Sorting tasks, analyses

A second analysis of the sorting tasks yielded additional information about how accurately students perceived similarities and differences among the problems they classified. Notice that these are not tests per se about whether students learned to associate names with figures. Rather, they assessed aspects of the students learning about the situations. 

We can also compare the schema-based measures with a different outcome measure, still using the same general approach described above. One candidate for comparison is the first sorting task of M. A. Brewer s experiment (1988), described in chapter 9. The dependent variable in the analysis now is CLASSIFICATION, the... 

Verification (refer Fig. 1.3) can be accomplished via some sort of human reliability analysis based on specific task analysis or test data. 

The sorting task can also be applied to achieve a selection of products prior to a descriptive analysis. Piombino et al. (2004) developed a two-step strategy for the analysis of wine flavor. The sorting of a large number of samples by untrained subjects led to the selection of a representative subset of products that were evaluated, in a second step, by descriptive techniques. 

The psychometric rationale behind cluster analysis is to assume that there is a latent partition of the products that each subject is trying to retrieve in the course of the sorting task. Thus, it follows that an appealing feature of such a strategy of analysis is that the outcomes (i.e. a partition or a hierarchically organized set of partitions) are of the same nature as the inputs (i.e. partitions of products associated with the various subjects). 

Courcoux, R, Faye, P. and Qannari, E. M. (2014). Determination of the consensus partition and cluster analysis of subjects in a free sorting task experiment. Food Quality and Preference, 32, 107-112. 

Becne-Bertant, M. and Le, S. (2011). Analysis of mnltihngnal labeled sorting tasks application to a cross-cultural study in wine industry. Journal of Sensory Studies, 26, 299-310. 

If this is not the case, but if you already have knowledge or expertise about the sensory space derived from your product set, use it You can, for example, process a classification of your products on the basis of a quantitative analysis or a free sorting task. After classification, you can choose a prototype of each cluster as the best representative of this cluster (the closest to the barycentre, for example). 

Task analysis forms the basis for the use of techniques for error prediction, which are currently in an early stage of development. They all depend upon one or other sort of checklist. For example each sub-task can be subjected to the following standard list of questions to specify what would happen if these types of error occurred and how such an error could happen (cf. HAZOP) ... 

Once a task has been identified as hazardous, controls need to be applied. For complex tasks, analysis may be required to sort out the component parts of what is actually going to be done. This can result in task improvements and risk reduction. A formal technique for doing this is known as job safety analysis, which has developed from work study practices. Essentially, the task is observed, and broken down into steps or stages which are then examined for the level of risk involved (Table 6.1). Control measures are then worked out, and the paperwork is then used to generate a written safe system of work. Naturally, the tasks selected need to be reviewed at intervals to make sure that the analysis and solutions remain valid over time. 

Exploration analysis is not adequate when the task of the analysis is clearly defined. An example is the attribution of each measurement to a pre-defined set of classes. In these cases it is necessary to find a sort of regression able to assign each measurement to a class according to some pre-defined criteria of class membership selection. This kind of analysis is called supervised classification. The information about which classes are present have to be acquired from other considerations about the application under study. Once classes are defined, supervised classification may be described as the search for a model of the following kind ... 

Most people have seen the cartoon — a big complex diagram of some outfit s Methodology it s got all sorts of purposeful-sounding tasks, numerous feedback loops and checkpoints, extensive forms and documentation templates to fill out, matrices to put check marks against and metrics to tell you how well you are doing. Inputs to analysis at one end, code out at the other. But in the middle of all this busy activity, all arrows converge sooner or later on the box that says And Then A Miracle Happens ... 

All MS technologies require the establishment of method-specific mass libraries so that compounds in the spectra can be identified [212], a tedious task that has been restricted to large laboratories. Nevertheless, some of these efforts are driven by the metabolomics community, thereby requiring some sort of standardization to conduct comparable experiments, as has been proposed with the ArMet standard [216], Last but not the least, metabolomics experiments generate large amounts of data that need sophisticated analysis methods to extract biological information, usually based on multidimensional statistics [3, 5, 58, 209, 217, 218]. Metabolomics experiments as the basis for an analysis of the possible dynamics of metabolic networks are discussed in Section VIII. 

However, the graphical approach is not appropriate for finding the absolute accuracy between more than two properties. The well-established statistical technique of regression analysis is more pertinent to determining the accuracy of points derived from one property and any number of other properties. There are many instances in which relationships of this sort enable properties to be predicted from other measured properties with as good precision as they can be measured by a single test. It would be possible to examine in this way the relationships between aU the specified properties of a product and to establish certain key properties from which the remainder could be predicted, but that would be a tedious task. 

---