On January 2, 1920, a frigid day even by Pittsburgh’s standards, nineteen year-old Henry Silverstein began his first day of work as a census enumerator.  He got off the street car in Homestead near Fifteenth and West, his official census portfolio tucked under one arm, and turned left instead of right.  Things went downhill from there, and within days he became so overwhelmed that he came up with an illegal scheme to be done with the whole mess.  He got away with it for almost a century, too, until a single-minded sleuth (that’s me) decided it would be fun to read through all 140 enumeration districts of the Homestead District, 1880-1940.[1]

Henry Silverstein was the American-born son of two Russian Jewish immigrants who ran a small confectionary in Pittsburgh’s Hill District.  Henry’s older sister married in 1910 and his older brother in 1912, leaving him alone at the age of 12 with his parents and his “deaf and dumb” younger brother, who was eventually sent to the Polk State Institution for Feeble-minded of Western Pennsylvania, never to return.  Henry graduated high school and began work as a traveling salesman for a manufacturer, peddling dry goods all over Western Pennsylvania and West Virginia.  Perhaps he liked being away from home for the required six to eight weeks at a time.[2]

Having availability in January 1920 to work as a census enumerator for comparatively low pay, however, suggests a break in his employment. To become an enumerator, Henry had to submit an application and pass an exam.  Allegheny County struggled to find enough people; they had to raise the pay and re-open the application period twice, the second time just days before the census was due to start.[3]

On January 2, the day Henry and all the other enumerators nationwide began their work, freezing weather overtook the Pittsburgh region.  A long, cold street car ride, with several transfers, took Henry straight from his home to his starting point at the bottom corner of Enumeration District 144 of Homestead’s Third Ward.  Henry’s turning left onto East Fifteenth, instead of right onto West Fifteenth, was not entirely his fault; the map the Census Bureau provided him (see below) is upside-down from Homestead’s actual geography, although the map indicates that nowhere.  Henry turned left to go west because he assumed the map was oriented normally.[4]

Before Henry realized his mistake, he did two full days of work, persevering through extremely cold weather to traverse steep, icy streets and knock on dozens of doors.  In those initial days, he gathered reasonably accurate and complete information.  That said, cross-outs on these early entries, plus a fair number of households he missed, suggest he struggled somewhat.[5]

It seems census enumerators throughout the county were running into difficulties.  There were the usual issues like people mistaking the strangers at their door for creditors, burglars, or scammers, but ongoing “Red raids” throughout this period, which restarted on the second day of the census with massive, nation-wide arrests, led people to hide from enumerators they mistook as federal agents.  In an anti-union milltown like Homestead, the ongoing suppression of the Great Steel Strike (Sep. 19, 1919–Jan. 8, 1920), which tarred many as Bolsheviks, created an atmosphere of fear that may have particularly hindered Henry. The Homestead canvassers reported the usual trouble with “foreigners”—both language and distrust.  Meanwhile, the weather remained a factor—it made roads and sidewalks harder for travel, and it even caused ink to freeze in fountain pens. One Homestead District canvsasser never showed up for work at all. Across the county, the work was falling behind pace.[7]

Henry had Sunday off, and when he resumed work on Monday, January 5, he essentially had to start from scratch to count the actual residents of ED 144.  Nevertheless, his deadline to complete the work remained unchanged, just 10 days away.

The actual ED 144 was not a particularly difficult district to canvass, largely consisting of wide streets with detached homes where 1-2 families lived. There were a couple alleyways with homes and one diagonal street that broke the grid, but Henry’s map laid all this out clearly.  Monday he knocked on a large number of doors, but Tuesday he lost momentum and knocked on far fewer.  Both days, he produced reasonable results, but the mis-ordering of his pages and mistakes in addresses suggest the work continued to challenge him.[8]

But Henry had cracked.  On Wednesday, he no longer attempted to do the job properly.  The first household he claimed to have visited that day, he fabricated.  A few households on the next block he also fabricated, as well as the last two households of the day.  Were the actual residents not home?  Were they reluctant to speak to him?  Whatever the reason, Henry ended the day with six fakes hidden amongst the twenty-two new households he had recorded.

