A Tribute to Bill Kruskal

Statistic al Scienc e 2007, V ol. 22, No. 2, 262– 263 DOI: 10.1214 /0883423 06000000358 Main article DO I: 10.1214/0883 42306000000420 c  Institute of Mathematical Statisti cs , 2007 A T ribute to Bill Kru sk al No rman M. Bradburn Bill Krusk al w as a dedicated citizen of the Unive r- sit y of Ch icago, of the br oader statistic al communit y and of the coun try . It is difficult to conv ey fully Bill’s commitment to and con tributions to the comm unit y life of the Univ ersit y. The Univ ersit y of Chicago , unusual among Amer ican unive rsities, b eliev es that the main- tenance and enhancemen t of a sc holarly comm un it y is part of its mission. Mainta ining a true sc holarly comm unit y is difficult and rests heavily on those who are willing to commit some of their time and energy to the b ettermen t of the communit y . Bill was an extraordinarily goo d citizen, He c h aired a n um- b er of facult y committees that dealt with some of the most delicate issu es of un iv ersit y life, n otably the Committee on App ointmen t In equities; he was a memb er of the Council of the Univ ersity S enate and of the Committee of the Council; he serv ed as c h airman of the Statistics Department and in a b old mo v e, for he w as not a mem b er of its f acult y , serv ed as dean of the So cial Sciences Division for t w o terms, as w ell as s er v in g as dean pr o temp ore dur ing the Harris Sc ho ol’s inaugural y ear. This co ncern for the Univ ers it y extended to the practices of the library . When h e could not find a particular issue of a Censu s Bur eau serial pu blica- tion, but instead found a note tap ed in the b ound v olume sa ying that the library had discarded n um- b er 8 at the requ est of the Go v ernment Prin ting Office b ecause the rep ort had to o man y mistake s, he w rote in some in dignation, “Ho w could our li- brary discard a docum ent that ipso facto h eld suc h so ciologica l, ethical, statistical and h istorical in ter- est? Horrors. . . . After all w hen the Russians send us replacemen t pages f or their great encyclo p edia, we Norman Br adburn is a memb er of the National Opinion R ese ar ch Center, University of Chic ago, Chic ago, Il linois 60637, USA e-mail: br adburn-norman@n or c.uchic ago.e du . This is an electro nic reprint o f the or ig inal article published by the Institute of Mathematical Statistics in Statistic al Scienc e , 2007, V ol. 22 , No . 2, 262 –263 . This reprint differs fr om the o r iginal in paginatio n a nd t yp ogr aphic detail. add them but k eep the old ones unlik e (I think) the librarians in Mosco w, Minsk and Lenin grad.” Bill had a passionate commitment to the health of the federal statistical system and to the in telligen t use of statistics in the formation and imp lemen ta- tion of pub lic p olicy . I t w as this commitmen t to the federal statistical system that w as the b eginning of m y close co llegial relatio nsh ip w ith Bill. In 196 9 he en ticed me to b ecome a mem b er of a National R e- searc h Coun cil panel on Problems in Census En u- meration, one of the fir st panels to inv estigate iss ues related to the census u ndercount. Although I d id not kno w it at the time, this activit y fl o wed into Bill’s w ork with Allen W a llis on the Pr esiden t’s Commis- sion on F ederal Statisti cs and then to the establish- men t of the Committee on National S tatistics (CN- ST A T) in 19 71. Bi ll b ecame its fi rst c h airman. The stamp he put on the Committee, a s w ith all activ- ities he w as in volv ed in shaping, w as still apparent man y y ears later when I b ecame a memb er and later c h airman of CNST A T. The most striking thing th at I th ink most of us will alwa ys remember ab out Bill w as his wide rang- ing corresp ond ence and the system h e d ev elop ed of sharing his though ts with others thr ough the effi- cien t mec hanism of sending copies of his letters and supp orting materials to a broad and d iv erse set of fello w inte llectual tra vele rs. Ho w man y of us hav e exp erienced, with a mix of deligh t and wonder, re- ceiving a co py of a let ter add r essed to someone w e did or d id not kno w , with a note in the upp er right hand corner in small, but legible handwriting, “File: cc:..” and then a list of names, man y of whom we ma y not ev en h a v e heard of ? Often I wondered what connection there m igh t b e b etw een m e and one or t w o of the others w ho w ere sharin g in Bill’s enlarge- men t of our horizons. Sometimes, of course, they w ere letters addressed directly to me with his characte ristically precise cri- tique of an article I had