I think it’s important to know these two first
* “Interval variable” values from where we can drive their “Mean”. e.g. measurements of B.P, Weight, Cholesterol levels etc
* “Nominal variable” different categories or “GROUPS” e.g:
* ”GENDER” Men and Women = Single Nominal
* ”GENDER” Men and Women & “RACE” Caucasian and African = Two Nominals
e.g we compare Blood Pressure measurements
*INTERVAL variable = “MEANS” of B.P measurements
*NOMINAL variable = “GROUPS”
So to compare MEANS of B.P. b/w two are more groups we do these tests
e.g. comparing the BP means b/w (Latinos) and (African Americans)
For a t-Test it will be a wrong statement to say: ”comparing B.P means b/w (Latinos), (African Americans) and (Asians)” ===> got 3 groups
Now what’s the difference B/W t-Test and ANOVA
* ANOVA = TWO Groups or more than Two Groups comparison
NOTE:these two will give u the Identical results if comparing “TWO Groups”
* t-Test = 1 Interval and 1 Nominal with TWO Groups
* One-way ANOVA = 1 Interval and 1 Nominal with TWO Groups or more
And for the above reason they rather use One-way ANOVA when more than two groups are there to be compared e.g as above
Comparing B.P means b/w (Latinos), (African Americans) and (Asians) = 3 groups
* Two-way ANOVA: 1 Interval + 2 Nominal
e.g. Gender (Men and Women) *AND Race (Caucasians and African Americans)
——>we do these tests to see If there is a Difference b/w groups ?
t-statistics (t-Test) and f-statistics(ANOVA) will be compared to get a p-value
If p-value = 0.05 or less ——-it is said to be “Statistically Significant’ and we say that there is a Difference in these Groups.
* Chi-square: we compare b/w “TWO NOMINAL” (with ANY # of GROUPS)
(Note; we are not using any Interval variable here , just the NOMINAL data)
best example….”Testing Drug Efficacy”
*** I think this is all u have to do is recognize what kind of data is presented , how many Nominals, How many groups and If Interval variable is used or not? ***
As answered by heigts, usmle guy and bebixGreat questions! I don’t know the answer to them all, but here are a couple basic quick tricks that I use (of course it’s more complicated but I’ve found i can answer most biostats questions this way):
- t-test and ANOVA are basically the same in that they both compare the means. The big difference is that t-test compares 2 means, ANOVA compares the means of of 3 or more groups.
Those are the basics.
- If you are not given means, then use a chi square test. This is used to compare categories (eg. number of people who responded yes/no, percentage of people who do xyz etc)
- A matched t-test is again just looking at the means between groups, except now you are matching variables. For instance, you want to look at the mean drug dose required to treat people who smoked for a mean number of years vs the mean dose in folks who smoked for a different mean number of years. So in this case drug dose and smoking is matched.Nice rule of thumb!
I would like to add: a matched-pairs t-test is used to test whether there is a significant mean difference between two sets of paired data (correlated data). For example:
- Pair of twins
- Where the same people are being measured in before-and-after comparison
- When the group is given two different tests at different times
TYPES OF SCALES (VARIABLES) AND ITS TESTS
Cross-Sectional Study: Chi-Square (x2)
Case-Control Study: Odds-Ratio
Cohort Study: Absolute Risk & Relative Risk