I can’t help but imagine the long, dark night of Henry’s soul as he lay in bed in his parents’ house that evening, finally warm after another frigid day on the unwelcoming streets of Homestead, wondering whether those six households represented his worst or his best idea.  Had he succumbed to a temptation he now needed to resist, or had he discovered a way out of the miserable census work?  And if he kept fabricating households, could he possibly get away with it?  He knew exactly what he risked from the instruction manual he had received: five years in jail and a $2,000 fine.[9]

Nevertheless, he awoke on Thursday determined to knock on no more doors.  In the coming days, he went on to fabricate 72 more households.  He even placed many of them at addresses that didn’t exist.[10]

To emphasize how absolutely bonkers this undertaking was, let me introduce you to the Hirsch family of 272 West Fifteenth Avenue:

Father Lewis, 60, was married to Ethel, 59, and they had four children ranging in age from 12 to 24.  All were born in Pennsylvania, but Ethel and Lewis were both the children of Yiddish-speaking parents from Russia.  Lewis had his own wholesale shoe store, and his oldest daughter, Gertrude, assisted him as bookkeeper.  They lived between the Mackrells of 260 West Fifteenth, an Irish family whose father worked at a machine works, and the Joneses of 274 West Fifteenth, a Welsh family, whose father worked at the steel mill.  None of these people were real, and there was never a 200 block of West Fifteenth.[11]

The details in Henry’s fabrications are wild.  Archibald Mackrell, the Irish machinist, lived with his brother-in-law, Ezra S. Leslie, an electrician at an electric supply company.  The Welsh Joneses shared their address with the German Huebbers, who lived with their brother- and sister-in-law, Rudolph and Margaret Hertig, who had immigrated in 1910 and naturalized in 1915.  The eldest Huebber son was a “qualified assistant” at a drug store.  Another neighbor on this non-existent block was the manager of a broom factory.  At the bottom of this census page was a boarding house at 305 Sylvan Avenue—although, again, there was no such block—where Bertha Griggs, a 52 year-old widow, kept 6 boarders, single men ranging in age from 23 to 40, mostly native-born, who all performed different kinds of labor at different industrial establishments.[12]

These people represent just a small subset of the 439 people Henry invented out of whole cloth.  Eventually, Henry tired of even this short-cut.  He didn’t bother to feign having canvassed the remainder of ED 144.  He retained the five mistaken pages of people who lived outside ED 144, since they covered up for the homes he had skipped.[13]

The following map visualizes the results of Henry’s work:

 

Note the coverage of the pins within Henry’s assigned ED—in particular, how many blocks have no pins on them at all.  For comparison, this additional map shows the 182 addresses that are represented in the 1918 directory for these blocks.  These pins give a sense of the number of people who never made it into the 1920 census on account of Henry’s fraud.

In order to get away with his deception, Henry not only had to fabricate families, but he also had to fake various kinds of internal documentation.  Every day, he had to submit a report to his supervisor about his daily canvassing activities.  At the end of his work, he had to submit a memorandum of families not enumerated on his first visit, a consolidated time report, and a certificate of completion.  To create these two parallel paper trails, Henry had to have exerted a significant amount of determination.  He dated his work as though he had finished on January 13, two days before the deadline, and submitted 21 pages with all the supporting documentation, hoping no one would notice.[14]

There is a small indication that his supervisor recognized his work as deficient:  an additional page, handwritten by someone else, includes one last household.  The page is prefaced by the handwritten note, “Suppl. – Copied from individual slips (form 9-301) sent by supervisor in Census Office 5/20/20.”  In the full context of the 1920 census, this situation is unusual.[15]

There is also an intriguing note on the side of this additional page:  “Research.”  Had someone in the office done this research, they might have noticed that the address for this family was outside of ED 144.  Had they pulled on that string, they might have noticed that there were five pages of addresses outside of ED 144.  Had they pulled further, they might have noticed that there were 9.25 pages of fake people, nearly half of them at fake addresses.

Or maybe they did pull on this string.  Maybe they also noticed that Anna Malone, who enumerated ED 143, did a poor job as well, skipping homes and visiting a whole block outside of her ED (hence the triplicate census entries noted above).  Perhaps they did discover that there was a gigantic mess in the Third Ward of Homestead and decided the easiest thing to do, five months after the enumeration was completed, was to act like they hadn’t seen anything and just pass it through.[16]

How, then, did I figure all this out a century later?  The string I pulled on was the family of Lewis Hirsch—and the families of Herman Ryave, Jacob Cohan, Philip Huckestein, Joseph Kamin, Morris Plesset, Harry Rosser, Thomas Goldman, and Paul Engelman.  These were the families of ED 144 that I had noted as possibly Jewish, based on their names, professions, countries of origins, and/or Yiddish language.  Of the 50 individuals in these families, I could find none of them in any other records—an unprecedented situation in the 140 enumerations districts I have reviewed in this way.  I noticed on Ancestry.com that for these eight families, there were no records linked to the census entries, no hints for the people once I saved them to my tree, nor any matches from these people in my tree to other people’s trees. In other words, there is no evidence of other genealogists researching these people because they are not real.[17]