written, or a p enetrat- ing question th at show ed all too plainly that I had missed an imp ortant b o dy of literature, with a p ar- en thetical reference to a b o ok or article that I sh ou ld certainly read or, if deserved, a bit of praise f or making s ome p oin t that h e particularly app ro v ed 1 2 N. M. BRADBURN of or that was n icely ph rased. If the lette r con tained some criticisms that were n ot only just, bu t su d- denly made one realize one’s o wn shortcoming, then the delica te sente nce that in v ariably ended suc h a letter, “I h op e y ou will n ot mind if I send a few copies of th is exchange to p ossibly intereste d col- leagues in addition to those listed,” wa s read with trepidation. Not only did Bill corresp ond with a wide range of p eople—in a curs ory r eview of my files, I found letters to colleagues in Vienna, Nov osibirsk, Can- b erra and Allen b ac h , Germany—but they co v ered what seems lik e an imp ossibly large set of topics— statistica l, historical, bibliographical puzzles, the ac- curacy of income s tatistics, the role of advisory com- mittees, an early use of the term “public opinion” b y T rollop e, the misuse of significance tests—to men- tion just a few th at are in m y files. Although Bill’s critical eye w as sh arp and he neve r w as one to p u ll pun c hes, he did manage to con v ey his commen ts in a delightful language that (mostly) to ok th e s ting out of h is critique. F or exa mple, “W ould it ha v e b een tot ally impractical to do the randomization b efore brin ging those sw eet women in to sigh t of the P romised Land?” or by usin g a memorable metaphor, “Y et there is one b one that stic ks in my throat, and I write to pic k it, thus p rob- ably mixing t wo osteo-meta ph ors .” Sometimes his irritation at some p ersistent misuse of statistic s would b oil o v er and h e w ould b e more direct, as with the author of an article that used p - v alues to assess the imp ortance of differences, a topic that wa s esp ecially dear to his heart. “So I’m sorry that th is ubiquitous practice receiv ed the accolade of use by yo u and y our d istinguished coauthors. I am thinking these da ys ab out the man y sen ses in w hic h relativ e im p ortance gets considered . Of these senses, some seem reasonable and others n ot so. S tatistical significance is lo w on my ordering.” Then, ever the gen tleman, he adds, “Do forgiv e m y blu ntness.” Bill also had a pr ecise sense of language that makes the author of Eats , Sho ots and L e aves lo ok like an amateur. In resp on s e to m y request for his opinion on someone w e were thinking of hiring, h e replied, “[A]s of now I do not hav e a crisp opinion.” When I ask ed him to b e th e acting d ean of the then new Har- ris Sc ho ol of Public P olicy Studies, he replied, after taking a resp ectable amount of time to think ab out it, that he w ould do it on ly on the condition that he b e d esignated the “dean pr o temp ore.” I hav e al- w a ys thought that this reflected a sense of senatorial dignit y and courtliness that must hav e b een part of his self-image. Bill was mod est and d ev oid of v anity himself. He w as Ap ollonian rather than Dion ysian. He had a wry sense of humor, but I only once r emem b er him re- ally laughin g. One day du ring a con v ersation ab out an academic app oin tment while he was dean of So- cial S ciences, I told him the Max Beerb ohm story ab out En o c h Soames, a minor English p o et wh o sold his soul to the devil in exc hange for immor- talit y . T o prov e that he had k ept his side of the bargain, the devil sends Soames forw ard a h undr ed y ears in time to lo ok himself up in th e catalog of the British L ib rary . There he fi nds the en try: “Eno c h Soames, an imaginary c haracter in a short story of Max Beerb ohm.” Bill laughed uproariously . I think the story must hav e epitomized for him all the v an i- ties and u ncertain ties among the facult y with whic h academic administrators ha v e to deal. After his retiremen t, Bill remained act iv e profes- sionally . He esp ecially enjo y ed b eing a visiting fello w at the Program Ev aluation and Methodology Divi- sion of the General Accoun ting Office, whic h ga v e him a new v en ue in wh ic h to pur sue his life long in terest in impro ving th e uses of statistics by the federal go vernmen t. His legacy in national statistics, as wel l as in the discipline and in the Univ ersity , is manifest in all of us, his colle agues, his studen ts and his myriad friend s. W e sh all m iss the con versat ions and the corresp ondence, but will eve r remem b er him as an extraordinary h uman b eing.