Presented by Joe Kazemi

- What is CSCAR?
(1) Service unit under the Office of Vice President for Research (OVPR)
(2) Provides statistical assistance to faculty members, researchers, and grad students
(3) Has statisticians and bio-statisticians

- Services provided by CSCAR: free consultation and fee-charged workshops
(1) Study design & methodological consultation
(2) Grant proposal assistance
(3) Statistical analysis advice: email for quick help usually will get reply within a same day
(4) Software training
(5) GIS support
(6) Emotional support

- Examples of services in detail
(1) Proposal & hypothesis

a. Make sure that hypothesis is clearly articulated and statistically testable
b. Calculate sample size (power analysis)
c. Design sampling frame
d. Choose appropriate measurement scales

(2) Data collection & cleaning

a. Create data entry template
b. Choose statistical software package
c. Identify data entry error
d. Perform data manipulation

(3) Data plotting & analysis

a. Pick the best plot to visualize your data
b. Choose appropriate statistical model
c. Perform diagnostics checking (see if your dataset meets the assumptions of model you are trying to employ)

(4) Result interpretation

a. Result interpretation and write-up
b. Incorporate reviewers’ comments

- Workshops (a full list is available on their website)
(1) General knowledge on statistics

a. Statistics review: one-day workshop with lecture & lab session (SPSS)
b. Linear regression
c. Logistic regression

(2) Specific techniques

a. Hierarchical linear regression, linear mixed mode (e.g. for data on clustered group)
b. Structure equation modeling (e.g. for identifying nature of relationship of interest)
c. Survival analysis (e.g. rate of event such as adoption of new technology)
d. Sample size calculation

(3) Qualitative data analysis (for contextual data)

a. NVIivo
b. Text mining on common digital texts
c. Survey data analysis

(4) Software training: SPSS, Stata, SAS, R and GIS

- Workshops are
(1) To teach people to be independent investigators and analysts
(2) Not meant to compete with stat courses provided by other departments
(3) Not for credit
(4) Offered at reduced fees for UM members
(5) Usually each workshop is offered once a year (popular ones may be offered twice a year, in Fall and Winter)

- How to access CSCAR?
(1) Location: 3560 Rackham building
(2) Website: www.umich.edu/~cscar
(3) Time: regular service M-F 9 am – 5 pm, evening service M-Th 7 pm – 9 pm
(4) Call to reserve free consultation – 1 hour/week limits, have a fairly long waiting list (2-3 weeks of waiting)
(5) Remote meetings available for off-campus researchers
(6) Walk-in help available by graduate students from Statistics and Bio-statistics departments
(7) Email

General question on statistics – cscar@umich.edu
S/W specific questions: spss.help@umich.edu, stats.help@umich.edu, sas.help@umich.edu

- Tips for using CSCAR
(1) CSCAR has statisticians at different levels; master-level, PhD-level, and graduate students
(2) As it tends to have a fairly long waiting list for an appointment, the best way to start would be to use email hot-lines.
(3) Emails received from hot-lines go to one of master-level consultants. As CSCAR has maintained an internal wiki for frequently asked questions, it is likely to get an answer to your question quickly unless your question is very unique.