From there, I compared every head of household in ED 144 to Homestead’s 1918 and 1921 directories.  A clear line emerged, the Wednesday breakdown I laid out earlier, before which most families appeared in other genealogical records, and after which no families appeared in other records.  Joe Price and his team at BYU’s Record Linking Lab ran a procedure of their own based on their Census Tree Project and confirmed my findings.[18]

So far as I can tell, Henry got away with his fraud.  Afterwards, he decided government work was to his liking.  The 1930 census recorded him as a laborer working for the city.  The 1940 census recorded him as a clerk in the District Attorney’s office.  From 1944 on, he was a clerk in the prothonotary’s office.

He lived at home with his parents for two more decades beyond the 1920 census until he married at 40 years of age.  He never had children.  In 1961 he took his own life—gunshot to the head.  He did not leave a note.

As for the 1920 census, it became mired in controversy.  The enumeration to which Henry contributed showed the continued movement of people to urban areas.  Rural representatives were unhappy at the prospect of losing political power, and for a full decade they blocked the reapportionment of the House of Representatives and the Electoral College.  In 1929 Congress passed a bill to prevent this mess from ever recurring.

And another census mess that never recurred?

Enumerating people in January.

Census Day has been April 1 ever since.[19]

 

Special thanks to my favorite GIS expert, Lauren Winkler, for her help in geocoding Third Ward addresses.  Thank you to Renée Carl and Lina Goldberg for reviewing a draft of this essay.  This essay was updated in November 2022 to clarify the sequence of Henry’s Wednesday breakdown.

---

Notes:

[1] n.b. This article was written before the release of the 1950 census, which, I assure you, I have by now read thoroughly.  The Homestead District includes, variously, Homestead, Munhall, West Homestead, Whitaker, Hays, Lincoln Place, and Mifflin Township.

[2] His 1918 draft card records his employer as the F. F. Dalley Company.  A want ad for the F. F. Dalley Corporations, The Pittsburgh Press, May 9, 1919, explains what his job entailed.  A Facebook post in the Vintage Hamilton group records that F. F. Dalley sold shoe polish, flavoring extracts, patent medicines, dye stuffs, baking powder and spices, oils, and drug sundries.

[3] Pittsburgh newspaper coverage from October-December 1919.

[4] Yes, as a life-long Pittsburgher living north of the Monongahela River, Henry surely should have realized that Homestead spread southward from the river, not northward, but clearly if he been more thoughtful about it—or even looked at a street sign—none of this would have happened.  (To be fair, it is speculation on my part that the map led Henry astray, but that map is pretty confusing!)

The map image is from  Enumeration district maps for the twelfth through sixteenth censuses of the United States, Roll 51 Oregon Lincoln – Yamhill, Pennsylvania Adams – Allegheny (1930, Pittsburgh), 1900-1940.

[5] We can assess the quality of Henry’s work on these first two days because we can compare how these same streets appear in their proper enumeration district, where they also appear.  In fact, because Henry wasn’t the only befuddled Third Ward enumerator, some sections are covered in two other enumeration districts!  From these comparisons we can see that Henry’s work largely aligns with the duplicate and triplicate entries for these same residences.  We can also see that those other enumerators caught 12 households on those blocks that Henry skipped (18.8%).

[6] Instructions to Enumerators, p. 14

[7] Pittsburgh newspaper coverage, January 1920.  Notably, the newspaper suppressed coverage of the strike, but it collapsed while the census was being conducted.  The report of the Homestead canvassers came from The Daily Messenger, 1/8/1920 and 2/9/1920.

[8] To clarify the mis-ordering of pages, he dated pages 4A (front), 4B (back), and 5B (back) as Monday, January 5, but page 5A (front) as Tuesday, January 6.  For addresses, I counted nine places where he mistook the house number (353 W 15th instead of 323 W 15th), street name (1310 Joseph instead of 1310 Sarah), or both (800 West instead of 108 W 14th).

[9] Fourteenth Census of the United State, January 1, 1920:  Instructions to Enumerators, p. 11.  Such arrests did happen.  If he had quit outright, he risked a $500 fine (p. 9).

[10] He both created house numbers that didn’t exist on real blocks and created entire blocks that never existed.  He likely used the map, along with the real house numbers he had canvassed, to make assumptions about the house numbers he had not; however, there were many oddities in house numbering in this part of Homestead that a person could only detect by walking the neighborhood.

[11] The machine works referenced the nearby Mesta Machine Works and the steel mill the Homestead Works of U.S. Steel.  These were the two largest employers of Homestead residents.

