Planning and analysis of observational studies

Preface......Page 4
Acknowledgments......Page 7
Contents......Page 8
1.1 INTRODUCTION......Page 11
1.2 STRATEGY IN CONTROLLED EXPERIMENTS-SAMPLEDAND TARGET POPULATIONS......Page 13
1.3 THE PRINCIPAL SOURCES OF VARIATION IN THERESPONSES......Page 17
1.4 METHODS OF CONTROL......Page 18
1.5 EFFECTS OF BIAS......Page 22
1.6 SUMMARY......Page 23
REFERENCES......Page 24
2.1 DRAWING CONCLUSIONS FROM DATA......Page 25
2.2 TESTS OF SIGNIFICANCE......Page 28
2.3 CONFIDENCE INTERVALS......Page 30
2.4 SYSTEMATIC DIFFERENCES BETWEEN THE POPULATIONS......Page 32
2.5 THE MODEL WHEN BIAS IS PRESENT......Page 35
2.6 SUMMARY......Page 39
REFERENCES......Page 41
3.1 INTRODUCTION......Page 42
3.2 THE STATEMENT OF OBJECTIVES......Page 43
3 3 THETREATMENTS......Page 45
3.4 MEASUREMENT OF TREATMENT LEVELS FOR INDIVIDUAL PERSONS AND THE EFFECTS OF GROUPING......Page 48
3.5 OTHER POINTS RELATED TO TREATMENTS......Page 50
3.6 CONTROL TREATMENTS......Page 52
3.7 THE RESPONSES......Page 54
3.8. Timing of Measurements......Page 56
3.9 SUMMARY......Page 58
REFERENCES......Page 59
4.1 SAMPLE SIZE IN REJATION TO TESTS OF SIGNIFICANCE......Page 60
4.2 SAMPLE SIZE FOR ESTIMATION......Page 65
4.3 THE EFFECT OF BIAS......Page 68
4.4 MORE COMPLEX COMPARISONS......Page 69
4.5 SAMPLES OF CLUSTERS......Page 71
4.6 PLANS FOR REDUCING NONRESPONSE......Page 75
4.7 RELATIONSHIP BETWEEN SAMPLED AND TARGETPOPULATIONS......Page 78
4.8 PILOT STUDIES AND PRETESTS......Page 80
4.9 THE DEVIL’S ADVOCATE......Page 81
4.10 SUMMARY......Page 82
REFERENCES......Page 83
5.1 CONFOUNDING VARIABLES......Page 84
5.2 MATCHING......Page 88
5.3 THE CONSTRUCTION OF MATCHES......Page 91
5.4 EFFECT OF WITHIN-CLASS MATCHING ON x......Page 94
5.5 EFFECT OF CALIPER MATCHING ON x......Page 97
5.6 EFFECT OF “NEAREST AVAILABLE” MATCHING ON x......Page 98
5.8 EFFECTS ON BIAS OF PI- yz......Page 100
5.9 EFFECT OF MATCHING ON THE VARIANCE OF 7, - 7 2......Page 103
5.10 INTRODUCTION TO STATISTICAL ANALYSIS OFPAIR-MATCHED SAMPLES......Page 104
5.12 SUMMARY......Page 107
REFERENCES......Page 110
6.2 y CONTINUOUS x’s CLASSIFIED......Page 112
6.3 y BINOMIAL x’s CLASSIFIED......Page 115
6.4 TREATMENT DIFFERENCE VARYING FROM CELL TO CELL......Page 116
6.5 y AND x’s QUANTITATIVE ADJUSTMENTS BY REGRESSION(COVARIANCE)......Page 118
6.6 REGRESSION ADJUSTMENTS WITH SOME n’s CLASSIFIED......Page 120
6.7 EFFECT OF REGRESSION ADJUSTMENTS ON BIAS IN j j , - jj2......Page 122
6.8 EFFECT OF CURVATURE ON LINEAR-REGRESSIONADJUSTMENTS......Page 123
6.9 EFFECTIVENESS OF REGRESSION ADJUSTMENTS ONPRECISION......Page 125
6.10 EFFECT OF ERRORS IN THE MEASUREMENT OF X......Page 127
6.11 MATCHING AND ADJUSTMENT COMPARED IN EXPERIMENTS......Page 129
6.12 MATCHING AND ADJUSTMENT COMPARED INOBSERVATIONAL STUDIES......Page 131
Matching on x......Page 133
Regression Adjustment......Page 134
6.14 SUMMARY......Page 136
REFERENCES......Page 138
6.13 A PRELIMINARY TEST OF COMPARABILITY......Page 135
7.2 THE SINGLE GROUP MEASURED AFI\'ER TREATMENT ONLY......Page 140
7.3 THE SINGLE GROUP MEASURED BEFORE AND AFTERTREATMENT......Page 141
7.4 THE SINGLE GROUP SERIES OF MEASUREMENTS BEFOREAND AFTER......Page 142
REFERENCES......Page 148
Index......Page 150