confidence intervals article
what is the most common CI used
95% confidence interval
-if there were 100 CI’s from 100 randomized controlled trials, the true between-group difference would fall within 95 of those
confidence intervals article
90% confidence interval
-less confident
-more precise
-range is smaller
confidence intervals article
99% confidence interval
-more confident
-less precise
-range is larger
other names that mean may be referred to as
point estimate AKA population parameter
KNOW
if given a table with
-mean: 0.4035
-95% confidence interval
-lower: 0.3708
upper: 0.4362
INTERPRET
43.6 upper, 40.3 mean, 37.1 lower
-the mean disability from ____ is 40.35% for patients seeking treatment at ____
-for patients seeking treatment at ____ there is a 95% chance that the population parameter for ____ ranges from 37.08% to 43.62%
statistical significance/clinical meaningfulness article
statistical significance
statistical summary of data would be equal to or more extreme than its observed value
statistical significance/clinical meaningfulness article
clinical meaningfulness
mean difference between groups that is large enough for patients to consider the difference important
probability
the likelihood the described samples came from different populations
-“statistical signifiance”
short answer example of probability/statistical significance
statistically significant differences were observed such that the patients receiving treatment A were better off than patients receiving treatment B by (enter magnitude)
assume NORMAL distribution
95% CI in terms of standard deviation
2 SD away from middle
one-tailed hypothesis
one side of bell-curve
two-tailed hypothesis
BOTH sides of bell-curve
null hypothesis Ho
no difference
alternative hypothesis Ha
there is a difference
alpha
threshold to detect statistical signifiance
-typically 0.05
p-value
index for measuring strength of evidence against null hypothesis
traditional way
- complete study, select statistical test, select alpha (alpha = 0.05)
- p-value is generated
- compare p-value to alpha
-if p < 0.05, REJECT null hypothesis -> there is a difference in ____ for patients receiving treatment A when compared to patients receiving treatment B
-if p > or equal to 0.05, FAIL TO REJECT null hypothesis -> there is no difference in ____ for patients receiving treatment A when compared to patients receiving treatment B
preferred way
steps 1 + 2 are the same as traditional way
1. complete study, select statistical test, select alpha (alpha = 0.05)
2. p-value is generated
3. interpret p-value against null hypothesis
-alpha level does NOT matter
-ex: p-value = 0.15 -> there is a 15% chance there is no difference in ____ for patients receiving treatment A compared to patients receiving treatment B
KNOW
what does a statistically significant finding mean
-statistically significant finding is not likely to have occurred if Ho is true
-does not necessarily mean that results are clinically important
-is not an indication that the difference is large
KNOW
interpretation guidelines-
small magnitude, lower p-value
it depends -> potentially meaningful
KNOW
interpretation guidelines-
small magnitude, higher p-value
no brainer -> not likely to be meaningful
KNOW
interpretation guidelines-
large magnitude, lower p-value
no brainer -> likely meaningful finding
KNOW
interpretation guidelines-
large magnitude, higher p-value
it depends -> potentially meaningful
