Army officers, and many others, I hadnspent 18 months posing as a homelessnman. 1 wandered across the country,neating, sleeping, and waiting on thenstreets and in shelters in big cities andnsmall communities. This investigationnover a number of years turned upnalmost no “families as couples.” Thusnit was always a shock to see so muchnpublicity about homeless families andnfor them to loom so large in ournnational deliberations. Jonathan Kozol’snRachel and Her Children: HomelessnFamilies in America is an excellentncase in point.nKozol’s story of women and theirnchildren in welfare hotels in New YorknCity focused the spotlight on the multitudenof stories and journal articles onn”homeless families” being produced atnthe time. But the reality is that thesenfragments of relahonships are not families,nthat they do not reflect the sort ofnkinship arrangements most peoplenthink of when they hear or use thenword “family.” Kozol’s title shows hownthe confusion is created: he found onlynRachels — with or without children,nand he subtitled them “America’snhomeless families.”nAdvocates and shelter operators usenthe word “family” in a variety of ways.nIn fact, they use it to indicate almostnany arrangement where a homelessnperson is with someone else. RecentlynI asked a Salvation Army worker hownmany families were staying in theirn”Family Shelter.” She said, “Oh, therenaren’t any here. There was a firenSunday night and we had to place ournwomen in other shelters, hotels, andnplaces.” I asked what she had donenwith the husbands. “We don’t allownany men in the shelter,” she said.nIn the same town, Nashville, Tennessee,nthere is a “Family Life Center”nwhere men are also barred. And, believenit or not, there is a shelter that isnallowed by metro space regulations tonhouse one and a half families. Nashvillenhas a population of nearly onehalfnmillion and about nine hundrednhomeless people, typical, I believe, ofnthe homeless population in most bigncities. After contacting every shelternthat serves families and each agencynthat I thought would have informationnabout homeless families in Nashville, Inwas able to count eight man-womanncouples, who may or may not havenbeen husband and wife.nDennis Duggan, who is now directornof the South Bend Center for thenHomeless, served as director of a shelternin San Antonio for three and a halfnyears. His records indicate that onnaverage 22 “families” came to hisnshelter every day. Of these, he said 19nwere single women, nine of whomnwere teenagers with a baby, leavingnthree women with “either a boyfriendnor a husband.” Dennis said that shelternoperators engaged in fund-raising usenthe term “family” very loosely “tonperpetuate the myth of the homeless,nhard-working, man-wife family and tonappeal to the culture of those who havenmoney.”nThe mission in my hometown ofnMurfreesboro, Tennessee, in its quarterlynreport for January-March 1990,nlisted 195 “total nights of shelter.” Inasked the director, who cooks suppernfor these people so he should know,nhow many of these were families. Hensaid, “Oh, about four, I imagine.” Andnthese are transient families at that,nbecause the rules of the mission do notnallow local families (but do allow localnsingle men) to stay at the mission.nThus, the information I have collectednabout homeless families, from SouthnBend to San Antonio and beyond,nindicates that a man and a womanncalling themselves a family is the rarestnarrangement on the streets.nTurning to the work of others, whatndo we find about homeless families?nThe data in an excellent new book,nMartha Burt and Barbara Cohen’snAmerica’s Homeless (published by thenUrban Institute, 1989), are based onnthe “first nafionally representative surveynof homeless people and providersnin large cities in America.” Of allnhomeless adults in the survey, 79 percentnwere single men, 10 percent werensingle women, and 9 percent werenwomen with children. “The remainder,”nthe authors report, “about 2,000nor 1 percent, were men with children.”nThat’s it, “the remainder.” No manwomannpairs, no couples, no husbandsnwith wives. Not a mention of spouses.nIn short, no word of families as wengenerally think of the term. Further,nBurt and Cohen use “families” innquotation marks in the early discussionnbut discard this qualification of thenterm and use families unadornednthroughout the remainder of the book.nTheir figure of 10 percent of thennnhomeless being “families,” they say, isnlower than most of the reported figures.nBut they claim it is consistentnwith the 9 percent given by the onlynother study of homeless persons thatnincluded a national but non-randomnsample, J.D. Wright and E. Weber’snHomelessness and Health (1987).nThe source of the most reliablenstatistical information about the presencenof families among the homeless,nhowever, is very likely Peter Rossi’snDown and Out in America: The Originsnof Homelessness, published lastnyear. The value of Rossi’s book derives,nin part, from his use of the results of 40nother surveys of homeless peoplen(other, that is, than his own classic andncontroversial study of the homeless innChicago). Having the scholarly virtuesnof objectivity and reliable numbers, thenbook is well-liked even by conservatives,nespecially — or perhaps only —nthe first six chapters. Rossi’s diagnosis isnright on; his therapy, beginning atnchapter seven, is way off.nHomeless families are as scarce innRossi’s study as they are in the othernreliable ones, which is to say they arenmissing. He writes, “The homelessnhave failed in the marriage marketnabout to the same extent as they havenfailed in the labor market,” and thatn”homelessness is almost identical withnspouselessness.” And he says this forngood reason.nFrom the Chicago study, he analyzednall those homeless adults whonwere with relatives and found that onlyntwo were with spouses. Ninety-onenpercent were alone, with a relativelynlarge number of homeless women withnchildren, a very few homeless personsnwith some other relative, and merelyntwo pairs of husbands and wives.nOf the persons who acknowledgednhaving relatives living in the Chicagonarea, Ross asked two questions. First:n”Assuming that they would want tonand have the room, would you benopposed to or in favor of going to livenwith them?” He then asked his subjectsnif they believed their Chicagonrelatives would be opposed to or innfavor of their moving in. The resultsnare as significant as they are totallynignored by the homeless advocates. “Itnis clear,” Rossi writes, “that they themselvesnbelieve they are not wanted asnhousemates [and] it is obvious that theynwould not want to live with theirnOCTOBER 1990/53n