Methods

Analysis of designed experiments
Designed experiments are often employed in basic sciences, medicine and agriculture to study the effects of a particular intervention. Randomization is carefully implemented to balance the observed and unobserved covariates. We have expertise in:

  • Analysis of Variance (ANOVA)
  • Analysis of Covariance (ANCOVA)
  • Factorial design and analysis
  • Nested and multi-level design
  • Blocked design and randomization
  • Sample size calculation and power analysis
  • Multiple comparison adjustments

Multivariate analysis
Multivariate statistical techniques are used increasingly in data mining and analytics to discover patterns and gain insights. We have expertise in the following multivariate methods:

  • Cluster and factor analysis
  • Principal components analysis (PCA)
  • Classification and tree-based methods
  • Generalized additive models
  • Neural and Bayesian networks
  • Multifactor dimensionality reduction (MDR)

Survival data analysis
Survival data are frequently observed in medical studies and industry experiments when time to event is of interest. A unique feature of survival data is that observations can be censored due to lost to followup or early termination of the study. We have extensive experience in many special techniques of analyzing survival data:

  • Nonparametric estimation of survival and hazard functions
  • Compare survival probabilities between different populations
  • Proportional hazards regression model
  • Accelerated life-time models
  • Parametric survival models: Weibull, lognormal, etc.
  • Competing risks data

Clustered and longitudinal data analysis
Many research studies involve clustered data due to study design and sampling methods, such as data from complex surveys, multi-center clinical trials, and repeated measurements. Longitudinal data is a special type of clustered data in which data is collected over time. Specialized statistical techniques are required to deal with clustering or correlation. We have expertise in the following methodologies:

  • Linear Mixed Model (LMM) and Generalized Linear Mixed Model (GLMM)
  • Growth-curve modeling
  • Repeated measures ANOVA
  • Generalized Estimating Equations modeling (GEE)
  • Analysis of complex survey data
  • Cluster/latent class Analysis

Categorical data analysis
Data are often observed in categories rather than continuity in social and medical research. Analysis with categorical data requires specialized methods. We have expertise in:

  • Logistic regression
  • Loglinear models for count data
  • Nonparametric and exact tests: Friedman's test, McNemar's test, Cochran's test, trend test, Wilcoxin signed rank test, etc.
  • Matched pairs modeling and analysis
  • Ordinal and multinomial outcomes modeling

262-720-3397
EuclidixConsulting@gmail.com