Transform (individual) advanced

Unlike previous versions, in PRIMER 7 the Transform(individual) routine has been moved to a more convenient – and logical – position in the Pre-treatment menu. Its routine use is therefore covered in Section 4, and its application has been seen several times already. However, in order not to break up the presentational flow of a typical analysis pathway in this earlier section, the more complex features of this routine were deferred to this section. As a brief recap, Pre-treatment> Transform(individual) operates on highlighted, not selected, portions of the data sheet (if there is no highlighting it takes place on the entire sheet) and produces a new sheet according to a BASIC language-type (Expression:       ) provided by the user, in which V stands for the existing value in each cell which is being operated upon. A Pick>Type list aids in the construction of expressions by providing a suite of possible functions (•Function), some of which are standard BASIC definitions (LOG(V), EXP(V), INT(V), ... – note that the difference between upper or lower case is ignored) and some are designed specifically for commonly-used operations (e.g. ARCSINE(V) is the often seen arcsin transformation – more often seen than is justified in fact! – in which the exponent is first square-rooted before arcsin, the ASIN(V) function, is applied; these are new to PRIMER 7). The Pick>Type list also has the facility to use the values of an existing (•Sample), (•Variable), (•Factor), (•Indicator) or even whole (•Worksheet), so there is much flexibility to manipulate a data matrix to a new form, totally within PRIMER. Having said that, many users will still find it more convenient for very complex operations to use the tools they are already familiar with outside the package – e.g. in Excel – but saving data to Excel, manipulating and re-opening it in PRIMER is a relatively painless procedure, since Excel moved away from its 255 column limit! (PRIMER v6 and beyond do not have any fixed restrictions on data sheet sizes but are inevitably limited by the available RAM and by execution time for routines such as MDS and SIMPROF, as noted earlier).