Hi, my name is Sean McKenzie, and I am up here in Anchorage Alaska.
I am a new user. My previous experience is mainly using GAUSS to program, a
PASCAL like language especially suited to matrix operations, and EVIEWS and
its precursor TSP, pretty much Econometrics stuff. I am by background, and
Economist (booo, hiss).
But now I am working on Child Care issues (hip, hip, hurray!), and we
finally bought SPSS, which I actually recommended despite my lack of
knowledge, because I know that it is commonly used in the Social Sciences
arena outside of Economics.
Most of my experience with lists etal are mainly from playing wargames. (So
if any of you are up here in Alaska send me an email at:
email@example.com and may be we can play.)
Given my backround, I believe for me it will be easier to work with matrix
operations, with which I am essentially thoroughly familiar.
I am just starting, and a routine for me in learning a new program is first
to read in data and then spit it back out and confirm that the spit out data
is exactly the same as the spit in.
I did that easily enough.
What I want to do in general is load data from an excel spreadsheet and put
it into a matrix.
So far to me it looks like the easiest way to do this is to upload the data,
save it as an spss.sav matrix file and then upload it again. I am looking
for a more efficient way. To directly upload data from excel as a matrix or
to convert excel data to a matrix without first having to save/re-upload.
My usual second experiment is to upload data, pull out the field names and
then spit those out. Something I could not figure out how to do without
One of the desirable things is for me not to have to explicitly know the
names of my fields or their order prior to uploading data/running a program,
and not to need to know the row or column length of my matrices, as even
when using the "same data" these typically change, even from day to day.
Typically, thinking in GAUSS, the commands are something like
The above is pulling out the field names in row 1 of my Excel SS.
The above is just a test, but in practical terms I would then assign
variable names using the first row of the matrix...since I have perhaps a
100 or coupla hundred variables in my database, I load a matrix of variable
names into the program and using do loops I compare the list of variable
names to the names I uploaded, so when
Variablenames[i,1]=Names[1,i] and suppose the string there was GDP
In the end I will want to load data directly from databse, which is an
access database, but it is just a file on my directory, no password et al.
For now I'll settle for downloading data into excel and then taking it into
SPSS. I can also import my data from text files.
I know that was long, but can any body please help?
As an aside, one of the things I find somewhat difficult to deal with is:
In GAUSS the = sign is an "ASSIGNMENT" operator and works the same for
MATRICES and Scalars. In general in GAUSS there was no special need to
declare something to be a matrix, in general it treated everthing like a
matrix, and a scalar is a [1,1] matrix.
Is there something similar in SPSS, an assignment operator?
Is it possible to operate in matrix style/matrix notation without going
through the whole matrix procedure?