Preparing the data file - G-String-Legacy/GS_MV GitHub Wiki

Return

• The numeric, ASCII-format data file lists alls the ratings of the the candidates' responses in a highly structured form, ensuring unequivocal data entry.
• The data format is that of a table containing only data, consisting of optional leading index columns, followed by score columns of identical data type and similar range. The table must be free of title, header and formatting rows.
• Most of the time index columns are ignored, except in the case of so called 'replication studies'. Replication studies are useful, when the sample size of the first facet of generalization differs from candidate to candidate (e.g. teachers being evaluated by students, or other samples of convenience).
• The index columns uniquely specify the row; the order of scores within a row is determined by the residual, implicit indices.
• If all residual-index facets are crossed, i.e. there is no nesting of these facets, all rows must be of equal lengths.
• Row items must be separated by either commas (csv) or tabs (tsv). Spreadsheets make this kind of formatting easy.