Objective - The objective of this systematic review was to assess which library instruction methods are most effective for improving the information skills of students at an introductory, undergraduate level, using cognitive outcomes (measuring changes in knowledge). The study sought to address the following questions: 1) What is the overall state of research on this topic? 2) Which teaching methods are more effective? Methods - This project utilised systematic review methodology. Researchers searched fifteen databases and retrieved 4,356 potentially relevant citations. They reviewed the titles and abstracts for relevance, and of those, 257 complete articles were considered in-depth using a predetermined inclusion/exclusion form. There were 122 unique studies that met the inclusion criteria and were subjected to an extensive data extraction and critical appraisal process. Of these studies, 55 met author-defined quality criteria to provide information on the effectiveness of different teaching methods. From this review there was a final group of 16 studies with sufficient information to enable meta-analyses and calculations of standardized mean differences. Results - The overwhelming majority of studies were conducted in the United States (88%). Experimental or quasi-experimental research methods were used in 79 studies (65%). Teaching methods used in the studies varied, with the majority focused on traditional methods of teaching, followed by computer assisted instruction (CAI), and self-directed independent learning (SDIL). Studies measured outcomes that correlated with Bloom's lower levels of learning ('Remember', 'Understand', 'Apply'). Sixteen studies compared traditional instruction (TI) with no instruction, and twelve of those found a positive outcome. Meta-analysis of the data from 4 of these studies agreed with the positive conclusions favouring TI. Fourteen studies compared CAI with traditional instruction (TI), and 9 of these showed a neutral result. Meta-analysis of 8 of these studies agreed with this neutral result. Another group of 6 studies compared SDIL with no instruction, and meta-analysis of 5 of these agreed that the result was positive in favour of SDIL. Conclusion - Based on the results of the meta-analysis, there is sufficient evidence to suggest that CAI is as effective as TI. Evidence also suggests that both TI and SDIL are more effective than no instruction. Additional comparative research needs to be done across different teaching methods. Studies comparing active learning (AL), CAI, and SDIL would greatly enrich the research literature. Further studies utilizing appropriate methodologies and validated research tools would enrich our evidence base, and contribute to the growth of knowledge about effectiveness of particular teaching methods.
