I need some help converting data from the 1980's from the SPSS oldtimers

[Björn Türoque]

Most secondary datasets downloaded off of ICPSR come with a setup file that
labels and imports data into SPSS for you, I have a dataset from the 1980's
that has a spss setup file like I have never seen before. For the life of
me I cant figure out how to get it to work, it talks about having tape as
your storage medium - tape! they haven't used that since the days of big
blue and poliester suits. I will include a snippet below if anyone has seen
this before please let me know how to get this dataset up and running.
Luckily things like varlabels and other commands have not changed, but SPSS
does not seem to get to these commands, it errors out pretty much
immediately. Any help would be greatly appreciated. Just out of curiosity,
when they talk about cards, are they talking about punch cards ala 1971?

Thanks a bunch for you help,
Don

RUN NAME SPSS CONTROL FILE DESCRIPTION
FILE NAME STUDENTS
DATA LIST FIXED

RECDTYPE 1-2 SCHLTYPE 3
SCHOOLID 4-7 STUDNTID 8-9 CASEID 4-9
CENRGN 10 TWINDATA 11 LANGDATA 12
**************** Bunch of variables here**************
EBQT5O 1020 EBQT5P 1021 EBQT5Q 1022
EBQT5R 1023 EBQT5S 1024 EBQT5T 1025


COMMENT THE FOLLOWING INDICATES THE DEVICE USED FOR INPUT
INPUT MEDIUM TAPE
N OF CASES UNKNOWN
VAR LABELS

RECDTYPE RECORD TYPE/
SCHLTYPE SCHOOL SAMPLE TYPE/
SCHOOLID FOUR DIGIT SCHOOL ID/
**************** Bunch of variables labels here**************
EBQT5S FELT VERY TENSE/
EBQT5T FELT NERVOUS OR JITTERY


VALUE LABELS

RECDTYPE (04) STUDENT QUEX/
SCHLTYPE
(0) REGULAR SAMPLE (1) ALTERNATIVE SCHOOLS
(2) CUBAN HISPANIC PUBL (3) OTHER HISPANIC PUBL
(5) REGULAR CATHOLIC (6) BLACK CATHOLIC
**************** Bunch of variables values here**************

READ INPUT DATA
COMMENT INSERT YOUR SPSS PROCEDURE CARDS HERE
FINISH

[Richard Ristow]

Ah, geezerdom!  At 07:55 PM 3/25/2007, Don Asay wrote:

>I have a dataset from the 1980's that has a spss setup file like I
>have never seen before. For the life of me I cant figure out how to
>get it to work, it talks about having tape as your storage medium -
>tape! they haven't used that since the days of big blue and poliester
>suits.

Wellll..... there've been half-inch mag tapes around a *little* later
than that. In fact, you wouldn't believe it, but there are still a few
people living who've used one.

>Just out of curiosity, when they talk about cards, are they talking
>about punch cards ala 1971?

Ah, let me tell you! When I was young, we didn't HAVE any of those
fancy editors. We cut the little holes out of the cards with toenail
clippers, and we stood barefoot in the snow whimpering for our output
while the system crashed and all the CS students got ahead of us in
line with program decks two feet thick...

Yes, that's what they were talking about, and they were use to around
1990.

>If anyone has seen this before please let me know how to get this
>dataset up and running.

OK. I'll write as if your data is in c:\Don\My
Documents\DATA\ICPSR.DAT. And commenting out with '*xx..xx.' means
"skip this one."

*--  Replaced by 'TITLE', not that SPSS pays much .
*--  attention to 'TITLE', either.                .
*xx  RUN NAME SPSS CONTROL FILE DESCRIPTION  xx.


*-- Replaced by file specification on "SAVE" statement .
*xx FILE NAME STUDENTS                       xx.


*-- Note modern-day indenting of continuation lines .
*-- And remove the "Bunch of variables" line, of    .
*-- course.                                         .
DATA LIST file='c:\Don\My Documents\DATA\ICPSR.DAT'
      FIXED    /       /* <<-- Added '/' required */
    RECDTYPE 1-2 SCHLTYPE 3
    SCHOOLID 4-7 STUDNTID 8-9 CASEID 4-9
    CENRGN 10 TWINDATA 11 LANGDATA 12
    **************** Bunch of variables here**************
    EBQT5O 1020 EBQT5P 1021 EBQT5Q 1022
    EBQT5R 1023 EBQT5S 1024 EBQT5T 1025.



