City of the big disjunctions

« previous post | next post »

Continuing in another connection with the exploration of real-estate listings that I discussed earlier ("Long is good, good is bad, nice is worse, and ! is questionable", 6/12/2013; "Significant (?) relationships everywhere", 6/14/2013), I stumbled on this curious factoid about the use of and and or in's listings for the ten cities I've harvested so far:

So the descriptions in Chicago real-estate listings use "or" 2.4 times as often as New York City listings do; but NYC listings use "and" almost twice as often as Chicago listings.

For those who worry about statistical significance, here are some of the 95% confidence intervals for the rates (estimated by R's binom.confint(…,methods="exact")):

"and" lower "and" upper "or" lower "or" upper
Chicago 20774 21542 4830 5208
Phila 34286 35131 2678 2922
LA 33344 34462 2703 3035
Atlanta 26004 27324 3144 3442
NYC 38847 39438 2056 2197

As usual, I have no idea what this means, except that there can apparently be striking differences in surprising aspects of local linguistic subcultures.



  1. dw said,

    June 29, 2013 @ 6:27 am


  2. Nick Lamb said,

    June 29, 2013 @ 6:49 am

    Factoid: A trivial fact or, sometimes, an untruth which is popularly accepted as fact

    (Imagine a "Why not both?" reaction image here)

  3. Rod Johnson said,

    June 29, 2013 @ 9:45 am

    dw, are you truly unfamiliar with "factoid" (something I would find surprising). Or are you questioning its use here?

  4. Y said,

    June 29, 2013 @ 10:57 am

    Does the Chicago/NY effect evident in a small random sample, of 10-20 ads? If so, can you extract some examples which might show what's going on? Mine the Or and And ore, as it were?

  5. MikeJ said,

    June 29, 2013 @ 1:26 pm

    They just build funny looking houses in Chicago. 3 bedrooms or 2 baths.

  6. suntzuanime said,

    June 29, 2013 @ 9:14 pm

    Note that even in Chicago "and" appears more than four times more often than "or". This is around the same ratio between the words that appears in random web text. So really, there's nothing to explain about Chicago, the anomaly is the other cities.

    My guess is that real-estate-listing-domain text tends to contain a lot of "and"s and few "or"s because it tends to have a lot of lists of features of the property in question. And for whatever reason the real-estate culture in Chicago has evolved to include more narrative text and less dry feature-listing, while New York has evolved in the opposite direction. You could come up with a story about New Yorkers being slaves to efficiency while Chicagoans take the time to deal with one another on a human level, or something.

  7. D.O. said,

    June 30, 2013 @ 2:39 am

    Prof. Liberman, did you include ampersand sign for your counts of and?

    [(myl) No — I'll add ampersand percentages a bit later today.

    There's clearly a locally-variable trade-off in choice of "&" and "and", which will attenuate the local variation in frequency of "and":


  8. D.O. said,

    June 30, 2013 @ 2:48 am

    Also, with ~700 words for average NYC announcements there is an average of 1.4 ors per announcement and for Chicago (~400 words) it's about 2 per. And why would you need to use or in a housing ad? Something like "e-mail or call"? Inquiring minds etc.

  9. James said,

    June 30, 2013 @ 6:26 am

    You wouldn't need to use 'or', but you know, "… ideal for use as nursery, guest bedroom, or study."

    [(myl) Some examples:

    Finished bedroom in basement with large unfinished space great for workshop or workout room.
    Average approval within 58 days or less.
    … bonus room on the main level can be used for a play rm or office …
    Entertain with flair or relax by the pool.
    For more information, contact XXX XXX at 678-XXX-XXX or XXX@EmailAddress

  10. Suburbanbanshee said,

    June 30, 2013 @ 8:48 am

    How much do the listings cost per word?

    How much does it benefit the seller to sound like he doesn't count words?

  11. Michael Carasik said,

    June 30, 2013 @ 10:23 am

    Rather than saying, "I am a Chicagoan," Chicagoans would say, "Am I a Chicagoan, or what?" But it's hard to imagine this usage cropping up too frequently in a real estate ad.

  12. Christian Hege said,

    June 30, 2013 @ 1:42 pm

    To James at 6:26am, and myl's reply: having lived in New York for 15 years, and looked at a number of apartments whose floor plan just shouldn't have existed, I can picture the realtor saying "this could be your x, OR you know, maybe your Y, OR then again…"

    Does it feel like that to you or what?

  13. D.O. said,

    June 30, 2013 @ 4:37 pm

    Another way to slice the data may be not by the city, but by the type — single family/multi-family/townhouse/condo/apartment/land.

  14. Tracy said,

    July 3, 2013 @ 3:37 pm

    My first reaction was that Acceptance and Commitment Therapy had discovered large corpus research. (ACT is known for teaching people, informally at least, to avoid unnecessary disjunctions.)

RSS feed for comments on this post