Question by LibraryFan: Question about SPSS regarding numeric/string?
I have a question regarding the SPSS program. Let me explain my usage of it. As part of my job, I do data entry of surveys and send our client an Excel and an SPSS file of the same data. I do not have any training in computer software, nor was I given any in-house training at work regarding how to extensively work with SPSS. My usage of it has basically been following the example of how other people at my job have used the SPSS program.
I do my data entry in Access, then perform a regular export to Excel. In transferring data to SPSS, I just set up my SPSS form to match my Access query using the variables tab, then copy and paste the data from Access to SPSS (again, this is just how it’s been done; a better way probably exists).
My question is regarding the rounding up of numbers that occurs whenever you put a “.5” in a cell that has a “numeric” type. It automatically rounds the half up to the next whole number unless you include a decimal place. For this client, we have never included a decimal place, as most of the data is the “excellent to poor” rating scale. One question, however, regards a number of hours. Some survey respondents reply with halves (e.g., 6.5 hours). As the data entry operator, I have been rounding these up because I knew that the SPSS would automatically round them. Lately, though, our client has been on our backs about complete, accurate adherence to what has been recorded on the paper surveys. I believe I addressed this several years ago to a supervisor, and they said to round them, but I’m not sure if this came from them or from a communication with the client. I’m actually kind of nervous about broaching the subject with the client, as they’ve been touchy, had an issue with my data entry a few months ago and I don’t want them to lose faith in our company. It’s just this one question about hours that’s worrying me, but I don’t want it to bite us.
Anyway, what I’d like to ask someone more in the know about SPSS is whether I could change the data type for that question to “String” and have that be fine. It would allow me to enter single digits as well as halves. I do not know what our client does with the data once they get it, how they process it. Does the data type in the SPSS form have to be “Numeric” for it to process as a number, or can data entered as “String” also process as a number? I’m asking this because I’m wondering whether I can just change the data type without telling our client and have it process on their end without any hitch. Can someone give me an answer to this?
Answer by Jon KP
1) SPSS does not round numbers when importing. If your Access field contains a decimal or has a decimal setting, that should be preserved in SPSS. After importing, go to the Data Editor Variable View and see how many decimals are set for your numeric variables. There is a different between the displayed values and the full precision values that are used for computation.
2) If you import a string datatype, SPSS will create a string variable. If the contents are numeric, it can be converted to a number in a variety of ways, but if the values are numeric, it is better to use a numeric type in the database.
3) SPSS can read Excel files directly, so unless you are using Access to restructure the data in some way, you might as well import it directly into SPSS from Excel.
What do you think? Answer below!