Statistical significance (*) on both variables??

[Gavanshir]

Hello, I'm presenting a paper from Circulation and a number of the graphs have * on BOTH bars.

For example, a simple graph showing protein levels as % of WT shows an * on the WT bar (which is of course at 100%) and a * on the KO bar (170%). What does this mean? What is the WT significant towards? Since this is repeated on a number of other graphs and that Circulation is a reputable publication, I doubt that it is a mistake but so far everyone I've asked has been unable to answer why. Thanks for any help.

---

Members don't see this ad.

 

[Myempire1]

What's the article? It should say it somewhere in the legend

 

[Gavanshir]

It does not.

PMID: 17224479

"Sarcoplasmic Reticulum Calcium Overloading in Junctin Deficiency Enhances Contractility but Increases Ventriular Automaticity" By Yuan & Kranias. Thanks.

 

[fantasty]

Without looking at the article:

Do they said what the statistical test is? If it's a chi-square or another test comparing whether or not two proportions are the same, then the test does technically relate to both numbers.

In general, the null hypothesis would be that the proportions are the same: Prop(a) = Prop(b)

And the alternative hypothesis would be that they are significantly different: Prop(a) not = Prop(b)

 

[Gavanshir]

That doesn't really help dantay. They seem to have used the Student t test to compare between groups.