LoadingLoading...

Multivariate Procedures (II) in NCSS

Introduction

Loglinear models techniques are used to analyze the relationships among the factors in a multi-way contingency table. Several model selection techniques, goodness of fit tests, and parameter estimation reports are provided.

Multidimensional scaling is a technique that creates a map displaying the relative positions of a number of objects, given only a table of the distances between them. The program calculates either the metric or the non-metric solution.

Correspondence analysis is a technique for graphically displaying a two-way contingency table by calculating coordinates representing its rows and columns. These coordinates are analogous to factors in a factor analysis.

t_productimg_290

Features

Loglinear Models
> Automatic model selection
> Delta value for zero cells
> Estimated effects
> Goodness of fit statistics
> Hierarchical models
> Likelihood ratio tests
> Pearson chi-square tests
> Residual analysis
> Up to seven factorsMultidimensional Scaling
> Classical (metric) scaling
> Dissimilarity and correlation input
> Eigenvalue report
> Kruskal’s stress statistic
> Non-metric scaling
Correspondence Analysis
> Chi-square contributions
> Column/row profiles
> Coordinates report
> Eigenvalues
> Mass values
> Point quality index
> Supplementary rows
> Variable/profile correlationClick here to see more Multivariate procedures available in NCSS.
129