Interpretting data, TPR

[misterwiggles]

When they give you data tables and such without referring to * indicating a significant finding (i.e. probability not due to chance) how do you determine if the results for example, of a treatment drug were effective or not. Some TPR questions love to ask "Which of the following is an accurate interpretation of the data?" but how does one know what a significant or insignificant improvement's are if the numbers in the tables are arbitrary or cannot be given a scale to compare to. D: Or do I just lack to comprehension skills necessary to do this...

---

Members don't see this ad.

 

[El-Rami]

Look for error bars in graphs and asterisks/crosses in the data tables. If you see neither, look for a p value (p<0.05 or p<0.0001), which tells you if it is significant or not. If you see error bars in a graph, look to see if one of the bars is higher/lower than others with its error bar not overlapping with others; that tells you it is significant. Asterisks/crosses are sometimes included in tables when p values aren't necessarily given.

 

[misterwiggles]

well what if a specific graph + table has none of those. Guess it's just intuition after that.

 

[El-Rami]

I've noticed on some of my practice exams (either TPR or Kaplan), there was maybe a graph or two that didn't seem to provide this info. If that's the case, then it should usually be a clear difference between two or more variables. You go with the best information available to you. If they provide stats info that tells you if something is significantly different, look for the associated symbols/p values/error bars, etc. in order to interpret graphs. If not, then go with intuition.