-operatively, and the role of technology in the school libraries. Most of the data were also collected before the release of Information Power, the national guidelines for school library media programs, whereas more recent studies undertaken by Lance and colleagues have been able to use the guidelines as a framework for examining the multiple roles of the school librarian.

The findings of the first Colorado study in relation to the upper grade levels could also have been strengthened, Lance suggests, as the numbers of schools involved in the analysis at these upper grade levels was sometimes quite small (Lance 1994). Another limitation that Lance was keen to address was that although the first Colorado study acknowledged the importance of the instructional role of the librarian, it did not define what this meant or address issues related to the value of having principal and teacher support. Nor did it demonstrate the relationship of information technology to student achievement (Lance 2001a).

The second Colorado study sought to redress some of these shortcomings and to determine whether the results of the first Colorado study held up over time, across states, and when a state's standards-based test was substituted for a norm-referenced test (Lance 2001a, p. 3). Unlike the first Colorado study, which focused on grades 1, 2, 4, 5, 7 and 10, the second Colorado study looked only at grades 4 and 7, which Williams, Wavell and Coles (2001) suggest could be a possible weakness, as it means that the findings are not necessarily applicable to secondary education, 'especially when the very different teaching methods are taken into account' (p. 9). Other studies undertaken by Lance have been more balanced in this respect, although the body of research on school libraries and student achievement as a whole seems to have focused more on primary school students than on secondary school students.

A series of library variables was used to predict student performance on the Colorado Student Assessment Program (CSAP) reading test. The study found that CSAP reading scores increase with increases in school library program development (such as library staff hours per students, print volumes per students, library expenditures per student); information technology (where networked computers link with other instructional sites); collaboration between teachers and library specialists (including the amount of professional development time school librarians spend with teachers); and individual visits to the school library (particularly where flexible scheduling allows students to visit the school library as an individual rather than as part of a group). The indirect predictor of student achievement was the leadership involvement of the library staff; that is, the more involved the library specialist, the higher the level of collaboration, which in turn has an impact on test scores (Lance 2000b).

Using more comprehensive data than in the first Colorado study, the second study weighed the relative effects on academic achievement of librarian staffing, other school characteristics (such as per pupil spending, teacher-pupil ratio) and community conditions (such as adult educational attainment, poverty). 'While community conditions proved to have the strongest impact, and instructional expenditures per pupil proved a strong factor at the secondary level, the librarian-pupil ratio outweighed both per pupil expenditures and teacher-pupil ratio at the elementary level and added to the effects of teacher-pupil ratio at the secondary level' (Lance 2000b, p. 7). As with other studies conducted by Lance et al., the findings cannot be explained away by school or community differences.

According to Lance (2001a), the second Colorado study was the first 'to distinguish between the leadership and collaboration activities of library media specialists and to demonstrate the critical pro-active contribution of leadership activities to setting the stage for collaboration and, in turn, higher achievement levels for students' (p. 5). Examples of such leadership and collaboration include spending time attending faculty and curriculum meetings, meeting with the principal, providing information skills training for teachers, meeting with other library professionals, planning programs with teachers and supporting computer technology in the library. Although the study points to the importance of networked computers linked to library resources, it 'does not discuss to what extent the resources were used by students or whether ICT skills which are unrelated to test scores might have a bearing on test scores' (Williams, Wavell and Coles, p. 10).

The Alaska study (Lance 2000a) was based on a 1997-98 survey of school libraries in Alaska and used multiple analysis techniques to assess each library program characteristic as a potential predictor of academic achievement. The focus was on students in Years 5, 8 and 11, thus offering more adequate representation at the secondary level. Its findings support and build on the earlier studies. The study found that it is as teachers, information specialists, and program administrators that school librarians generate their influence on academic achievement. A significant finding of the Alaska study was that schools with more library staff spend more time teaching information literacy, 'resulting in more student visits to library media centers and, in turn, higher reading scores' (Lance 2001a, p. 4). Lance notes that the Alaska study was the first to identify the importance of the school librarian as a teacher of information literacy and the impact of the achievement of the library specialist as an in-service trainer of staff (Lance 2001a).

Similar results were found in the Pennsylvania study (Lance 2000C), where Pennsylvania System of School Assessment (PSSA) reading scores were shown to increase with increases in staffing, information technology, and the integration of information literacy into the curriculum. It was also found that as library staffing, information resources and information technology increase, so too does school librarian involvement in teaching information skills to students, a relationship that cannot be explained away by other school (such as expenditures per student, teacher characteristics, teacher-pupil ratio, student characteristics) or community conditions (such as adult educational attainment, families in poverty, racial/ethnic demographics) conditions. The Pennsylvania study shows the importance of an integrated approach to information literacy teaching.