*-- You don't need any of the following stuff.  The .
*-- "INPUT MEDIUM" is replaced by the "file="       .
*-- specification on "DATA LIST".                   .
*xx THE FOLLOWING INDICATES THE DEVICE USED FOR INPUT xx.
*xx INPUT MEDIUM TAPE                                 xx.
*xx N OF CASES UNKNOWN                                xx.
*xx FILE NAME STUDENTS                                xx.


*-- Note modern-day indenting of continuation lines .
*-- And labels must now be in quotes.               .
VAR LABELS
    RECDTYPE 'RECORD TYPE'/
    SCHLTYPE 'SCHOOL SAMPLE TYPE'/
    SCHOOLID 'FOUR DIGIT SCHOOL ID'/
    **************** Bunch of variables labels here**************
    EBQT5S 'FELT VERY TENSE'/
    EBQT5T 'FELT NERVOUS OR JITTERY'.

*-- Same as VAR LABELS:                             .
*-- Note modern-day indenting of continuation lines .
*-- And labels must now be in quotes.               .
VALUE LABELS
    RECDTYPE (04) 'STUDENT QUEX'/
    SCHLTYPE
    (0) 'REGULAR SAMPLE' (1) 'ALTERNATIVE SCHOOLS'
    (2) 'CUBAN HISPANIC PUBL' (3) 'OTHER HISPANIC PUBL'
    (5) 'REGULAR CATHOLIC' (6) 'BLACK CATHOLIC'
    **************** Bunch of variables values here**************

*-- Historical note: this filled about the same    .
*-- role as present "BEGIN DATA/END DATA". Not     .
*-- needed if data isn't read from the syntax file .
*xx READ INPUT DATA                              xx.


*xx COMMENT INSERT YOUR SPSS PROCEDURE CARDS HERE xx.
*-- At this point, put any transformation code you  .
*-- want to add and, probably, SAVE the file.       .
*-- Put in SPSS procedures as well, if you like.    .


*xx FINISH                                        xx.
*-- Good luck!  -Richard                            .

[statisticsdoc]

>Just out of curiosity, when they talk about cards, are they talking
>about punch cards ala 1971?

Ah, let me tell you! When I was young, we didn't HAVE any of those
fancy editors. We cut the little holes out of the cards with toenail
clippers, and we stood barefoot in the snow whimpering for our output
while the system crashed and all the CS students got ahead of us in
line with program decks two feet thick...

Yes, that's what they were talking about, and they were use to around
1990.
*-- Good luck!  -Richard

***** /  Fogey Mode ON  / ********* .

Richard - we mere statisticians were pushed aside by the CS students because
we were supposed to take our cards to the card-sorting machine, which would
dutifully sort them into piles according to combinations of two variables.
We mere statisticians would then count our cards, and dutifully compute our
contingency statistics by hand (without a calculator).

Ah, the joys of Crosstabs!

Best,

Steve

P.S.  It might amaze you to know that we even had a machine that would find
the inverse of a 10 by 10 correlation matrix in less than five days.  Well
worth the price of the buggy ride.


***** /  Fogey Mode OFF / ********* .

For more ancient remininscences, visit www.statisticsdoc.com

[Richard Ristow]

At 09:03 PM 3/26/2007, Statisticsdoc wrote:

>>When I was young, we didn't HAVE any of those
>>fancy editors. We stood barefoot in the snow
>>whimpering for our output while the CS students
>>got ahead of us in line with program decks two feet thick...
>
>***** /  Fogey Mode ON  / ********* .
>
>Richard - we mere statisticians were pushed
>aside by the CS students because we mere
>statisticians were supposed to take our cards to
>the card-sorting machine, which would dutifully
>sort them into piles according to combinations of two variables.
>
>Ah, the joys of Crosstabs!

Ah, I thought I could out-fogey just about
anybody; but those methods had nearly gone out,
when I came to the business.

Summer of 1967, US Federal Reserve Board,
Washington, D.C. The Fed had just switched to an
IBM 360 running OS/360, thereby being an early
and frustrated adopter of both. (Windows gives me déjà vu.)

The group I was with (I was a very junior
research assistant) were doing econometrics,
multiple regressions with a few niceties like
Durbin-Watson statistics, using specially-written FORTRAN programs;

I don't know whether anybody at the Fed did
crosstabs, or how, but there was a complete
card-handling room that could have been used for
statistics from card decks.

