Types of Services
There is no charge for Stat Lab consulting with USC faculty, staff and students. Due to the small size of the staff, and the relatively large number of clients, we do ask that you keep the number of repeat visits on the same problem to the minimum number necessary. At the same time we do not want to limit those who need intensive assistance. In order to accommodate these competing interests, each client will receive 3 meetings free of charge, after which some form of payment may be necessary. The Stat Lab staff member assigned to a problem will provide the client with written recommendations briefly describing the problem, followed by an appropriate statistical solution and the computer code for the solution. We are able to assist with statistical computing in a wide variety of packages, including SAS, SPSS-X, Microsoft Access, Minitab, R, etc.
In the initial interview, we review the current status of the project, discuss data collection and analysis methods and reach an understanding of the specific problems the client would like addressed. The Stat Lab then prepares a project description including an estimate of the effort and cost of each phase of the project (fee schedule available upon request). If the client chooses to proceed with the project, the Stat Lab prepares a proposal package detailing the scope of work and budget, which is forwarded to the University's office of Sponsored Awards Management (SAM). SAM then forwards a contract to the client's appropriate representative for approval. Billing is conducted through the University's Office of Budgets and Contracts.
In order to get the most out of the assistance provided by the lab, the client should come to us as early in the research process as possible. Ideally, interaction with the Stat Lab should begin during the preparation of a research plan and data collection protocol. Often, issues that could present problems in analysis later can be solved during this early stage of the process. However, we are glad to help researchers at any stage of their work.
When you contact the Stat Lab, we will generally respond by the next day. Often, we can set up a meeting within the week, though delays of two weeks are not uncommon. This means that we will likely be unable to provide help if you are seeking immediate statistical assistance. As mentioned above, we encourage clients to contact us early in their thesis process.
In addition to working with students and faculty at the university, we also do contract consulting with industrial, governmental and other private clients.
We have experience in all phases of a research project including:
- experimental design
- data collection
- data entry and management
- data description and presentation
- data analysis
- reports and presentations
We have also accumulated extensive experience with survey projects and can provide help in designing survey instruments, developing survey sampling plans and conducting surveys. Our recent projects have included the following clients:
We developed statistical methods to aid in the spatial and temporal modeling of several analytes in a variety of environmental conditions. We developed spatiotemporal models for analytes in monitoring wells at waste storage and landfill sites and developed tests for trends over time.
We have been retained to provide consulting on miscellaneous freshwater fisheries projects, including growth curve modeling of largemouth bass in SC’s reservoirs, population estimation of anadromous species in the Santee-Cooper lakes, and sample size recommendations for surveys of stocked fish.
We designed and analyzed the annual survey of statewide seat-belt compliance for over ten years. Field surveys in 16 counties are organized and drivers and front-seat passengers are monitored for compliance. The Stat Lab uses stratified multistage sampling and computes Horvitz-Thompson-type estimators and standard errors of compliance rates.
We worked with the Jones Ecological Research Center in Ichauway, GA to analyze fire behavior in longleaf pine savannas. Analysis has included standard spatial methods to characterize fuel "cells" and Bayesian spatiotemporal modeling to analyze fire temperature as a function of fuel characteristics and environmental conditions.
We have analyzed PACT (Palmetto Achievement Challenge Tests) data for SC students to identify test items that distinguish between students with disabilities and the general student population. The DIF (Differential Item Functioning) analyses we conducted will be used to modify individual learning plans for students with disabilities.