Categorical scale, variable

While correlation may not be the optimal measure of the usefulness of the Epworth scale for discrimination or prediction (32), there also appears to be no categorical association between ESS score and sleep latency (30,33,34). Thus, although the ESS may be predictive of other variables such as general health status (16) or respiratory disturbance index (12), most of the evidence suggests that its association with sleep latency is tenuous. Similar results emerge from studies comparing ESS with the other (newer) objective measure of EDS, the maintenance of wakefulness test (MWT). In a large narcolepsy sample, the correlation between ESS and MWT was weak (r = -0.29, p <. 001) (35). 

Molecular Dynamics simulation is one of many methods to study the macroscopic behavior of systems by following the evolution at the molecular scale. One way of categorizing these methods is by the degree of determinism used in generating molecular positions [134], On the scale from the completely stochastic method of Metropolis Monte Carlo to the pure deterministic method of Molecular Dynamics, we find a multitude and increasingly diverse number of methods to name just a few examples Force-Biased Monte Carlo, Brownian Dynamics, General Langevin Dynamics [135], Dissipative Particle Dynamics [136,137], Colli-sional Dynamics [138] and Reduced Variable Molecular Dynamics [139]. 

Factor is a controllable variable of interest. The factor can be either quantitative or qualitative. A quantitative factor can be measured on a numerical scale. Some examples of qualitative factors include the temperature of a furnace, amount of a chemical, ratio of a material portion, weight of a snbstrate, etc. A qualitative factor can be categorized into a group. Examples include type of material, suppliers, operators, etc. 

In order to explore the dose-response pattern of the full-scale IQ-PbB association we have used PbB as categorical variable with five levels (< 14.9,15.0-24.9, 25.0-34.9, 35.0-44.9, and > 45.0 /ig/dl). The full-scale IQ difference between the two extreme PbB levels was 11.1 units (Figure 5). After allowance for the covariates of the optimal model the full-scale IQ difference was 9.1 units. A consistent decrease of the IQ by PbB was noted in levels higher than 25.0 jUg/dl. 

---