Test construction basic issues

The actual hypothesis to be tested must be carefully chosen. The basic issue is to ensure a safe working environment and we interpret this in the sense that safety must be statistically demonstrated or the environment is considered unsafe. Thus a hypothesis is constructed that the 95th percentile of the distribution of air concentrations is equal to the 95th percentile of the limiting distribution BQ (null hypothesis) and an environment is not considered safe until statistical evidence demonstrates that < BQ(alternate hypothesis) in a lower one-sided test. (Note BQ= ygagd-645) where yg and ag are parameters of the limiting distribution.)... 

Therefore, as the above paragraphs suggest, schema-based assessment requires a reconsideration of the basic principles behind test construction and use. The issues just mentioned outline some of the differences between schema-based tests and traditional tests. The remainder of this chapter examines some of the major implications of schema theory for test development and test interpretation. Some of the implications are far-reaching and must await new methodologies. Others are candidates for immediate implementation. 

Research has been carried out to test the psychometric properties of the Safety Attitudes Questionnaire and AHRQ Hospital Safety Culture Survey (Nieva and Sorra 2003 Sexton et al. 2006 Sorra and Nieva 2004). However, there have been important misapplications of survey tools in healthcare. A proliferation of climate surveys now exist, including many where the factor stmcture and construct validity have not been tested. Many hospitals have developed their own bespoke survey tools and these are poorly designed because basic survey design rules have not been followed. One common issue is surveys that do not counter-balance positive and negative statements, thus increasing the risk of response set bias where the... 

---