A careful review of historic Sanborn maps and city directories confirms there was never a 200 block of West Fifteenth; the numbers have always skipped from 100 to 300.  He did canvass the 300 block of West Fifteenth; perhaps he was trying to represent the 100 block, which he did not canvass.

[12] Sylvan does not extend far enough to have a 300 block.  There is a 1300 block of Sylvan, but it is outside of Henry’s ED.  I believe he misunderstood where W 15th ended, although W 15th’s 300 block didn’t include the house numbers he used.

[13] The other EDs in Ward 3 had 18 or 22 pages.  Henry submitted 21 pages.

[14] This summary of administrative documentation is reconstructed using the Instructions to Enumerators. The Census Bureau did not design a process where supervisors were actually involved in enumerators’ daily work; they tracked their progress remotely by way of the daily reports.  Enumerators had to rely on the mail “for all ordinary purposes,” but “should any emergency arise in which you need immediate counsel and instruction, use the telegraph or telephone” (p. 13).

[15]

Form 9-301, the Individual Census Slip, was left behind for people who were not home during the day and for whom no one else could provide information (“Appendix B: Instructions to Enumerators” from Fourteenth Census of the United States in the Year 1920, Volume II, “Special Instructions to Enumerators in Cities,” pp. 1385, 1387.)

I believe this slip may relate to how Henry realized his mistake.  The couple from this slip were real people who lived at 249 E Fourteenth Avenue, near the edge of the area Henry mistakenly canvassed (it’s the lone top-right pin on the map above).  Henry’s final accounting of ED 144 includes the entire even-numbered side of East Fourteenth, but no one from the odd-numbered side.

It’s unclear how Henry’s supervisor got a hold of this slip—whether Henry mistakenly submitted it, or via some other route—but it brought at least part of ED 144’s problems awfully close to being discovered.

[16] It seems like the local census office became aware of at least one of the issues in Homestead:

There has been one trouble which arose in Homestead only. Some of the enumerators have trespassed on another’s territory. While this trouble has been practically settled…the proper enumerator of the district will have to pay a visit and take the names. So any person who may receive a second visit will understand that it was through this mistake and should overlook the trouble caused them. —The Daily Messenger, 1/8/1920

There is no way to know if they discovered this issue as it pertained to Henry and Anna, whose duplicates remained in their submissions, or in some other enumeration district(s), where the duplicates were omitted before submission.

As the enumeration went on, the county’s census headquarters admitted that they had “[dismissed] a few of the enumerators for indifference and incompetency” (The Pittsburgh Post, January 14, 1920, p. 8). Henry and Anna got to complete their work.

[17] It was also just really weird to find so many Jewish households in this part of town.  In reality there were seven Jewish households in the Third Ward: two in ED 144 (Henry counted one and skipped one) and 5 in ED 143 (Anna counted four and skipped one).  Henry’s additional 8 Jewish-seeming households doubled the count from 7 to 15.  A closer look at the other ethnicities represented might reveal more unusual patterns.

[18] It turns out that the 1921 Homestead directory has its own issue:  the 1921 directory contains a significant number of duplicate entries from the 1918 directory for people who were definitely not in Homestead in 1921. For example, Chaim Lembersky died in November 1918, and Fanella Mervis married and moved away, but both appear in the 1921 directory with entries that are an exact copy of their pre-death/pre-marriage entries from 1918.  This error becomes very apparent with someone like Homer Stern, who was mis-named Omar in the 1918 directory, and therefore shows up twice in the 1921 directory with the copied entry for Omar and a new entry for Homer.  As a result of this duplication, the appearance of a person with consistent information in both the 1918 and 1921 directories is not as meaningful as it would seem.

How prevalent in this problem?  Amongst Homestead’s Jewish residents of this period, 11% of the directory entries are duplicates.  Of the duplicates, 25.6%-38.5% wrongly record a person staying in Homestead who left, and another 25.6% may obscure a person’s change of address within Homestead.

Homestead knew their 1921 directory was a dud. “Long awaited, the directory doesn’t seem to fill that long want because it isn’t correct.  Homestead had the glooms. Polk has handed us a lemon. Oh, death, where is thy sting.”

[19]  The United States Census Bureau: 1920 Overview,  What day was the census taken each decade?  Previous censuses had been June 1or August 1, but “for the 1920 census, the Department of Agriculture presumed they would obtain more accurate information about the value of crops if the census were taken on Jan. 1. They feared farmers would forget financial details over the winter,” but after the census was taken, “representatives of farm states contended that the new Jan. 1 census date meant that many men who spent most of the year working on farms were counted in cities where they spent just a few winter months” (source).