12-31-2014 01:23 PM
I am new to SAS. My dissertation analysis will compare students’ reading and math scores from 6th and 8th grade. Then I will do the same comparison but with attendance rates. Lastly, I want to control for variables such as race/ethnicity, gender, etc.
I was advised to use linear regression, ANOVA, and ANACOVA. But, I’m having trouble getting started. My first question is how to make data – right now organized into Excel spreadsheets – conformable to SAS. Do I need my data arranged in a particular way, or with specific column headings, so that the SAS program will know what to do when it uploads my data? The spreadsheets also include data about the schools, their districts, district codes, etc – do I need to remove that data from the spreadsheet before uploading it into SAS to run our analysis?
I’ve been reading online resources about SAS, but haven’t found anything that seems applicable to my type of data, analysis, and concerns. If anyone could offer some advice, or perhaps suggest a good website to guide me through this, I’d greatly appreciate it.
12-31-2014 01:29 PM
If you go to the SAS University Edition site, you'll see two links to free SAS courses, including one on statistical analysis in SAS.
These will go through how to perform these analysis in SAS.
Without knowing how your data is currently organized its hard to comment on how you might need to change it.
12-31-2014 02:55 PM
Your first challenge getting data into SAS from Excel is possibly conceptual.
SAS datasets consist of observations (rows) where each variable(column) has the same characteristics such as numeric or character data and relates to the same topic (all values in the column are age or height or first name for example). If your Excel data conforms to that then you've got a good start. But if the data in Excel mixes meanings, such as a School Name above an enrollment count then you'll have lots of issues. It helps if column names are on the first row and contained within a single Excel Cell. Headings that involve 2 cells mean that the second or subsequent stacked cells is treated as data. Also if the headings are identical in different columns then there is an issue as SAS has each variable with a unique name. So you may have names modified from Total to Tota2
If you search this forum you will find that getting data from Excel into SAS in one way or another is possibly the single most discussed topic.
There are several ways to get data from Excel into SAS and some have an issue with numeric codes being transformed into actual numeric values that you may not want. This mainly involves removing leading zeroes in district or codes.
If practical, I would try to have all of the sheets look basically the same if they should have the same data, make sure that each type of record has a unique identifier (and since school names I work with change from year to year getting a single code is required to track across years).
And keep all of that school information.