The second change is that as data collection methods have become more automated and data storage has become inexpensive, datasets have dramatically increased in size. As a result, research projects have become more ambitious, collecting many measurements from large samples of subjects. SEM, which has made possible a range of complex analyses, has the potential to be an extremely valuable approach to these new challenges due to i) greater statistical power (less variance in study outcomes), and ii) greater precision (less bias in the results).