Behavioral Science Flashcards
sensitivity =
=TP/(TP + FN)
=1-FN
*sensitivity rules OUT
specificity =
=TN/(TN + FP)
=1-FP
*specificity rules IN
Positive Predictive Value = PPV =
= TP/(TP + FP)
- PPV = proportion of test results that are true positive
- if increased prevalence, then increased PPV
Negative Predictive Value = NPV =
= TN/(TN+FN)
- proportion of negative test results that are truly negative
- increased prevalence, decreased NPV
Point prevalence =
=total cases in population at a given time/total population at a given time
Prevalence =
=incidence X disease duration
Incidence =
=new cases in popl over a given period of time/total popl at risk during that time period
Odds Ratio = OR =
=(a/b)/(c/d) = ad/bc
*use OR for case-control studies
Relative Risk = RR=
= [a/(a+b)]/[c/(c+d)]
*use RR for cohort studies
Attributable Risk =
= [a/(a+b)] - [c/(c+d)]
*AR is the proportion of disease occurences attributable to exposure to a risk factor
Absolute Risk Reduction
the reduction in risk associated with a treatment as compared to a placebo
Number needed to treat = NNT =
= 1/absolute risk reduction
= 1/[a/(a+b)]
Number needed to harm = NNH =
= 1/attributable risk
Precision, Accuracy, Reliability, Validity, Random error, Systemic error
Precision = Reliability Accuracy = Validity
Random error - reduces precision in a test
Systemic error - reduces accuracy in a test
Standard error of the mean = SEM =
=σ/sqrt of n
used in Normal/Guassian/Bell-Shaped curves (where mean = mode = median)
where:
σ = standard deviation
sqrt of n = square root of sample size
*note: SEM decreases as n (sample size) increases
In a normal/gaussian/bell-curve (where mean=median=mode), what percent of the population is 1 σ (standard deviation) to either side of mean? 2σ to either side of mean? 3σ on either side of mean?
What percent of the population correlates wtih a σ = 1.645 on either side of the mean?
1σ on either side of mean = 68% of popl
2σ = 95%
3σ = 100% (99.7%)
1.645σ = 90% of popl
relationship of mean, median,mode in a positively-skewed statistical distribution?
positive skew: asymmetry with tail on the right
mean > median > mode
relationship of mean, median,mode in a negatively skewed statistical distribution
negative skew - asymmetry with tail on left
mean < median < mode
Null hypothesis = H0 =
hypothesis of no difference; there’s no association between disease and the risk factor in the population
Alternative hypothesis = H1
hypothesis that there is some difference; there is some association between the disease and the risk factor in the population
type 1/alpha error = false positive error
stating there is an effect or difference when none exists; accepting H1 (rejecting H0) when H0 is really true
*ie convicting an innocent man
p value
p = probablity of making a type 1 (alpha) error
ps not actually there)
Type 2/Beta error = False negative error
stating there is not an effect or difference when one exists; not rejecting H0 when it actually is false (so choosing H0 when H1 is true)
*ie setting a guilty man free
Beta
probability of making a type 2 (beta) error