females males Analyses with discrete variables

Slides:



Advertisements
Similar presentations
Chapter 18: The Chi-Square Statistic
Advertisements

Analysis of Categorical Data Nick Jackson University of Southern California Department of Psychology 10/11/
Lecture (11,12) Parameter Estimation of PDF and Fitting a Distribution Function.
Presentation title Date Responder endpoint and continuous endpoint, logistic regression or ANOVA? DSBS 24 OCT 2013 Søren Andersen.
Logistic Regression Psy 524 Ainsworth.
Log-linear and logistic models Generalised linear model ANOVA revisited Log-linear model: Poisson distribution logistic model: Binomial distribution Deviances.
Log-linear and logistic models
1 SOC 3811 Basic Social Statistics. 2 Reminder  Hand in your assignment 5  Remember to pick up your previous homework  Final exam: May 12 th (Saturday),
Handling Categorical Data. Learning Outcomes At the end of this session and with additional reading you will be able to: – Understand when and how to.
Review for Exam 2 Some important themes from Chapters 6-9 Chap. 6. Significance Tests Chap. 7: Comparing Two Groups Chap. 8: Contingency Tables (Categorical.
Chapter 12 Inferential Statistics Gay, Mills, and Airasian
The Chi-Square Test Used when both outcome and exposure variables are binary (dichotomous) or even multichotomous Allows the researcher to calculate a.
The Chi-square Statistic. Goodness of fit 0 This test is used to decide whether there is any difference between the observed (experimental) value and.
1 of 27 PSYC 4310/6310 Advanced Experimental Methods and Statistics © 2013, Michael Kalsher Michael J. Kalsher Department of Cognitive Science Adv. Experimental.
PBG 650 Advanced Plant Breeding
Week 10 Chapter 10 - Hypothesis Testing III : The Analysis of Variance
Lecture 8: Generalized Linear Models for Longitudinal Data.
Multiple Regression and Model Building Chapter 15 Copyright © 2014 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin.
When and why to use Logistic Regression?  The response variable has to be binary or ordinal.  Predictors can be continuous, discrete, or combinations.
Week 5: Logistic regression analysis Overview Questions from last week What is logistic regression analysis? The mathematical model Interpreting the β.
Section Copyright © 2014, 2012, 2010 Pearson Education, Inc. Lecture Slides Elementary Statistics Twelfth Edition and the Triola Statistics Series.
Education 793 Class Notes Presentation 10 Chi-Square Tests and One-Way ANOVA.
Educational Research Chapter 13 Inferential Statistics Gay, Mills, and Airasian 10 th Edition.
CHI SQUARE TESTS.
Multiple Logistic Regression STAT E-150 Statistical Methods.
1 Follow the three R’s: Respect for self, Respect for others and Responsibility for all your actions.
Intermediate Applied Statistics STAT 460 Lecture 20, 11/19/2004 Instructor: Aleksandra (Seša) Slavković TA: Wang Yu
IMPORTANCE OF STATISTICS MR.CHITHRAVEL.V ASST.PROFESSOR ACN.
Chi Square & Correlation
Nonparametric Statistics
Biostatistics Class 3 Probability Distributions 2/15/2000.
Statistics and probability Dr. Khaled Ismael Almghari Phone No:
I. ANOVA revisited & reviewed
The Chi-square Statistic
Nonparametric Statistics
32931 Technology Research Methods Autumn 2017 Quantitative Research Component Topic 4: Bivariate Analysis (Contingency Analysis and Regression Analysis)
Non-Parametric Statistics
Chapter 12 Chi-Square Tests and Nonparametric Tests
Chapter 7. Classification and Prediction
Statistical Modelling
Logistic Regression When and why do we use logistic regression?
Chapter 9: Non-parametric Tests
Advanced Quantitative Techniques
5.1 INTRODUCTORY CHI-SQUARE TEST
Lecture Slides Elementary Statistics Twelfth Edition
INTRODUCTORY STATISTICS FOR CRIMINAL JUSTICE
Correlation – Regression
Hypothesis testing. Chi-square test
CHOOSING A STATISTICAL TEST
Basic Statistics Overview
Statistics for the Social Sciences
Hypothesis Testing Using the Chi Square (χ2) Distribution
Introduction to logistic regression a.k.a. Varbrul
PPA 501 – Analytical Methods in Administration
Data Analysis for Two-Way Tables
Nonparametric Statistics
Introduction to Statistics
The Chi-Square Distribution and Test for Independence
Basic Statistical Terms
Review for Exam 2 Some important themes from Chapters 6-9
Discrete Event Simulation - 4
Hypothesis testing. Chi-square test
Statistical Analysis Chi-Square.
Two-way analysis of variance (ANOVA)
Data Analysis Module: Chi Square
Making Use of Associations Tests
Chapter 18: The Chi-Square Statistic
Karl L. Wuensch Department of Psychology East Carolina University
Hypothesis Testing - Chi Square
Presentation transcript:

