Video created: May 7, 2013
While the overall result of an intervention is important, the range of effect is just as critical. That’s where confidence intervals come in. They describe the range of results found in a research study. Understanding a Confidence Interval describes the 95% confidence interval, which tells us that 95 times out of 100, the effect of an intervention will be within the range specified.
Knowing the confidence interval provides a measure of certainty of the effectiveness of an intervention in the real world. To explain, the video describes the effect of a hypothetical social media campaign to prevent cyber-bullying in schools. While the research shows that the campaign did indeed reduce cyber-bullying, the range of results, as determined by the 95% confidence interval, may not be so clear-cut.
The video also explores what it means if the range crosses the ‘line of no difference’. In that case, we should expect that the cyber-bullying could actually increase. Whenever making a decision on whether the results of an intervention warrant the cost and effort of implementing it, it’s important to take a close look at the confidence interval.
The National Collaborating Centre for Methods and Tools is funded by the Public Health Agency of Canada and affiliated with McMaster University. The views expressed herein do not necessarily represent the views of the Public Health Agency of Canada.
NCCMT is one of six National Collaborating Centres (NCCs) for Public Health. The Centres promote and improve the use of scientific research and other knowledge to strengthen public health practices and policies in Canada.
NCCMT - URE - Confidence Interval
Теги
NCCMTClinical SignificanceConfidence IntervalForest PlotOdds RatioStatsStatistical TermsStatisticsEvidence-Informed Public HealthPublic HealthEvidence-Informed PracticeResearch EvidenceUnderstanding ResearchNational Collaborating CentreNursingDobbinsMaureen DobbinsEIDMNumber Needed To TreatNNTP-ValueP Value6s PyramidRelative RiskTypes of ReviewsMaking sense of an SMDstandardized mean differenceSMD