Book contents
- Frontmatter
- Contents
- Preface
- 1 Introduction
- 2 Interventions
- 3 Evaluating an intervention
- 4 Randomized designs
- 5 Nonrandomized studies
- 6 Statistical analysis of intervention trials
- 7 Methods for adjusting for baseline differences between treatment groups
- 8 Time series analysis
- 9 Special topics
- 10 Research to action
- 11 Conclusion
- Index
- References
6 - Statistical analysis of intervention trials
Published online by Cambridge University Press: 10 May 2010
- Frontmatter
- Contents
- Preface
- 1 Introduction
- 2 Interventions
- 3 Evaluating an intervention
- 4 Randomized designs
- 5 Nonrandomized studies
- 6 Statistical analysis of intervention trials
- 7 Methods for adjusting for baseline differences between treatment groups
- 8 Time series analysis
- 9 Special topics
- 10 Research to action
- 11 Conclusion
- Index
- References
Summary
How do I test whether my intervention has had a statistically significant effect?
Up until this section, I have not explicitly addressed how to test whether the effect of an intervention is statistically significant. We have, however, covered the basic ingredients.
First, the study hypothesis should be stated in the null. To review, the null hypotheses for the three major questions of intervention studies are:
1 There is no change between the pre-intervention and the post-intervention assessment.
2 The change between the pre-intervention and the post-intervention assessment in the intervention group is no greater than the change in the comparison group.
3 The outcome for the intervention group is no different than that of the comparison group.
Next we use statistics to determine the probability that the null hypothesis is correct. If the probability is very low, we consider the alternative hypothesis: that the intervention is associated with a significant change.
To choose the correct statistic you need to identify:
(a) the nature of your outcome variable,
(b) the number of times you have measured the outcome,
(c) the number of study groups,
(d) whether you need a bivariate test or a multivariable test, and
(e) whether your data are longitudinal cohort or serial cross-sectional.
To help you identify the type of outcome variable you have, I have listed them with examples in Table 6.1.
- Type
- Chapter
- Information
- Evaluating Clinical and Public Health InterventionsA Practical Guide to Study Design and Statistics, pp. 73 - 100Publisher: Cambridge University PressPrint publication year: 2010