>We would then count our cards,

I hope you don't mean you had to count the cards
after you sorted them into cells. I recall, the
sorters at least had mechanical counters that
gave the number sorted into each bin.

(Punched cards and sorters were invented by
Herman Hollerith for precisely this.)

>and dutifully compute our contingency statistics
>by hand (without a calculator).

That is pretty bad. Desktop rotary mechanical
calculators were widely available by 1960,
anyway; but they weren't cheap, and maybe your shop had a tight budget.

>P.S.  It might amaze you to know that we even
>had a machine that would find the inverse of a
>10 by 10 correlation matrix in less than five
>days.  Well worth the price of the buggy ride.

It does amaze me. I'd love to know what the
machine was; I love old computing machinery.
Let's see: off the top of my head, Gaussian
elimination for matrix inversion should be about
n^3, so about 1E3 additions/subtractions, perhaps
1E2 divisions. (Can anybody correct me?) That
works out to several minutes per arithmetic
operation; on the other hand, managing a kiloword
or so of storage likely took most of the time.

Did you program it with a plugboard?

>***** /  Fogey Mode OFF / ********* .

And I'll do the same; but it's been fun.

[Björn Türoque]

Just wanted to send a big thank you out to everyone, it worked. In addition
your comments about the past were absolutely hilarious, I am considering
sharing them with my adviser to see if he gets a kick out of them.

The dataset I needed to get up and running consisted of more than 500
variables and 8,000 cases, from what you say of previous computing
techniques, it sounds like something like this would take a week and a half
to get up and running. Just to let you know my "old" 1.2 gigahertz machine
still took almost 10 minutes to import the file, but I was surfing the web
on my wireless broadband connection while the computer was doing all the
work... Just kidding, I don't mean to rub it in.

Thanks again,

Don

On 3/27/07, Richard Ristow <[hidden email]> wrote:

>
> At 09:03 PM 3/26/2007, Statisticsdoc wrote:
>
> >>When I was young, we didn't HAVE any of those
> >>fancy editors. We stood barefoot in the snow
> >>whimpering for our output while the CS students
> >>got ahead of us in line with program decks two feet thick...
> >
> >***** /  Fogey Mode ON  / ********* .
> >
> >Richard - we mere statisticians were pushed
> >aside by the CS students because we mere
> >statisticians were supposed to take our cards to
> >the card-sorting machine, which would dutifully
> >sort them into piles according to combinations of two variables.
> >
> >Ah, the joys of Crosstabs!
>
> Ah, I thought I could out-fogey just about
> anybody; but those methods had nearly gone out,
> when I came to the business.
>
> Summer of 1967, US Federal Reserve Board,
> Washington, D.C. The Fed had just switched to an
> IBM 360 running OS/360, thereby being an early
> and frustrated adopter of both. (Windows gives me déjà vu.)
>
> The group I was with (I was a very junior
> research assistant) were doing econometrics,
> multiple regressions with a few niceties like
> Durbin-Watson statistics, using specially-written FORTRAN programs;
>
> I don't know whether anybody at the Fed did
> crosstabs, or how, but there was a complete
> card-handling room that could have been used for
> statistics from card decks.
>
> >We would then count our cards,
>
> I hope you don't mean you had to count the cards
> after you sorted them into cells. I recall, the
> sorters at least had mechanical counters that
> gave the number sorted into each bin.
>
> (Punched cards and sorters were invented by
> Herman Hollerith for precisely this.)
>
> >and dutifully compute our contingency statistics
> >by hand (without a calculator).
>
> That is pretty bad. Desktop rotary mechanical
> calculators were widely available by 1960,
> anyway; but they weren't cheap, and maybe your shop had a tight budget.
>
> >P.S.  It might amaze you to know that we even
> >had a machine that would find the inverse of a
> >10 by 10 correlation matrix in less than five
> >days.  Well worth the price of the buggy ride.
>
> It does amaze me. I'd love to know what the
> machine was; I love old computing machinery.
> Let's see: off the top of my head, Gaussian
> elimination for matrix inversion should be about
> n^3, so about 1E3 additions/subtractions, perhaps
> 1E2 divisions. (Can anybody correct me?) That
> works out to several minutes per arithmetic
> operation; on the other hand, managing a kiloword
> or so of storage likely took most of the time.
>
> Did you program it with a plugboard?
>
> >***** /  Fogey Mode OFF / ********* .
>
> And I'll do the same; but it's been fun.
>

[Mark A Davenport MADAVENP]

Oh for the advantage of punch cards.  We had to code our data on clay
slabs with a wooden stylus, then stand in line for hours waiting to have
our slabs baked to hardness.  Try carrying a box of those around the
temple.

