Class Notes (1,100,000)
US (470,000)
Ball State (2,000)
JOUR (300)
Lecture 23

# JOUR 601 Lecture Notes - Lecture 23: Confidence Interval, Relative Risk

Department
Journalism
Course Code
JOUR 601
Professor
Robin Blom
Lecture
23

This preview shows half of the first page. to view the full 2 pages of the document.
Understanding a Confidence Interval
When using research to determine the effectiveness of an intervention, we need to look at
more than just the overall result because it is only half the story.
If an investment broker were to say that you can expect a return on your investment of 5% you
might be likely to invest, however, if the investment broker told you the return could be as low
as 0.05% or as high as 7% you may decide the low end of the scale is too low to warn investing.
The same concepts apply to research evidence. Every study should report and overall effect as
well as a range in which the true effect really lies.
To make informed public health decisions, we need to take into account both the overall effect
of an intervention as well as the range of effect.
The confidence interval tells us what that range is.
When we conduct a study, we usually assemble a small group of people to participate with the
intention of applying what we learn from the small group to the whole population.
However, if we did the same study 100 times using new groups of people to participate, we
would get different results each time.
But which result is the right one to apply to the population?
We can’t say, so what we do instead is look at the full range of results indicated by the 95%
confidence interval.
It tells us that 95 times out of 100 the effect of an intervention will be somewhere within this
range.
The 95% confidence interval gives us a large degree of certainty as to what the effectiveness of
an intervention would be in the real world.
Example:
The results of a study find that schools that adopt the social media campaign report a relative
risk of 0.6 or 40% reduction in cyber bullying.
At first glance, this may seem like a large effect supporting the implementation of social media
campaigns. However, the 95% confidence interval illustrates and shows that the true effect
ranges from a relative risk of 0.9-0.3. Or 10%-70% of cyberbullying if this intervention were to
be implemented. We would need to see that we could see an impact as small as a 10%
reduction or as high as a 70% reduction.