Similarly, the Oregon study (Lance 2001b) found that Oregon reading test scores increase with increases in total staff hours per 100 students (including both professional and support staff), print volumes per student, periodical subscriptions per 100 students, and library media expenditures per student. Program development alone accounted for a three to five per cent variation in Oregon reading scores, with the exception of the high school level, where community factors are said to have masked the impact of the library program. This trend, in which the impact of library programs decreases as high school students near graduation, is consistent with the pattern revealed in the original Colorado study.

More recently, and following in the same tradition as the earlier state-based studies, research was conducted into the impact of school library programs in Iowa. The main focus was on the reading test scores of students in grades 4, 8 and 11. As well as looking at the characteristics of library staff that affect student academic achievement, and the contribution of collaboration between school librarians and teachers to the effectiveness of the library programs, the study also sought to investigate 'the growing role of information technology' in school library programs, especially in regard to licences, databases and the Internet.

As with the earlier Lance studies, the research shows that Iowa reading test scores increase with the development of school library programs (Rodney, Lance, Hamilton-Pennell 2003). Taking other factors into account, this development alone is said to account for about 2.5 per cent of the variation in the reading test scores for students in grades 4 and 8. The school librarian, the Iowa study shows, exerts 'a complex web of effects' on the library programs (p. ix). An effective school library program is said to be one that:

• is adequately staffed, stocked and funded; the relationship between this aspect and achievement is incremental: as staffing, funding and size of collection rise, reading scores increase;
• has staff who are active leaders in the 'teaching and learning enterprise' of the school;
• enjoys a close working relationship with classroom teachers through collaborative planning of instructional units, the provision of information literacy skills to students, and the provision of in-service training to teachers; and
• embraces networked information technology (p. ix).

Similar studies were carried out in Massachusetts (Baughman 2000) and Texas (Smith 2001). The Massachusetts study found a strong correlation between school libraries and student achievement. Mean Massachusetts Comprehensive Assessment System (MCAS) scores were higher in schools that had school library programs at all levels as opposed to schools that do not have such programs. The study showed that 'the highest achieving students attend schools with good school libraries' (Baughman, p. 10.) Other findings were that at each level, higher MCAS scores were associated with increased use of the school library, higher per pupil book count, and longer opening hours. At the elementary and middle/junior high school levels, students also scored higher on the MCAS when there was a library instruction program and larger per pupil expenditures for school library resources.

Similarly, Smith's study (2001), which examined data from a random sample of 600 Texan school libraries, at elementary, junior and high school levels, as well as community socioeconomic data, in order to determine the impact of school libraries on student performance, found that students achieved higher scores on the Texas Assessment of Academic Skills (TAAS) at each level in schools with teacher librarians than in schools without librarians. The study used more than 200 variables to examine the relationship between libraries and TAAS performance, including thirty-four identified as significant indicators of library performance. These variables were grouped into the areas of program development, leadership, collaboration through teaching, library technology and school technology.

In a similar pattern to the Colorado studies, the Texas research showed that whereas socioeconomic variables-such as the percentage of white students, Hispanic and economically disadvantaged students-explain most of the variance in TAAS performance at all educational levels, library variables explained about four per cent of the variance in TAAS performance at the elementary and middle/junior high school levels and just over eight per cent at the high school level (Smith p. 2). The study also found that library variables were generally more important in explaining TAAS variance than school variables, such as the number of computers per student, teacher experience, and teacher turnover rate. The study reveals that 'library staffing levels, collection sizes, librarian interaction with teachers and students, and library technology levels have a positive association with TAAS performance at the elementary, middle/junior high, and high school levels' (p. 3).

At the same time, test performance was found to be associated with different library variables at each level. At elementary school the most influential library variables were collection expenditure, Internet connection and software packages; at middle/junior high level, the significant library variables were identifying materials for instructional units developed by teachers and providing information skills instruction to individuals or groups; and at high school level the library variables found to be important were library hours of operation and staffing levels, collection size and numbers of subscriptions, planning instructional units with teachers, and providing staff development to teachers. Smith concludes that while a causal relationship between school library activity and TAAS performance cannot be 'unequivocally proven', such a relationship is 'highly plausible' (p. 2). One significant finding of the study was that 'libraries can play a very special role in providing enrichment to those students who come from economically disadvantaged backgrounds and who need additional help to develop the skills they will need to succeed' (p. 3).