***************************************************************************************************************************************************************
Mark A. Davenport Ph.D.
Senior Research Analyst
Court of Nebuchadnezzar II
Hanging Gardens University
336.256.0395
[hidden email]

'An approximate answer to the right question is worth a good deal more
than an exact answer to an approximate question.' --a paraphrase of J. W.
Tukey (1962)






Statisticsdoc <[hidden email]>
Sent by: "SPSSX(r) Discussion" <[hidden email]>
03/26/2007 09:03 PM
Please respond to
Statisticsdoc <[hidden email]>


To
[hidden email]
cc

Subject
Re: I need some help converting data from the 1980's from the SPSS
oldtimers






>Just out of curiosity, when they talk about cards, are they talking
>about punch cards ala 1971?

Ah, let me tell you! When I was young, we didn't HAVE any of those
fancy editors. We cut the little holes out of the cards with toenail
clippers, and we stood barefoot in the snow whimpering for our output
while the system crashed and all the CS students got ahead of us in
line with program decks two feet thick...

Yes, that's what they were talking about, and they were use to around
1990.
*-- Good luck!  -Richard

***** /  Fogey Mode ON  / ********* .

Richard - we mere statisticians were pushed aside by the CS students
because
we were supposed to take our cards to the card-sorting machine, which
would
dutifully sort them into piles according to combinations of two variables.
We mere statisticians would then count our cards, and dutifully compute
our
contingency statistics by hand (without a calculator).

Ah, the joys of Crosstabs!

Best,

Steve

P.S.  It might amaze you to know that we even had a machine that would
find
the inverse of a 10 by 10 correlation matrix in less than five days.  Well
worth the price of the buggy ride.


***** /  Fogey Mode OFF / ********* .

For more ancient remininscences, visit www.statisticsdoc.com

[Mark A Davenport MADAVENP]

For you AMOS and Mixed Models users out there:

I am trying to assess pastor/spouse responses on 2 factors (means,
really): Emotional Exhaustion and Depersonalization.  I have fit a
multi-goup (pastor/spouse model in AMOS and get great fit with no post-hoc
modifications needed.  My sense is that the analysis that I really want to
run is not simply a comparison of a pastor group to a spouse group but a
dyad analysis linking a particular pastor to a particular spouse.  Has
anyone set one of these up in AMOS or in Mixed Models before?  I have seen
examples in LISREL and in multilevel software but I read that SEM has
certian advantages over HLM-style modeling in terms of dyad analysis.  Can
anyone point me to an article on the proper set-up.  I just can't seem to
wrap my head around this one.

Mark

***************************************************************************************************************************************************************
Mark A. Davenport Ph.D.
Senior Research Analyst
Office of Institutional Research
The University of North Carolina at Greensboro
336.256.0395
[hidden email]

'An approximate answer to the right question is worth a good deal more
than an exact answer to an approximate question.' --a paraphrase of J. W.
Tukey (1962)

[Stevan Nielsen]

We had to smash hydrogen and oxygen together and wait for the resulting
water to erode the rocks so that we could make clay!

Stevan Lars Nielsen, Ph.D.
Clinical Professor
Clinical Psychologist
2518 WSC, BYU
Provo, UT 84602

801-422-3035; fax [phone]

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Mark A Davenport MADAVENP
Sent: Tuesday, March 27, 2007 9:02 AM
To: [hidden email]
Subject: Re: I need some help converting data from the 1980's from the
SPSS oldtimers

Oh for the advantage of punch cards.  We had to code our data on clay
slabs with a wooden stylus, then stand in line for hours waiting to have
our slabs baked to hardness.  Try carrying a box of those around the
temple.

************************************************************************
************************************************************************
***************
Mark A. Davenport Ph.D.
Senior Research Analyst
Court of Nebuchadnezzar II
Hanging Gardens University
336.256.0395
[hidden email]

'An approximate answer to the right question is worth a good deal more
than an exact answer to an approximate question.' --a paraphrase of J.
W.
Tukey (1962)






Statisticsdoc <[hidden email]>
Sent by: "SPSSX(r) Discussion" <[hidden email]>
03/26/2007 09:03 PM
Please respond to
Statisticsdoc <[hidden email]>


To
[hidden email]
cc

Subject
Re: I need some help converting data from the 1980's from the SPSS
oldtimers






>Just out of curiosity, when they talk about cards, are they talking
>about punch cards ala 1971?

