Most people aged 50-plus remember 1968 as the year of student revolt, but personality psychologists remember it as the year Walter Mischel published his book on Personality and Assessment, the year that everything changed for personality research. Personality and Assessment has been described as arguing that there is no such thing as personality, and there is no way of assessing it. Mischel reviewed previous research and drew four conclusions, which, if they were confirmed, would largely demolish a traditional trait and factor model of personality:

1. People are not consistent over time.

2. People are not consistent across different places, or situations.

3. Any consistency that can be found is ‘constructed’ by the tests, or by fallible human observers.

4. Personality tests do not work; they cannot predict any interesting or important outcomes.

In the second half of the book he outlined his alternatives to traditional personality models, which are covered in this book's Chapters 4 and 6.

Consistency over time

Personality questionnaires, like the Eysenck PQ or the NEO, usually achieve reliabilities around 0.80 over short periods. A large body of research reports data on the stability of personality over longer periods, of up to 40 years, and is summarised in two recent reviews. Roberts and DelVecchio (2000) find:

• an average stability correlation of 0.50 for 18–21-year-olds;

• stability increases with age, reaching 0.70 by the 50s; but

• stability is much lower in small children;

• stability is the same in males and females;

• stability does not vary across the big five factors;

• stability seems to be the same for PQs, projective tests, and others’ reports, but test type is confounded with age, because children are usually assessed by parent or teacher report, while adults are usually assessed by PQs or projective tests.