A few questions about statistical significance and p values

[deleted647690]

I know that in a graph, when you see error bars, overlap indicates that there is not a statistically significant difference between the data points you are comparing.
With data presented in the form of the attached passage from the P/S section bank, where error is presented as + or - some number indicating the amount of error, can you similarly compare data points?

For example, in the data given, for the measure of "Distractibility task: Impulsivity", cocaine exposure gave a value of 29.5 + or - 12.8, while the control gave a value of 8.8 + or - 13.8
Is there something you can say here based on whether the error for condition overlaps or not?

Or would you just look at the p value? To answer the question attached to this passage which asked about the Impulsivity measure, I just looked at the p value, saw that it was below 0.05, and concluded that the difference between the control and cocaine exposure were therefore significant.

---

Members don't see this ad.

 

[Nugester]

I would look at the p-value. To be considered significant, most choose the value of p<.05. That is the only thing statistically significant. The ranges overlap for the measures so you can't really get much information from that.

 

[aldol16]

Since they gave the p-value, I would look at the p-value only because the p-value is the result of a direct statistical comparison between two numbers. The +/- number here is most likely the standard error of the mean, which is a related but slightly different metric: Standard error: meaning and interpretation | Biochemia Medica

 

[deleted647690]

Since they gave the p-value, I would look at the p-value only because the p-value is the result of a direct statistical comparison between two numbers. The +/- number here is most likely the standard error of the mean, which is a related but slightly different metric: Standard error: meaning and interpretation | Biochemia Medica

Thank you.
Alright, so SEM isn't the same as normal error, and I guess the same rules for SE don't apply to SEM (IE, overlap doesn't indicate an absence of statistical significance)

 

[deleted647690]

I would look at the p-value. To be considered significant, most choose the value of p<.05. That is the only thing statistically significant. The ranges overlap for the measures so you can't really get much information from that.

Thanks!