Editing Oil drop experiment (section)

==Controversy==
Some controversy was raised by physicist [[Gerald Holton]] (1978) who pointed out that Millikan recorded more measurements in his journal than he included in his final results. Holton suggested these  data points were omitted from the large set of oil drops measured in his experiments without apparent reason. This claim was disputed by [[Allan Franklin]], a high energy physics [[experimentalist]] and [[philosopher]] of science at the [[University of Colorado at Boulder|University of Colorado]].<ref>{{cite journal|last=Franklin |first=A. |title=Millikan's Oil-Drop Experiments |journal=[[The Chemical Educator]]|volume=2 |issue=1 |year=1997 |pages=1&ndash;14 |doi=10.1007/s00897970102a|s2cid=97609199 }}</ref> Franklin contended that Millikan's exclusions of data did not substantively affect his final value of ''e'', but did reduce the [[errors and residuals in statistics|statistical error]] around this estimate ''e''. This enabled Millikan to claim that he had calculated ''e'' to better than one half of one percent; in fact, if Millikan had included all of the data he had thrown out, the standard error of the mean would have been within 2%. While this would still have resulted in Millikan having measured ''e'' better than anyone else at the time, the slightly larger uncertainty might have allowed more disagreement with his results within the physics community. While Franklin left his support for Millikan's measurement with the conclusion that concedes that Millikan may have performed "cosmetic surgery" on the data, [[David Goodstein]] investigated the original detailed notebooks kept by Millikan, concluding that Millikan plainly states here and in the reports that he included only drops that had undergone a "complete series of observations" and excluded no drops from this group of complete measurements.<ref>{{Cite journal  | last = Goodstein  | first = D.  | author-link = David Goodstein  | title = In defense of Robert Andrews Millikan | journal = Engineering and Science  |volume=63  | issue = 4 | pages = 30&ndash;38 | publisher = Caltech Office of Public Relations  | location = Pasadena, California | year = 2000 | url = http://calteches.library.caltech.edu/4014/1/Millikan.pdf }}</ref><ref>{{cite journal  | bibcode = 2001AmSci..89...54G | doi = 10.1511/2001.1.54| title = In Defense of Robert Andrews Millikan| journal = American Scientist| volume = 89| issue = 1| pages = 54| year = 2001| last1 = Goodstein| first1 = David| s2cid = 209833984| url = http://calteches.library.caltech.edu/4014/1/Millikan.pdf}}</ref> Reasons for a failure to generate a complete observation include annotations regarding the apparatus setup, oil drop production, and atmospheric effects which invalidated, in Millikan's opinion (borne out by the reduced error in this set), a given particular measurement.