Ah, let me tell you! When I was young, we didn't HAVE any of those
fancy editors. We cut the little holes out of the cards with toenail
clippers, and we stood barefoot in the snow whimpering for our output
while the system crashed and all the CS students got ahead of us in
line with program decks two feet thick...

Yes, that's what they were talking about, and they were use to around
1990.
*-- Good luck!  -Richard

***** /  Fogey Mode ON  / ********* .

Richard - we mere statisticians were pushed aside by the CS students
because
we were supposed to take our cards to the card-sorting machine, which
would
dutifully sort them into piles according to combinations of two
variables.
We mere statisticians would then count our cards, and dutifully compute
our
contingency statistics by hand (without a calculator).

Ah, the joys of Crosstabs!

Best,

Steve

P.S.  It might amaze you to know that we even had a machine that would
find
the inverse of a 10 by 10 correlation matrix in less than five days.
Well
worth the price of the buggy ride.


***** /  Fogey Mode OFF / ********* .

For more ancient remininscences, visit www.statisticsdoc.com

[Marks, Jim]

We used to dream of clay!

We had to quarry our own granite blocks-- using only pine needles-- and
carry them 3 leagues from the pit head (uphill of course).

Then we had to scratch our data into the granite using our fingernails,
while the CS students broke our granite slabs into gravel on our
foreheads.

We called it getting familiar with the data.

--jim

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Mark A Davenport MADAVENP
Sent: Tuesday, March 27, 2007 10:02 AM
To: [hidden email]
Subject: Re: I need some help converting data from the 1980's from the
SPSS oldtimers

Oh for the advantage of punch cards.  We had to code our data on clay
slabs with a wooden stylus, then stand in line for hours waiting to have
our slabs baked to hardness.  Try carrying a box of those around the
temple.

************************************************************************
************************************************************************
***************
Mark A. Davenport Ph.D.
Senior Research Analyst
Court of Nebuchadnezzar II
Hanging Gardens University
336.256.0395
[hidden email]

'An approximate answer to the right question is worth a good deal more
than an exact answer to an approximate question.' --a paraphrase of J.
W.
Tukey (1962)






Statisticsdoc <[hidden email]>
Sent by: "SPSSX(r) Discussion" <[hidden email]>
03/26/2007 09:03 PM
Please respond to
Statisticsdoc <[hidden email]>


To
[hidden email]
cc

Subject
Re: I need some help converting data from the 1980's from the SPSS
oldtimers






>Just out of curiosity, when they talk about cards, are they talking
>about punch cards ala 1971?

Ah, let me tell you! When I was young, we didn't HAVE any of those fancy
editors. We cut the little holes out of the cards with toenail clippers,
and we stood barefoot in the snow whimpering for our output while the
system crashed and all the CS students got ahead of us in line with
program decks two feet thick...

Yes, that's what they were talking about, and they were use to around
1990.
*-- Good luck!  -Richard

***** /  Fogey Mode ON  / ********* .

Richard - we mere statisticians were pushed aside by the CS students
because we were supposed to take our cards to the card-sorting machine,
which would dutifully sort them into piles according to combinations of
two variables.
We mere statisticians would then count our cards, and dutifully compute
our contingency statistics by hand (without a calculator).

Ah, the joys of Crosstabs!

Best,

Steve

P.S.  It might amaze you to know that we even had a machine that would
find the inverse of a 10 by 10 correlation matrix in less than five
days.  Well worth the price of the buggy ride.


***** /  Fogey Mode OFF / ********* .

For more ancient remininscences, visit www.statisticsdoc.com

[Richard Ristow]

At 10:28 AM 3/27/2007, Don Asay wrote:

>Just wanted to send a big thank you out to everyone, it worked.

Great!

>In addition your comments about the past were absolutely hilarious, I
>am considering sharing them with my adviser to see if he gets a kick
>out of them.

For goodness sakes, yes. Advisors get little enough fun.

>The dataset I needed to get up and running consisted of more than 500
>variables and 8,000 cases, from what you say of previous computing
>techniques, it sounds like something like this would take a week and a
>half to get up and running.

