Cross Validation: Analysis of Pretested Items
All of the results presented up to this point were based on items that have appeared in final forms of the GRE. It is therefore natural to wonder how well the results derived from these items hold up with items that have not been administered before. A significant level of prediction would show that the predictors of difficulty we have studied up to this point with items that have appeared in final forms are sufficiently stable to withstand scrutiny on “raw” items. Such a finding would obviously be of theoretical interest but would be even more interesting for practical reasons.
KeywordsCage Flare Tempo Astigmatism Nale
Unable to display preview. Download preview PDF.