Another important source of evidence showing the positive impact of school libraries on student achievement relates to the Library Power initiative. This initiative was set up in 1988 when the DeWitt Wallace-Reader's Digest Fund provided funding to nineteen communities in the United States to improve their school library programs. The studies draw on both qualitative and quantitative data. The data from the schools involved in the studies demonstrate that 'improvements to the library facilities and collection and an emphasis on collaboration between the librarian and teachers are contributing factors to the impact on academic achievement, but that school readiness to accept changes of roles and teaching styles also play a part in sustaining the impact' (Williams, Wavell and Coles, 2001, p. 12).

Oberg's (1999) case study of the Library Power initiative at Lakeside Elementary School in Tennessee was conducted in 1996-97, and uses data gathered from observations, interviews and school documents, as well as scores on the Tennessee Comprehensive Assessment Program (TCAP) tests. The Library Power initiative at Lakeside Elementary began in 1994-95 with the aim of improving student learning by improving school library programs. Although the subsequent 'dramatic improvement in student performance' cannot be attributed solely to the implementation of the Library Power initiative, there is enough evidence to suggest it has made a significant contribution to the positive trend in student achievement that has taken place. Some of the changes to occur as part of the Library Power initiative included more collaborative planning by the librarian and teachers, greater curriculum alignment, flexible scheduling, which allowed students the opportunity to demonstrate independent research skills in the library, a more student-centred approach to learning, improved library collections and facilities, and mandated professional development for teachers.

Some of the evidence attributable at least in part to the impact of the Library Power initiative at Lakeside Elementary School included greater library use by students for both free voluntary reading and curriculum-based projects, more challenging items being selected for personal reading, an improvement in the quality of children's work in the classroom (such as writing reports in their own words), and positive comments made by teachers in regard to the collaborative partnership with the library. When the librarian at Lakeside Elementary compared library use statistics for each classroom with the average TCAP scores for each classroom, 'she found that reading and reference scores began to climb as library use increased' (Oberg, p. 72). Pharr (2002) similarly reports on the impact of the Library Power initiative at the Lakeside Academy of Math, Science and Technology in Tennessee. She notes that after one year of flexible scheduling, when all library projects had been developed through teacher-librarian collaboration, there was a direct correlation between library usage and improved tests scores. Those teachers with the highest rate of library usage also had the highest test scores whereas the teacher who used the library least also had the lowest mastery scores (Pharr 2002).

A study of the impact of the Library Power initiative in seven schools in Philadelphia during 1996-97 drew on interviews with librarians and principals, observation, focus group discussions with librarians, book circulation figures, teacher collaboration information and document analysis (Philadelphia Education Fund 1997a). The most noticeable changes implemented under the initiative were refurbishment of the library; updating of collections and new resources; teachers' preparation of curriculum outlines that would enable the library to purchase relevant materials for topics; flexible scheduling so that students could use the library on an individual, group or classroom basis; student checkout of books and book circulation; degree of collaboration between librarians and teachers; partnerships with other libraries; and professional development for librarians. The extent to which these changes were implemented effectively depended on four variables: the principal's support for the initiative, the librarian's skills and commitment, the professional culture of the school, and school size.

The study found that implementation levels varied among the participating schools and that results tended to be mixed. It was found that although the library's resources remained under-utilised overall, 'the quality of student's use of the library has improved. Indeed, librarians cited students' new love of reading as the most rewarding aspect of the initiative' (Philadelphia Education Fund 1997a, p. 26). According to another Philadelphia Education Fund report (Philadelphia Education Fund 1997b), there is ample evidence to show the ways in which Library Power is supporting student achievement. Indicators of this positive impact include student use of the library to gather information for class projects and recreational reading, student research on projects that have been developed jointly by teachers and librarians, and student ability to recognise the availability of resources relevant to their needs.

A longitudinal study of the impact of Library Power on participating schools, undertaken by researchers from the University of Wisconsin, used data collected from more than 1000 teachers, 400 principals and 400 library media specialists (Zweizig et al. 1999). Conducted over four years from 1994 to 1998, the study found that, as a result of the changes generated by Library Power, schools were able to 'engage students in meaningful and educationally rich learning activities. Instead of being limited to classroom lectures and textbook assignments, by using library resources-books, CD ROMs and the internet, students were able to explore topics in more depth' (Zweizig, p. 14). The study also found that students increased their use of the school library, used the library more on their own initiative, and had a more positive attitude towards using the library. Perhaps the most significant finding of the research into the impact of Library Power on schools is the potential for school library reform 'to leverage important improvements in classroom practice and professional relationships', and to demonstrate to schools that library resources and practices can be used effectively to 'promote a shared curriculum and contribute powerfully to improve instruction' (Zweizig, p. 23).

[Back to review index]