Research Objectives, Data Characteristics & Types of Analyses
Key:
M = measurement (Mi = independent; Md = dependent)
O = ordinal (Oi = independent)
C = classification (Cd = dependent; (2) = two values; (>2) more than two values)
Data Description
Summarize many observations by expressing their statistical distribution according to a particular model. In some cases, model parameters are produced. In others, tables of summary values or plots are given. Some statistics are available that indicate the quality of the fit between the data and the model.
variable type type of analysis
M parametric statistics (mean, std dev); frequency analysis
O frequency analysis; median; interquartile ranges
C frequency analysis; mode
Differentiating Groups
Use either the model description (for parametric distributions) or frequency distributions (for non-parametric distributions) to identify groups that have significant differences in their measurement or frequency values.
variable type type of analysis
C(2) + C(2) frequency analysis; chi-square test
C(>2) + C contingency coefficient
C(2) + M F-test & t-test
C(>2) + Md Duncan-Waller; analysis of variance
C+C+M analysis of variance
Cd+Mi+Mi+ discriminant analysis
Mi+Mi+Mi+ cluster analysis
Measures of Association
How much of the variability between observations can be accounted for by the values of the variables.
variable type type of analysis
Oi + Oi Spearman's rho
Md+Mi regression analysis
Mi+ Mi correlation analysis
Md + Mi + Mi + multiple regression analysis
Mi + Mi + Mi + multiple correlation; factor analysis; duster analysis
Cd + Mi + Mi + discriminant analysis
Md+Md+Mi+Mi + canonical correlation