








USMLE Step 1 Forum USMLE Step 1 Discussion Forum: Let's talk about anything related to USMLE Step 1 exam 

Thread Tools  Search this Thread  Display Modes 
#1




Types of scales and Stats tests?
So, there are 4 types of scales in stats.
a. Nominal (categorical) b. Ordinal c. Interval d. Ratio 1. Can somebody explain me what's the difference between them. I mix up ordinal and interval scales as both of them can have different groups in sequence, or not? Any examples which can clear this out. 2. When to use ttest, chi square, one way ANOVA, matched pairs ttest and repeated measures ANOVA? 
#2




Great 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):
1) ttest and ANOVA are basically the same in that they both compare the means. The big difference is that ttest compares 2 means, ANOVA compares the means of of 3 or more groups. 2) 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) Those are the basics. A matched ttest 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. 
The above post was thanked by:  
drnirajmavani (06022011), excellence (05222014), Gulara (09092011), mle guy (05312011), struggle (06022011) 
#3




Quote:
I would like to add: a matchedpairs ttest 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 beforeandafter comparison  when the group is given two different tests at different times 
The above post was thanked by:  
drnirajmavani (06022011), excellence (05222014), Gulara (09092011), mle guy (05312011), struggle (06022011) 


#4




Are we supposed to know how to use Chi square test?
I didnt get any details in the High yield Biostatistics.. 
#5




No we don't have to use the chi square test or compute any other formula. We just have to know that WHEN we have to use them.

#6




Thanks
Thanks a lot heights and bebix.
Now if any can explain the differences between the scales: nominal, ordinal, interval and ratio. Thanks Last edited by mle guy; 05312011 at 11:37 PM. 
The above post was thanked by:  
benzo (06032011) 
#7




Types of Numerical data
Types of Numerical data
As data we consider the result of an experiment. A rough classification is as follows: Nominal data Numbers or text representing unordered categories (e.g., 0=male, 1=female // 0=death, 1=alive // 0=north, 1=south, 2=west, 4=east) Ordinal data Numbers or text representing categories where order counts (e.g., 1=fatal injury, 2=severe injury, 3=moderate injury, etc.) Discrete data: This is numerical data where both ordering and magnitude are important but only whole number values are possible (e.g., Numbers of deaths caused by heart disease (765,156 in 1988) versus suicide (40,368 in 1988). Continuous data Numerical data where any conceivable value is, in theory, attainable (e.g., height, weight, etc.) Hope this help Ref: Principles of Biostatistics. Marcello Pagano 
The above post was thanked by:  
Claus_CU (06022011), drnirajmavani (06022011), hassanahk (06022011), mle guy (06012011), struggle (06022011) 
Tags 
BiostatisticsEpidemiology 
Thread Tools  Search this Thread 
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
NBME tests!  ferdenand  USMLE Step 2 CK Forum  2  10162013 10:22 AM 
Step 2 CK Self Assessment Tests  meanmr.mustard  USMLE Step 2 CK Forum  1  05262011 11:08 AM 
I want details about NBME tests  MedicalDR14  USMLE Step 1 Forum  1  10132010 01:24 PM 
Pay Scales and Immigration Laws  hafu23  IMG Visa Issues  1  05112010 08:47 AM 
