Effective Experimentation: For Scientists and Technologists

Effective Experimentation: For Scientists and Technologists......Page 2
Contents......Page 8
Preface......Page 14
1.2.1 Develop a better product......Page 16
1.2.3 A complex project......Page 17
1.3 Good design and good analysis......Page 18
2.2 Towards a darker stout......Page 20
2.3 Summary statistics......Page 21
2.4 The normal distribution......Page 22
2.5 How accurate is my mean?......Page 23
2.6 Is the new additive an improvement?......Page 24
2.7 How many trials are needed for an experiment?......Page 26
2.8 Were the aims of the investigation achieved?......Page 27
2.9 Problems......Page 28
3.2 An example: who wears the trousers?......Page 30
3.6 Controlling the participants......Page 31
3.7 The paired design......Page 32
3.9 Problems......Page 34
4.2 An investigation......Page 36
4.3 Limitations of a one-variable-at-a-time experiment......Page 37
4.4 A factorial experiment......Page 40
4.5 Confidence intervals for effect estimates......Page 44
4.7 Were the aims of the investigation achieved?......Page 46
4.8 Problems......Page 47
5.2 An experiment with three variables......Page 50
5.3 The design matrix method......Page 52
5.4 Computation of predicted values......Page 54
5.5 Computation of confidence interval......Page 55
5.6 95% Confidence interval for an effect......Page 56
5.9 Were the aims of the experiment achieved?......Page 57
5.10 A four-variable experiment......Page 58
5.11 Half-normal plots......Page 59
5.13 Problems......Page 61
6.1 Introduction......Page 66
6.2 Obtaining half-fractional designs......Page 67
6.2.2 With defining contrast AC......Page 70
6.3 Design of ½(24) experiment......Page 71
6.4 Analysing a fractional experiment......Page 72
6.6 Did Wheelwright achieve the aims of his experiment?......Page 74
6.7.2 Four variables......Page 75
6.7.3 Five variables or more......Page 76
6.8 Problems......Page 78
7.2 Towards a better oil?......Page 80
7.3 The experiment......Page 81
7.4 An alternative procedure for estimating the residual SD......Page 84
7.7 Analysis of the design......Page 85
7.8 Conclusions from the experiment......Page 87
7.11 How to obtain the designs......Page 88
7.12 Other uses of saturated designs......Page 89
7.13 Problems......Page 90
8.2 Example: keeping quality of sprouts......Page 92
8.3 How good a fit has the line to the data?......Page 93
8.4 Residuals......Page 94
8.6 Correlation coefficient......Page 96
8.9 Confidence intervals for the regression statistics......Page 97
8.10 Assumptions......Page 98
8.11 Problem......Page 99
9.2 An experiment to improve the yield......Page 102
9.4 Selecting the first independent variable......Page 104
9.5 Relationship between yield and weight......Page 106
9.7 Selecting the second independent variable......Page 108
9.8 An alternative model......Page 111
9.9 Limitations to the analysis......Page 112
9.10 Was the experiment successful?......Page 113
9.11 Problems......Page 114
10.2 An example: easing the digestion......Page 116
10.3 Analysis of crushing strength......Page 119
10.4 Analysis of dissolution time......Page 122
10.5 How many levels of a variable should we use in a design?......Page 124
10.7 Problem......Page 125
11.2 An outlier in one variable......Page 128
11.3 Other outlier tests......Page 130
11.4 Outliers in regression......Page 131
11.5 Influential observations......Page 132
11.6 Outliers and influence in multiple regression......Page 133
11.7 What to do after detection?......Page 135
12.2 An example: design the crunchiness......Page 136
12.4 Estimating the effects......Page 138
12.5 Using multiple regression......Page 140
12.6 Second stage of the design......Page 141
12.9 Critique of central composite designs......Page 145
13.2 Mixtures of two components......Page 148
13.3 A concrete case study......Page 150
13.4 Design and analysis for a 3-component mixture......Page 151
13.5 Designs with mixture variables and process variables......Page 155
13.6 Fractional experiments......Page 157
13.7 Was the experiment successful?......Page 158
14.2 How it works......Page 160
14.3 An example......Page 161
14.5 Selecting the model and number of trials......Page 162
14.6 How the program chooses the design set......Page 163
14.8 Problems......Page 167
15.2 The principles behind EVOP......Page 170
15.3 EVOP: The experimental design......Page 171
15.5 The principles of simplex optimization......Page 175
15.6 Simplex optimization: an experiment......Page 176
15.7.5 Amount of judgement required......Page 179
15.7.6 Coping with nuisance variables......Page 180
16.1 Introduction......Page 182
16.2 Was the experiment successful?......Page 186
17.2 An example: which is the best treatment?......Page 188
17.3 Analysis of variance......Page 189
17.4 Multiple comparison test......Page 192
17.6 Cochran’s test for standard deviations......Page 193
17.9 Problems......Page 194
18.2 An example: kill the sweat......Page 198
18.3 Analysis of the data......Page 199
18.5 Was the experiment successful?......Page 201
18.8 The Latin square design......Page 202
18.9 Latin square analysis of variance......Page 204
18.10 Properties and assumptions of the Latin square design......Page 205
18.12 Was the experiment successful?......Page 206
18.14 Problem......Page 207
19.2 An example: improving the taste of coffee......Page 210
19.3 Two-way analysis of variance......Page 211
19.4 Multiple comparison test......Page 214
19.5 Was the experiment successful?......Page 215
19.6 Problem......Page 216
20.2 Example: an incomplete block experiment......Page 218
20.3 Adjusted means......Page 220
20.4 Analysis of variance for balanced incomplete block design......Page 221
20.6 A design with a control......Page 224
20.8 Problem......Page 225
21.2 23 ways of messing up an experiment......Page 228
21.3 Initial thoughts when planning an experiment......Page 229
21.5 Designing the experiment......Page 230
21.7 Analysing the data......Page 231
21.8 Summary......Page 232
Solutions to problems......Page 234
Statistical tables......Page 260
Index......Page 272