Oberlin Group Library Statistics
Download as PPTX, PDF1 like591 views
This document discusses analyzing 6 years of statistics from the Oberlin Group of liberal arts colleges. It describes transforming the data, checking for outliers, testing correlations between variables, and building a regression model. It finds that high group attendance impacts reference transactions but not circulation. High gate count results in high circulation. Circulation depends on volume count. The document also explores predicting US News rankings based on library-specific data like expenditures, staff salaries, collections, and transactions.
Oberlin Group Library Statistics
- 1. Claremont Colleges LibraryOberlin Group StatisticsJohn McDonaldClaremont Colleges Library
- 2. Claremont Colleges LibraryWhat and whyCheck validity of the measures
- 3. Are we all measuring the same thing?
- 4. Test the reliability of the variables
- 5. Are we all measuring consistently from year to year?In order to:Refine our measures collecting process
- 6. Provide stronger definitions of measuresClaremont Colleges LibraryOberlin Group Data6 years of statistics
- 7. From 2004 through 2009
- 8. All 80 schools represented, though not all years for all schools
- 9. 66 relatively consistent variables collectedClaremont Colleges LibraryTransform dataTransposed Excel spreadsheet
- 12. Add US News Ranking for 2009
- 13. Export to STATA
- 14. Run analysisClaremont Colleges Library
- 15. Claremont Colleges LibraryStatisticsLook for Outliers
- 16. Test Rank order correlation
- 17. Build regression modelClaremont Colleges LibraryClaremont Colleges
- 18. Claremont Colleges LibraryAtlanta University CenterClaremont Colleges
- 19. Claremont Colleges LibraryGroup Transactions (Attendance)Group Transactions
- 20. Claremont Colleges LibraryFaculty FTEUndergraduate FTE
- 22. Claremont Colleges LibrarySmith CollegeCirculation (Faculty)SimmonsFaculty FTE
- 23. Claremont Colleges LibraryHow many electronic subscriptions do you receive?Total Library Expenditures
- 24. Claremont Colleges LibraryRank Order Correlation of Undergraduate FTE
- 25. Claremont Colleges LibraryRank Order Correlation of Total Ejournals
- 26. Claremont Colleges LibraryDoes high group attendance impact reference transactions? [YES] - R-squared = 0.2709 - Group Trans = t= 5.92 P>|t| = 0.000 - Librarian FTE = t= 3.21 P>|t| = 0.000 Or impact circulation (students)? [NO] - R-squared = 0.4145 - Total Volumes = t= 14.42 P>|t| = 0.000 - Undergrad FTE = t= 5.16P>|t| = 0.000 - Total Volume = t=11.47 P>|t| = 0.000
- 27. Claremont Colleges LibraryDoes high gate count result in high circ? - R-squared = 0.2663, - Gate Count = t=8.74 P>|t| = 0.000 - Undergrad FTE = t=3.62 P>|t| = 0.000 Does circ depend on volume count? - R-squared = 0.4539 - Gate Count = t= 6.39 P>|t| = 0.000 - Undergrad FTE = t= 2.16 P>|t| = 0.031 - Total Volume = t=11.47 P>|t| = 0.000
- 28. Claremont Colleges LibraryJust for fun!Predicting US News & World Report rankings
- 31. Claremont Colleges LibraryLooking at just library-specific dataTotal Library Expenditures Professional Staff Salaries Total Collection Volume Count Total Books & Bound Volumes Added Books added by purchase Total journals Print journals EjournalsILL returnables borrowed for you ILL nonreturnables provided to you Circulation (Student)circulation (Faculty)Reference transactions (typical week)Group Transactions Group Transactions (Total attendance)Gate count (Typical week)
- 33. Claremont Colleges LibraryWhat affects our student’s view of us?How much you spend.How much you have.How much you can get.How much you don’t help them.
Editor's Notes
- #4: 66 variables that I could use – I added two more, year of survey and US News rank in 2009. Not all variables reported for each school each year.
- #10: Here’s a great correlation, with few outliers.
- #11: Example of a two highly correlated measures, as we would expect
- #12: This one is more concerning, because each of the 11 dots starting to separate themselves in 2008-09 are different schools.
- #13: Here’s an example of tight clustering, but with some concerns
- #14: Here’s a great example of a variable we clearly didn’t understand.
- #16: Shows slower adoption of ejournals – the rank order correlation from the previous year and following year is good, but two years later, it’s poor.