![]() ![]() However, in some cases the program suggested doses between 20% and 100% higher than normal for the same therapy, depending on how doctors entered patient information. The program was intended to calculate a safe and effective dose of radiation and had been in use for 8 years without incident. Other software issues can be less obvious but no less critical, such as Multidata International’s Treatment Planning Software used by doctors at the National Cancer Institute in Panama City, Panama. The error was found in a guidance-control system inherited from and thought to have been successfully vetted in the rocket’s previous iteration, the Ariane 4. Such was the case with the maiden flight of the $500-million Ariane 5 launch vehicle, which veered off course 40 s after takeoff and self-destructed (Jezequel & Meyer, 1997). Thorough testing is an essential step in software development and often seemingly working components can perform unexpectedly when used in real-life scenarios. The National Institute of Standards and Technology estimates that software bugs cost approximately $59.9 billion dollars annually and a third of such errors could be detected through testing (National Institute of Standards, 2002). Given the critical role data extraction software may play in quantitative syntheses of published data, it is important for researches to have access to a wide variety of tools with known reliability and validity. Results from the validity analysis also showed high correlations between values extracted by coders and values reported in the articles, ranging from. Pearson product-moment correlations between data extracted by two different coders showed a high correlation (mean r = .9592). They then conducted a validity analysis for 44 of these graphs by comparing data extracted by coders using UnGraph to condition means reported in text or tables of their respective publications. The investigators assessed the reliability of UnGraph by comparing data extracted by two independent coders for each of 91 published graphs. ![]() (2009) assessed the reliability and validity of one such computer program, UnGraph (Biosoft, 2004), compatible with computers running Microsoft Windows operating systems. ![]() One method of obtaining raw data sets from a published study is to extract them from graphs using a specialized computer program. A necessary step in many meta-analyses is to obtain several raw data sets. Doing so may provide information on the generality of interventions (i.e., replications across subjects), influence public policy, or affect the practices of a field (Scruggs & Mastropieri, 1998). While studies involving within-subject designs serve the purpose of identifying functional relations, often investigators may wish to synthesize the results of several studies. Results from within-subject design studies are generally reported graphically (i.e., without reporting numerical values), as visual inspection (i.e., evaluating the data graphically for differences in variability, trend, and level) is the primary method of data analysis with these types of designs (Kahng et al., 2010). Within-subject designs involve repeated measures and are differentiated from group designs in that each individual is exposed to at least one treatment and control condition.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |