Presentation is loading. Please wait.

Presentation is loading. Please wait.

EVOLUTION FROM EXCEL PIVOT TABLES TO

Similar presentations


Presentation on theme: "EVOLUTION FROM EXCEL PIVOT TABLES TO"— Presentation transcript:

1 EVOLUTION FROM EXCEL PIVOT TABLES TO
UPLOADING SAS-GENERATED FILES FOR IPEDS Robert June and Alexandra Riley, Marquette University Introduction SAS Code – Key Value Pair SAS Code – Fixed Width Data Verification As the workload of our office increased, we looked for ways to do more with less. The IPEDS surveys were one area in which SAS could help us increase: Accuracy – Efficiency – Reproducibility This poster focuses on Part A of the Completions Survey, and presents two possible methods for generating a data upload file: key value pair and fixed width. The general techniques presented can be used to generate data upload files for other IPEDS surveys. The data in the text file are specified by a set of key value pairs For each pair, the “key” is the name of a characteristic and the “value” is the level/category. (e.g. SEX=1) Key value pairs are separated by commas in each row The data in the text file are specified by their position in the columns of the text file Each field in the file has a starting column and a length The values for each field are put in the prescribed column position Important to confirm that data were accurately uploaded to the IPEDS survey. Verify that data appearing in IPEDS corresponds to uploaded data file Key value pair file is easier to read and compare Step 1: Data Preparation Step 1: Data Preparation Create SAS dataset where data fields are re-coded from institutional values to IPEDS specified values according to the IPEDS import specification Part A fields to re-code are: major number, cipcode, award level, gender, race/ethnicity (e.g. gender recoded from M,F to 1,2) Create a generic analysis variable where values all equal 1 Create SAS dataset where data fields are re-coded from institutional values to IPEDS specified values according to the IPEDS import specification Part A fields to recode are: major number, cipcode, and award level (e.g. award level recoded from Bachelor’s to 3) Need to create 18 binary analysis fields corresponding to each combination of gender and ethnicity Evolution of IPEDS at MU Step 2: Data Aggregation The responsibility and efficiency of submitting data to IPEDS has evolved over the years at Marquette. Calculate the number of degrees conferred for each combination of fields defined in Step 1 using the analysis variable (cnt = 1) Use the MEANS procedure in SAS Key holder in Public Affairs Acted as coordinator; each functional area compiled report Process was labor intensive, time consuming, and error prone Key holder in Institutional Research Reports prepared using SPSS with limited use of syntax and Excel pivot tables Data entry performed manually; time consuming; still error prone Adoption of SAS in IR Use SAS to generate aggregate data and upload files, minimizing data entry Significantly less time consuming and error prone Step 2: Data Aggregation PROC MEANS has 18 analysis variables corresponding to those created in Step 1 Resulting dataset will have fewer rows but more columns than key value pair method Structure of PART_A data set generated from PROC MEANS Is it worth it? Costs SAS has a steep learning curve, requiring a significant time investment in training Licensing costs can be significant, especially in year 1 Benefits Significant time savings; upload process takes minutes instead of hours Automated process reduces errors Code can be reused year after year. Step 3: Upload File Generation Step 3: Upload File Generation Create text file from PART_A_FW dataset from Step 2 @n moves the pointer to column n Zw. is a format which pads a number with leading zeros where w is the total width of the variable. Create a text file from the PART_A dataset in Step 2 DATA _NULL_ processes DATA step without creating a dataset FILE statement specifies the location and name of the text file PUT statement writes lines to the file specified in FILE statement as specified by the IPEDS Import specs +(-1) in PUT statement moves the pointer back one, eliminating unnecessary space Degree Data File Clear Path for Implementation The subsequent examples in this poster focus on the Part A of the Completions Survey. The structure of the data used is: Start with Completions. This survey provides the greatest return on investment. Human Resources is very challenging. Use the other surveys to gain experience before attempting this one. Partial implementations work. Missing information can be manually entered on the data entry screens. Read the IPEDS documentation very carefully. Don’t be afraid to call the Help Desk or reach out to other resources. Sample upload file Sample upload file Robert June, Alexandra Riley,


Download ppt "EVOLUTION FROM EXCEL PIVOT TABLES TO"

Similar presentations


Ads by Google