It is important to distinguish 1) the internal representation of a
binary
number and 2) the display format of that number as a
decimal
number. The translation is not exact.
What shows in the variable view is the decimal display format. It does
not have any influence on the internal binary representation.
Art Kendall
Social Research Consultants
P. S. It is usually better to post follow up questions to the list also
so that other who have similar difficulties will be able to see
responses in the archives.
Mani Subramani wrote:
Art
Thanks for responding... I thought of this...so I set trust to width 1,
0 decimals...it is an integer.
Just in case I was mistaken, I did the computation again...after
confirming that trust was an integer...I get the same results for the
frequencies of new the new numeric vars I created again just
now..(test*5 and test+5)..
I am attaching the spss dataset so that you can check for
yourself....or maybe something's wrong at my end.
Thanks a lot again, for your help.
regards
Mani
---------------------------------------------------
Mani Subramani, [hidden email]
IDSc Dept, Carlson School of Management
University of Minnesota
612-624-3522
On Tue, Jun 16, 2009 at 4:09 PM, Art Kendall
<[hidden email]>
wrote:
Are
you sure that TmTrust and CalcTrustAvg are exactly integers.
based on test2 it seems that this is more than the result of the
differences that would arise from the decimal formating of binary
numbers.
increase the number of displayed decimal places for TmTrust
CalcTrustAvg and test. then look at those columns in the data view.
Are there digits in the first 4 or 5 places after the decimal point?
Art Kendall
Social Research Consultants
Mani Subramani wrote:
Folks
This problem is driving me nuts..I can't figure out what is going on.
I created a variable called test to check if two variables I had in my
dataset were the same.
They were...so the value of the test var for all case was 0.
But when I checked the frequency table, this had multiple zeros.
I created test2 and test 5 to see what was going on with test and the
frequency tables for these are really strange.
Here is the code I used: The frequency table is below that. I would
appreciate any inputs.
regards
----------------- Code to create test, text2, test 3.
Compute Test=TmTrust-CalcTrustAvg.
exe.
Compute test=test*1.
exe.
COmpute test2=test*5.
exe.
compute test3=test+5.
exe.
---------------
Statistics
Test test2 test3
N Valid 254 254 254
Missing 0 0 0
Frequency Table
Test
Frequency Percent Valid Percent Cumulative
Percent
Valid 0 9 3.5 3.5 3.5
0 60 23.6 23.6 27.2
0 98 38.6 38.6 65.7
0 75 29.5 29.5 95.3
0 12 4.7 4.7 100.0
Total 254 100.0 100.0
test2
Frequency Percent Valid Percent Cumulative
Percent
Valid -.02 9 3.5 3.5 3.5
-.02 60 23.6 23.6 27.2
.00 98 38.6 38.6 65.7
.02 75 29.5 29.5 95.3
.02 12 4.7 4.7 100.0
Total 254 100.0 100.0
test3
Frequency Percent Valid Percent Cumulative
Percent
Valid 5.00 9 3.5 3.5 3.5
5.00 60 23.6 23.6 27.2
5.00 98 38.6 38.6 65.7
5.00 75 29.5 29.5 95.3
5.00 12 4.7 4.7 100.0
Total 254 100.0 100.0
Locked
