file drawer effect

Terms from Statistics for HCI: Making Sense of Quantitative Data

For every positive result published there are many more negative results (non-significant) that are withheld. Such selective publishing may be a deliberate attempt to deceive, but is more often a combination of ignorance and bias. This is exacerbated by journal reviewing practices that typically make it hard or impossible to publish non-significant results.

Defined on page 82

Used on pages 82, 87, 99