Choosing and using statistics : a biologist's guide / Calvin Dytham.

Format:

Book

Author/Creator:

Dytham, Calvin

Language:

English

Subjects (All):

Biometry.

Physical Description:

xiii, 248 pages : illustrations ; 25 cm

Edition:

Second edition.

Place of Publication:

Malden, MA : Blackwell Pub., 2003.

Contents:

1 Eight steps to successful data analysis 1

Observations 2

Hypothesis testing 2

P-values 3

Sampling 3

Experiments 4

Statistics 4

Descriptive statistics 4

Tests of difference 5

Tests of relationships 5

Tests for data investigation 6

3 Choosing a test: a key 7

Remember

eight steps to successful data analysis 7

The art of choosing a test 7

A key to assist in your choice of statistical test 8

4 Hypothesis testing, sampling and experimental design 22

Hypothesis testing 22

Acceptable errors 22

P-values 23

Sampling 24

Choice of sample unit 24

Number of sample units 25

Positioning of sample units to achieve a random sample 25

Timing of sampling 26

Experimental design 26

Control 26

Procedural controls 27

Temporal control 27

Experimental control 27

Statistical control 28

Some standard experimental designs 28

5 Statistics, variables and distributions 30

What are statistics? 30

Types of statistics 31

Descriptive statistics 31

Parametric statistics 31

Non-parametric statistics 31

What is a variable? 31

Types of variables or scales of measurement 32

Measurement variables 32

Ranked variables 33

Types of distributions 34

Why do you need to know about distributions? 34

Discrete distributions 34

Poisson distribution 34

Binomial distribution 35

Negative binomial distribution 37

Hypergeometric distribution 37

Continuous distributions 38

Rectangular distribution 38

Normal distribution 38

Describing normal distribution further 39

Transformations 41

t-distribution 43

Confidence intervals 43

Chi-square distribution 44

Exponential distribution 44

Non-parametric 'distributions' 44

Ranking, quartiles and the interquartile range 45

Box and whisker plots 45

6 Descriptive and presentational techniques 46

Displaying data: summarizing a single variable 46

Displaying data: showing the distribution of a single variable 46

Descriptive statistics 49

Statistics of location 49

Statistics of distribution, dispersion or spread 50

Other summary statistics 53

Using the computer packages 53

Displaying data: summarizing two or more variables 57

Displaying data: comparing two variables 58

Associations 58

Trends, predictions and time series 59

Displaying data: comparing more than two variables 63

Associations 63

Multiple trends, time series and predictions 64

7 The tests 1: tests to look at differences 66

Do frequency distributions differ? 66

G-test 66

Chi-square test (X[superscript 2]) 68

Kolmogorov-Smirnov test 77

Anderson-Darling test 79

Do the observations from two groups differ? 80

Paired data 80

Paired t-test 80

Wilcoxon signed ranks test 83

Sign test 85

Unpaired data 89

t-test 89

One-way anova 94

Mann-Whitney U test 101

Do the observations from more than two groups differ? 104

Repeated measures 104

Friedman test (for repeated measures) 104

Repeated measures anova 107

Independent samples 108

One-way anova 109

Post hoc testing: after one-way anova 116

Kruskal-Wallis test 119

Post hoc testing: after the Kruskal-Wallis test 121

There are two independent ways of classifying the data 122

One observation for each factor combination (no replication) 122

Friedman test 122

Two-way anova (without replication) 126

More than one observation for each factor combination (with replication) 133

Interaction 133

Two-way anova (with replication) 137

Scheirer-Ray-Hare test 145

There are more than two independent ways to classify the data 151

Multifactorial testing 151

Three-way anova (without replication) 151

Three-way anova (with replication) 152

Multiway anova 157

Not all classifications are independent 158

Nested or hierarchical designs 159

Two-level nested design anova 160

8 The tests 2: tests to look at relationships 165

Is there a correlation or association between two variables? 165

Observations assigned to categories 165

Chi-square test of association 165

Cramer coefficient of association 172

Phi coefficient of association 172

Observations assigned a value 173

'Standard' correlation (Pearson product-moment correlation) 173

Spearman rank-order correlation 177

Kendall rank-order correlation 179

Regression 181

Is there a 'cause and effect' relationship between two variables? 181

'Standard' linear regression 182

Kendall robust line-fit method 189

Logistic regression 189

Model II regression 192

Polynomial, cubic and quadratic regression 192

Tests for more than two variables 193

Tests of association 193

Correlation 193

Partial correlation 194

Kendall partial rank-order correlation 194

Cause(s) and effect(s) 194

Regression 194

Analysis of covariance (ancova) 195

Multiple regression 198

Stepwise regression 198

Path analysis 199

9 The tests 3: tests for data exploration 200

Observation, inspection and plotting 200

Principal component analysis (PCA) and factor analysis 200

Canonical variate analysis (CVA) 205

Discriminant function analysis 205

Multivariate analysis of variance (manova) 210

Multivariate analysis of covariance (mancova) 212

Cluster analysis 212

Decorana and twinspan 214

10 Symbols and letters used in statistics 216

Greek letters 216

Symbols 216

Upper case letters 217

Lower case letters 218

What if the assumptions are violated? 236

Using a computer 237

Sampling 238

Statistics 238

Displaying the data 239

14 A table of statistical tests 241.

Notes:

Includes index.

ISBN:

1405102438

OCLC:

50006623