Biostats (Segars) Flashcards
___ is a research perspective which states there will be no true difference between groups being compared
BIG STAR
null hypothesis
most conservative and commonly utilized
either accept or reject this perspective, based on statistical analysis
How many standard deviations..?
99.7%: __
68%: __
95%: __
3
1
2
positively skewed graph…
which direction is the tail pointing?
relationship of the mean to the median?
tail points to the right
mean higher than median (mean>median)
negatively skewed graph…
which direction is the tail pointing?
relationship of mean to the median?
tail points to the left
mean lower than median (mean
this data type is dichotomous/binary; non-ranked named categories
has NO magnitude/ NO consistency of scale
Nominal
Gender, occupation class, party affiliation, above 5 ft or below 5 ft tall
this data type has ranked categories; non-equal-distance
YES magnitude/ NO consistency of scale
Ordinal
> 3 mos, <3 mos, > 6 mos; strongly disagree, somewhat disagree, somewhat agree, strongly agree
this data type has order, magnitude and equal intervals of scale (units)
YES magnitude/ YES consistency of scale
Interval
living siblings (number), Age (in yrs), Temp, BP, creatinine, weight
this test for interval type data assesses for equal variances between groups
BIG STAR
Levene’s test
1st question to ask to select the correct statistical test…
What DATA LEVEL is being recorded?
a) Does the data have MAGNITUDE? (yes/no)
b) Does the data have a fixed, measurable INTERVAL along the entire scale (yes/no)?
this gets you on the right sheet
2nd question to ask to select the correct statistical test…
What TYPE OF COMPARISON/ASSESSMENT is desired?
1) correlation
2) Regression (prediction/association)
3) Survival comparison (time)
4) Group comparison (most common)
*frequencies/counts/proportions
this type of comparison /assessment provides a QUANTITATIVE measure of the STRENGTH and DIRECTION of a relationship between variables (range from -1.0 to +1.0)
Correlation
Name the CORRELATION test for…
Nominal: ___
Ordinal: ___
Interval: ___
nominal=contingency coefficient
ordinal=spearman correlation
interval=pearson correlation
___ correlation just assesses for LINEAR correlation
BIG STAR
Pearson
this type of comparison/assessment compares the proportion of events over TIME, or TIME-TO events, between groups (ongoing progression)
Survival tests
Survival tests commonly represented by a __ curve
Kaplan-meier
Name the SURVIVAL test for…
Nominal: ___
Ordinal: ___
Interval: ___
nominal=Log-rank
ordinal=Cox-proportional Hazards
interval=Kaplan-meier
all can be represented by a Kaplan-meier curve
this type of comparison/assessment provides a measure of the relationship between variables by allowing the prediction about the dependent, or outcome, variable (DV) knowning the value/category of independent variables (IV)
Regressions
Name the REGRESSION tests for…
Nominal: ___
Ordinal: ___
Interval: ___
nominal=logistic regression
ordinal=multinomial logistic regression
interval=linear regression
3rd question to ask to select the correct statistical test?
HOW MANY GROUPS are being compared?
2 or 3 or more groups
Name the tests for Nominal data, 2 groups of independent data
- (pearson’s) Chi-square
- Fisher’s exact –> greater than 2 groups w/EXPECTED cell count of <5
For nominal data, >3 groups with statistically significant findings, you must perform what analysis to determine which groups are different?
Post-hoc testing –> Bonferroni test of Inequality
Nominal test for 2 groups of paired/related data…
McNemar test
Nominal test for >3 groups of paired/related data
Cochran
Ordinal test for 2 groups of independent data..
Mann-whitney test
ordinal test for >3 groups of independent data..
Kruskal-Wallis test
Post-hoc test for ordinal data, >3 groups
Student-Newman-Keul
Dunnett
Dunn
Ordinal test for 2 groups of paired/related data…
Wilcoxon Signed Rank test
ordinal test for >3 groups of paired/related data…
Friedman test
Interval test for 2 groups of independent data…
Student t-test
Interval test for >3 groups of independent data…
ANOVA (1 DV) or MANOVA (>2 DVs)
Interval test for >3 groups of independent data w/confounders…
ANCOVA or MANCOVA
interval test for 2 groups of paired/related data…
paired t-test
interval test for >3 groups of paired/related data…
Repeated measures ANOVA (1 DV)
OR
Repeated measures MANVOA (>2 DVs)
post-hoc tests for interval data…
Bonferronie Tukey Scheffe Dunn Dunnett Student-Newman-Keul
__ is a CORRELATION TEST showing relationship or agreement between evaluators (consistency of “decisions”, “determinations”)
Kappa Statistic
Kappa interpretation:
__=the observers perfectly “classify” everyone exactly the same way
__=there is no relationship at all between the observers “classifications”, above the agreement that would be expected by chance
__=the observers “classify” everyone exactly the opposite of each other
+1
0
-1
4th question to ask to select the correct statistical test?
Is the data INDEPENDENT or RELATED (PAIRED)?
-Data from the same (paired) or different groups (independent)
__ is also known as alpha error and is a REJECTION of the null hypothesis when it is actually TRUE, and you SHOULD HAVE ACCEPTED IT
Type 1 error
__ is also known as beta error and is NOT REJECTING the null hypothesis when it is actually FALSE, and you SHOULD HAVE REJECTED IT
Type 2 error
The larger the sample size, the greater the likelihood (ability) of detecting a difference if one truly exists. This has an increase in ___
Power –> the ability of a study design to detect a true difference if one truly exists between group-comparisons, and therefore the level of accuracy in correctly accepting/rejecting the Null hypothesis
Add in what with Alpha (type 1) and Beta (type 2) error rates and confidence interval?
BIG STAR
Add in anticipated drop-outs or loss to follow-ups
p-value should be ___ to be statistically significant
<0.05 (less than 5%)
what does a 95% CI really mean?
We are 95% confident that the “true” difference (0) or relationship (1) between the groups is contained within the confidence interval range
how do you know if the CI is NOT SIGNIFICANT?
if CI crosses 1.0 for ratios (OR/RR/HR)
i.e., 0.78-1.01