By 1980, a good deal less than that. Punch cards were still definitely
in use, but a lot of shops (certainly Brown University) had interactive
accounts where you could create your 'card' decks with an editor, send
them directly to the batch system, and get your 'printed' output back
as a file. Made much, much shorter turn-around than if an operator had
to load the card reader, then unload and separate the printed output.

And your 8 meg dataset (I looked at your record length) was notably big
at that point, but no longer extraordinary.

(From the Federal Reserve days, in 1967: We passed around the word,
with kind of awe, of somebody who'd brought in a dataset on five CASES
of cards. Cards came 2,000 to a box, five boxes to a case, and five
cases made quite a pile - I saw them. But that was
2,000cards*80columns*5boxes*5cases=4megabytes.)

>My "old" 1.2 gigahertz machine still took almost 10 minutes to import
>the file [...]

I'm a tad surprised it took even that long. May have actually been CPU
bound, though that's rare for an SPSS job on a modern system.

But, oh, yes. Most of my time at Brown we were using a 360/67, a fine
machine for its day and a fairly large one. But it had 1 megabyte of
memory - not a misprint, and in fact notably large - and about a 1
microsecond memory cycle, translating to maybe 3 MEGAhertz clock.

I just re-estimated. When I joined Brown - about 1975, remember - we
had something under a gigabyte in mounted disk storage, in 30-megabyte
disk drives each about the size and shape. Now, I have a 1 gigabyte
thumb drive on my key chain.

[statisticsdoc]

Richard,

The underequipped lab I am thinking of was engaged in an underfunded project
on social gerontology back in the early 1970's.

The whizzing matrix inversion machine actually dates from the early days of
computing at the University of Illinois (Urbana-Champaign).  Not quite a
plugboard, but definitely very pre-Silicon!

BTW, in my college days, Psych majors were expected to learn to use
plugboards for control of learning and memory experiments (*gasp*).

Cheers,

Steve

For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: Richard Ristow [mailto:[hidden email]]
Sent: Tuesday, March 27, 2007 12:20 AM
To: Statisticsdoc; [hidden email]
Subject: Re: I need some help converting data from the 1980's from the
SPSS oldtimers


At 09:03 PM 3/26/2007, Statisticsdoc wrote:

>>When I was young, we didn't HAVE any of those
>>fancy editors. We stood barefoot in the snow
>>whimpering for our output while the CS students
>>got ahead of us in line with program decks two feet thick...
>
>***** /  Fogey Mode ON  / ********* .
>
>Richard - we mere statisticians were pushed
>aside by the CS students because we mere
>statisticians were supposed to take our cards to
>the card-sorting machine, which would dutifully
>sort them into piles according to combinations of two variables.
>
>Ah, the joys of Crosstabs!

Ah, I thought I could out-fogey just about
anybody; but those methods had nearly gone out,
when I came to the business.

Summer of 1967, US Federal Reserve Board,
Washington, D.C. The Fed had just switched to an
IBM 360 running OS/360, thereby being an early
and frustrated adopter of both. (Windows gives me déjà vu.)

The group I was with (I was a very junior
research assistant) were doing econometrics,
multiple regressions with a few niceties like
Durbin-Watson statistics, using specially-written FORTRAN programs;

I don't know whether anybody at the Fed did
crosstabs, or how, but there was a complete
card-handling room that could have been used for
statistics from card decks.

>We would then count our cards,

I hope you don't mean you had to count the cards
after you sorted them into cells. I recall, the
sorters at least had mechanical counters that
gave the number sorted into each bin.

(Punched cards and sorters were invented by
Herman Hollerith for precisely this.)

>and dutifully compute our contingency statistics
>by hand (without a calculator).

That is pretty bad. Desktop rotary mechanical
calculators were widely available by 1960,
anyway; but they weren't cheap, and maybe your shop had a tight budget.

>P.S.  It might amaze you to know that we even
>had a machine that would find the inverse of a
>10 by 10 correlation matrix in less than five
>days.  Well worth the price of the buggy ride.

It does amaze me. I'd love to know what the
machine was; I love old computing machinery.
Let's see: off the top of my head, Gaussian
elimination for matrix inversion should be about
n^3, so about 1E3 additions/subtractions, perhaps
1E2 divisions. (Can anybody correct me?) That
works out to several minutes per arithmetic
operation; on the other hand, managing a kiloword
or so of storage likely took most of the time.

Did you program it with a plugboard?

>***** /  Fogey Mode OFF / ********* .

And I'll do the same; but it's been fun.