Career Choices : Richness of Opportunities Nell Sedransk National Institute of Statistical Sciences North Carolina State University
Background Academia – Mathematics, Statistics & BioEngineering Departments, Medicine and Cancer Centers Industry – Engineering, R&D Training Research Administration – NSF, NIST, NISS Government Agency – BLS, NCES, NIH&NCI Panels to Assess Research Directions &Proposals - NIH, NIDR, NCI, NSF, BiNational RF
Contexts for Statistics Academia – Teaching, Research, Collaboration; Self- selecting & Self-defining, Dissemination/Publication Industry – Hard Problems in Design & in Analysis; Results Required, Implementation Research Administration – Weighing Impact of Results, Weighing Impact of Application, Publication and/or Deliverables for Implementation Government Agency – BLS, NCES, NIH&NCI, Relevance to Policy Panels to Assess Research Directions &Proposals - NIH, NIDR, NCI, NSF, BiNational RF, Technical Assessment of Potential for Significant Advances
Aspects of Statistics Academia – Classrooms of Students, Self-defined Research Objectives Industry – Applications needing Experiments – Designs, Analysis of Data (designed and observational), Algorithm construction and data representation Research Administration – Large-scale projects, Opportunities to define statistical approaches Government Agency – Data & Questions to answer from Data Bases, New Methodology for acquiring and organizing data (surveys) or modeling potential data Research Enterprises – Survey construction and development, medical/laboratory record analysis
Personal Motivation Statistical/Mathematical Impetus Theoretical (abstract) Insight Application requiring mathematical interpretation Social science theory for relationships in data Potential Impact of (applied field) results Fascination with patterns in data Computational tools Ingenuity with data/relationship representations Working Style Individual Collaborative
Example: New Kinds of Data Trajectories: Speech Production Analysis –“Lights” affixed to Reference Points on Jaw –Repeated Syllables (consonant-vowel patterns; accented / unaccented) –Degree of Lisp Light Tracings –Multiple Cameras and Sensors on Hand –Subject in Wheel-Chair –Extent of Controlled Motion
Example: Image Data Pressure Patterns for Seated Wheel-Chair –Grid of Sensors on Wheel-Chair Pad (Multiple Properties) –Serial Multivariate Readings –Non-informative Regions –Detection of Acute Pressure Areas –Measures of Variability within Session Averaging and Differencing Images –Distinguishing Variability from Significant Change –Characterizing Precursors to Deleterious Change
Criteria for Problems Abstract View –Higher Dimensions, Fewer Constraints –Smallest Problem without Solution –Generalization of Solutions to Smaller Problems –New Context for Fundamental Formulation –Proof of Importance Illustrative Example(s) Comparison Studies (often Simulations)
Criteria for Problems Concrete View –Motivating Context or Potential Impact - Example(s) –Comprehensive Problem Statement –Guarantee of Solution –Promise of Implementation –Match of Example with Statistical Formulation –Proof of Importance Impact of Single-Case Solution Transfer of Methodology for Repeated Application
Criteria for Solutions Abstract View –Model Failures or Constraints –New Kinds of Data –Interweaving Deterministic and Stochastic Models –Communication for Complex or High Dimensional Models –Distances: Statistical Formulation to Application; Efficiency for Approximation –Proof of Importance Example
Criteria for Solutions Concrete View –Motivating Example(s) or Data Set(s) –Information (Precision) Requirement for Solution –Time Constraint for Solution –Cost Constraint for Solution –New Context for Fundamental Formulation –Proof of Importance Impact of Solution to Scientific Problem Transfer of Methodology for Repeated Application