females males Analyses with discrete variables Data result from counting, not measuring. Do not forcibly apply methods used for discrete variables! females males One-way requency table, 1 x 2 table, the number of cells in a row does not affect dimensionality. One-way table: we can compare the observed frequency ratio to any ratio we may wish to compare it to, usually 1:1, ours differs from 1:1 st: ( = 19.9; p<0,0001), does not differ from 0.6:0.4 (3:2) ( = 0,59; p=0,44)

females males black white A two-way frequency table: we can study associations between two variables: females males black white let’s start from thinking what does it mean that there is no association.

Calculate marginal distributions: there is no association when cell frequencies are producs of row and cell frequencies

The same with absolute numbers: These are frequencies for the case when there is no association between the variables.

Chi-squared-test in a two-way table looks for the difference from such a table in which there is no association, - tests for a different thing compared to a one-way table.

A two-way frequency table: test for an associaton between values of discrete variables There is a statistically significant association between these variables, ( = 85,0; p<0,0001). The strength of the relationship is characterised by odds ratio: (59/103)/(155/28) = 0.103: the “risk” of a black animal to be female is lower.

No matter which way to look at it because: (a/b)/(c/d) = (a/c)/(b/d) = ad/cb if there is no association, odds ratio equals 1, one figure is sufficient to describe only in the case of a 2x2 table. Naturally, it is not like that that there should be equal numbers in all cells: 200 100 50 25

Assumption of chi-squared test: the expected frequency of any cell should not be below 1, and there should not be more than 20% of cells in which expected frequency is less than 5.

female white female black male black male white An analogous G-test is less sensitive to violation of assumptions, Fisher’s test is not sensitive at all, but Fisher’s test is designed for a very special case, the case when marginal distributions are known in advance: female black female white male black male white only in a 2x2 table.

A three-way frequency table – rectangular box (cuboid); : 37 34 39 43

A three-way frequency table chi-square – is there an association or not, but we want more! Log-linear analysis and interactions:

sex*para: division to parasitised and non-parasitised is not independent of sex; division to females and males is not independent of parasitism; sex*para*year: 1) association between para and year is not independent of sex; 2) association between sex and year is not independent of para; 3) association between sex and pära is not independent of year. There is no distinction between independent and dependent variables! Not frequently needed – too complex sex*year may not be of interest at all!

Usually we are interested only in the values of one variable, we can focus on it, declaring it dependent variable, all the rest are independent variables! When binary (has two values) - logistic regression! P=exp(bx+k)/(1+exp(bx+k)) log(P/(1-P) = bx+k; logit(P) = bx + k. probability body weight

P=exp(bx+k)/(1+exp(bx+k)) interpretation of parameters (y=0.5 is at –k/b):

General linear models (ANOVA, ANCOVA....) and generalized linear model – other distributions, we can do all the same: include several independent variables; include interactions ; nested, repeated, random. but just more recent – not necessarily available; in addition to binary variable, will consider one more – a variable with Poisson distribution.

Values obtained by counting: - bugs on plants; when few – discrete; whan many - continuous; more complicated in between. Poisson distribution: let’s throw grains on chessboard, how many in one field; see the image; small mu: special shape; large mu: approaches normal distribution. characteristic: variance equal to mean; if not: underdispersed or overdispersed, biological reasons.