<-- http://www.pitt.edu/~wpilib/statfaq.html -->
stat-FAQ, December 1998 index
- This is the Combined index to the same material that
existed in two pieces as of March. The articles in
Chapters (June, 1997) have been moved here, in the
same 'chapters' as before, but listed among the larger
amount of Answers and Comments that are in smaller files.
For convenience in downloading, some day I hope to have
a ZIP file including all of the separate files.
The "UNIX directory" Link will give you what the Pitt
Unix-Web server creates for all the files referred to
by this index (and which may be selected by clicking).
It gives information on file size and creation date, in
alphabetical order.
= Unix directory =
List of files in the FAQ: name, date, size
Links to advice and advisors, free data, FAQs, vendors
*--------
- the (really) Frequent Questions -
- (12K) Gauss / normal distribution
- (5K) Variances, N vs N-1 (1995)
- (7K) Mixture of gaussians (1993)
and (3K)– more REFs. Sarle (1997)
- (12K) Stepwise regression: 3 comments
- (2K) Ratio of two Normals (1997) REFs
- (4K) Code for Normal simulation (1997)
- (4K) Finite population formulas? (1997) and
(3K) - more FPC (1995)
- ( 8K) Neural nets – from FAQ by Sarle
- (22K) Cluster analysis - Sarle
- (22K) Bootstrapping - Sarle
- (11K) Compositional data (adds to 100%)
- (2K) Coefficient of variation (1995)
- (4K) Conf Intervals, refs, Steven Goodman (1995)
- (7K) Regression: moderation/mediation (1997)
- (2K) Risk of smoking (1997)
*--------
Topics and questions,
as organized below (more or less)
Some basic considerations
Data: outliers, transformation
Correlation. Reliability: alpha, kappa, ROC
Design and analyses, ANOVA
Regression - varieties
Regression - interpretation
Multivariate variance analyses
Contingency tables
"Ad hoc" testing
Power/sample size; meta-analysis
Factors; clusters
Structural equations methods
Other classifiers, classification
Various, miscellaneous
-
-
*-------- Some basic considerations
- Data entry.
- Data mining? a text. (see Stepwise, for generic warnings.)
- On-line data sources. Shoenfield. (See other Stat Web pages, generally.)
- on Teaching. Ward. (see other WEB sites).
- Dictionaries of terms?
- ... learn about Bayesian theory?
- What's a good random-number generator?
- Moments: definitions.
- Equations for the ? distribution? - Numerical Recipes, online. Normal.
- Using harmonic/geometric means?
- ML Maximum likelihood vs other statistics.
- Why Pearson's test, for contingency table?
- ... know where to round-off numbers?
*-------- Data: outliers, transformation
- ... compute K-S test of normality?
- (Q about the K-S test for normality).
- ... compare skew between samples?
- ... do T-scoring, z scoring?
- Why does the Mann-Whitney require "identically shaped" distributions?
- ANOVA and Unequal variances (In two parts).
- Transformations
*-------- Correlation. Reliability: alpha, kappa, ROC
- ... get an average of correlations?
- ... compare correlations?
- ... compare correlated correlations?
- ... compare correlation matrices?
- ... compare variances in correlated samples?
- ... compare (partial) regressions?
- ... compare two regressions? (Chow test)
- ... partial corr. with a categorical var?
- What is the test on correlation shown with the paired t-test?
- Missing values. Myles.
- Cronbach's coefficient alpha
- Testing Rausch assumptions
- Likert scaling.
- ... compute an ICC with unequal N? (intraclass correlation)
- Whatever happened to Guttman analyses?
- Should I use kappa, other than nominal data?
*-------- Design and analyses, ANOVA
- Observational studies. Kademan. O'Rourke.
- Bonferroni, Holm correction
- ... compute eta-squared?
- ... test for rho=1 ?
- Why is it called Jackknife?
- ... CI for relative risk?
- ... compute CI on annual rates?
- ... compare variables across time?
- ... analyze a survey?
- How good are meta/mega analyses?
*-------- Regression - varieties
*-------- Regression - interpretation
- alpha inflation in stepwise selection
- Propensity scoring (outcomes research).
- Regression, non-nested models
- Time-series , regression. Problems. Raffolovich.
- Why is this R-squared negative? (R^2 in regression through origin).
- ... estimate a break-point?
*-------- Multivariate variance analyses
*-------- Contingency tables
*-------- "Ad hoc" testing
*-------- Power/sample size; meta-analysis
- ... compute POWER? This starts with a formula, before some questions.
- ... compute power for Logistic regression? (several responses)
- ... find power for 2-way ANOVA?
*-------- Factors; clusters
- Cluster analysis texts
- Correspondence analysis
- ... find Procrustean Factor analysis?
- How many clusters? (Sarle)
- What is MDS (and refs)?
- ... choose number of factors?
- Nonlinear canonical corr, etc.
*-------- Structural equations methods
*-------- Other classifiers, classification
- CHAID, CART
- What is CHAID?
- SEM features of stat packs?
- Neural nets - SPSS claims? SAS varieties? weight decay? Sarle.
- Neural-net, weight decay. (refs). Sarle.
*-------- Various, miscellaneous
- Qualitative data analysis
- Control charts
- Circular statistics. Zar
- Case-matching software
- Environmental statistics.
- Linkage analyses (genetics), Where?
- Nonlinear confidence regions. Chandler.
*-------- off the topic
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Document by Rich Ulrich. mailto:wpilib+@pitt.edu
FAQ top. //
Richard Ulrich home page.
This document is the start of the Richard Ulrich stat-FAQ at
http://www.pitt.edu/~wpilib/statfaq.html