Chittaranjan Andrade1. 1. Department of Psychopharmacology, National Institute of Mental Health and Neurosciences, Bengaluru, Karnataka, India. E-mail: andradec@gmail.com.
Sir,The July–September 2016 issue of the Indian J Psychiatry carried eight original articles. Of these, one[1] provided only a range of values for the age of the participants, and another[2] provided no description of the sample, at all. These are serious limitations of a manuscript because, without descriptive information about the sample, readers cannot know to what population the results of a study may be generalized.Three other papers[345] presented data on age not as mean (standard deviation) values, but in class intervals or groups that were otherwise defined. This is completely illogical; age is a number, not a group. When continuous variables are categorized, precision is lost, and the ability of inferential statistical tests to identify statistical relationships is weakened. Continuous variables should be categorized only if there is a specific need, such as for administrative purposes, or when the data could not be accurately recorded, or when the distribution is skewed.[67]Similar considerations apply to other continuous variables, as well; for example, education can be operationalized in units of years rather than as specific levels of attainment as presented by at least two teams of authors.[35]