Quiz 2 Flashcards
why do we use stats in medicine
identify how common or rare a condition may be
understand how frequent a treatment is used
assess accuracy of a diagnostic test
evaluate how well tx’s work
describe risk factors for a disease
eval how well tx works
eval likelihood that tx will cause side effects
compare costs of diff tx strategies
describe two different types of statistics
descriptive and interential
what are descriptive statistics
they allow us to summarize data so it is easier to intrepret
ex. summarizes performance on an exam
how are descriptive stats commonly represented?
Graphically
Name different types of descriptive stats
means or average medians counts percentages ranges standard deviation
what are inferential statistics
allows researchers to make conclusions about a larger population from a smaller study sample
way of determining how the study results are affected by random error
What is random error
relates to the number of subjects measured. As you increase the number of subjects measured, the amount of random error decreases.
give an example of both descriptive and inferential
56% of polled respondents would vote in favor of the amendment
(plus or minus 5%)
56 is descriptive
plus or minus 5% is inferential
what are terms commonly used in inferential stats?
P-value
Statistical significance
Confidence interval
what are two types of data collected from research?
Continuous or categorical (dichotomous is subtype)
what is categorical data?
limits potential responses to certain categories like:
college major OR Color
What is dichotomous data?
Limits potential responses to two potential responses.
Death (yes/no)
Pregnancy (yes/no)
What is continuous data?
continuous data can take any number within a given range, for example blood pressure, age and pain (0-10)
What types of data are often consistent with averages or means?
continuous data
What are common statistics used to describe center of a data set
Mean or average
median or middle response
What are common statistics used to describe the spread or variabilllity of data?
Range
Innterquartile range
standard deviation
What is the most common way to describe the center of data?
Mean or average
How to calculate the mean of a data set?
Add the values of all observations in the data set
divide the sum by the number of observations
What is the definition of median?
The median represents the exact center point in a data set
How do you calculate the median?
Order all observations from smallest to largest
Choose the middle observation (if there is an even number of observations calculate the mean or average of the two middle values)
when data is symetrical, can mean and median be similar
yes
Median values are not affected by extremes. T/F
True
What is the limitation of the mean?
It is influenced by extreme values and misrepresent data
what are three different types of describing spread and variability?
range
interquartile range
standard deviation
Describe the ‘range’ in observable data?
smallest and largest values observed in data. gives a sense of spread of data
What measure is less sensitive to extreme observations? (the range)
interquartile range
What is the interquartile range?
The middle 50% of the data
How do i calculate the interquartile range?
Order all the observations from smallest to largest
Find the median
divide the data below and above the median in half
What does a small interquartile range indicate?
What does a large inteerquartile range indicate?
most of the data fell close to the median
There was a lot of variability in the data
What is the standard deviation?
What does a small standard deviation signify?
What does a large standard deviation signify?
how far the data are spread out from the mean
average distance from a mean
Small: data tightly grouped around mean
Large: data widely spread out from the mean
how to screen your search
does is have the right P I and O
is this about what I want
read thru abstract
only whether it assess what we want, not whether it supports our hypothosis
what are odds ratio
not intiutive
how strongly a given variable may be associated with the outcome of interest compared to other variables (play horses)
mostly restrospectively sometimes used for prospective studies.
exposure and outcome
What are odds?
probability that this event will occur vs. it will not
.80 80% P- probability
Odds _ prob/1-prob 1:4 .80 1-.80 .8/.2 80%/20% 8/2
prob of it happening over the prob it doesn’t happen
odds ratio
an odds ratio of 1 always means there was no difference between the groups
if the odds ratio is close to 1, then we know there isn’t a very large difference between the two
Statisical significance
If odds are substantially above 1
then the event is more common in the group numerator than denominator
Significant significance
if it crosses 1 it is not statistical significant
what is risk ratio?
the risk of developing in one group vs developing in the other group
when is RR and OR similar?
when a condition is rare
Risk ratio can only be used in what kind of study?
prospective studies only
how to calculcate risk ratio?
54/70
2/28 (Yes/total)
What is a hazard ratio
what is a hazard ratio close to 1
way of looking at probablity
looks at time to event
Death… time to death important to delay death
what is ‘number needed to treat’?
How many people would we have to treat for 1 person to have benefit
1/1000
we would need to treat 1,000 in order for 1 person to be spared the event
Number needed to treat
how many people would need to be on statins in order to prevent 1 person from having a heart attack
what is relative risk?
tell you if a relationship exists between event and outcome
and how great is that relationship
one group must be exposed to risk and outcome
relative risk moves forward
how to calculate an RR
1,000 children get dental ca
