Olive Jean Dunn
| Olive Jean Dunn | |
|---|---|
Olive Jean Dunn in 1998
|
|
| Born | 1 September 1915 |
| Died | 12 January 2008 (aged 92) |
| Alma mater | |
| Known for |
|
| Scientific career | |
| Fields |
|
| Institutions | |
| Thesis | Estimation problems for dependent regression (1956) |
Olive Jean Dunn (1 September 1915 – 12 January 2008) was an American mathematician and statistician, and professor of biostatistics at the University of California Los Angeles (UCLA). She described methods for computing confidence intervals and also codified the Bonferroni correction's application to confidence intervals. She authored the textbook, Basic Statistics: A Primer for the Biomedical Sciences in 1977.
Dunn studied mathematics at the UCLA, earning a BA in 1936 and an MA in 1951. She was awarded a PhD in mathematics in 1956 at UCLA, supervised by Paul G. Hoel.
The title of Dunn's doctoral dissertation was Estimation problems for dependent regression.
In 1956, she was appointed assistant professor of statistics at Iowa State College. Dunn returned to UCLA in 1959 as assistant professor of biostatistics and assistant professor of preventive medicine and health, later becoming full professor and serving in that role until her retirement. Dunn served on the editorial boards of several journals.
Some of Dunn's 1958 and 1959 work led to the conjecture of the Gaussian correlation inequality, which was only proved by German mathematician Thomas Royen in 2014 and was only widely recognized as proved in 2017.
Dunn's doctoral dissertation work formed the basis for her continuing development of methods for confidence intervals in biostatistics, and the development of a method for correcting for multiple testing. From the notes to her 1959 publication on confidence intervals:
"Most of the research for this paper was part of my doctoral dissertation. The idea of writing an article for the research worker who uses statistical methods was suggested to me by one of the non-statisticians on my doctoral committee at the time of my final examination. In working on the various confidence intervals for k means, I thought of the Bonferroni inequality ones quite early, but since they were so simple I thought they couldn't possibly be of any use. I spent a long time trying to prove that the confidence intervals which would be used in the case of independent variables could also be used or dependent variables. After failing to find a general proof for this, I finally noticed that the simple Bonferroni intervals were nearly as short".
...
Wikipedia