Simple/silly question

[Flying Penguin]

In studies, they always refer to certain variable values or something. Like: one-sample t (64) = 7.02, p < .0001. This overestimation occurred even though self-ratings of ability were significantly correlated with our measure of actual ability, r (63) = .39, p < .001 or This was true for the first set of self-appraisals, &Beta s(67) = - .40 to - .49, p s < .001, as well as the second, &Beta s(67) = - .41 to - .50, p s < .001

what are t, p, r etc. ? :confused:

I know I should know this but it's been a looooooong time.


Poop. Those "&Beta"s are meant to be Beta symbols.

[chroot]

Can you provide a link to the entire source material? It is possible that this is some kind of notation common in the field from which this quote came from.

- Warren

[Wong]

This sounds like an abstract from a psychological studies. I am not sure about the symbols. But in general in statistics, "r" refers to the "sample correlation coefficient" between two variables. "r" lies in the range from -1 to 1. A positive (negative) value indicates a positive (negative) *linear* correlation between the two variables. Since in studies we can only take a finite number of samples, we can only hope to estimate r using the data that are available to us. In this process of estimation there is likely to be "errors".

As to "p", it is called the "p value" and 1-p is called the "confidence level". It is tied with something called "hypothesis testing" in statistics. I do not see any hypothesis in the piece of writing so I cannot explain it.

As to the one sample t and beta, I think they are some "test statistics" for some hypothesises. But without knowing the hypothesises I cannot say any further.

[uart]

"t(n)" most likely refers to t-distribution of n degrees of freedom and "p" is the probability that the given "t" results could have occured purely by chance given the "null hypothoses". Note the a "t" distribution is a common probability density function that is quite similar to the Normal distribution. The quoted example is a little unusual as "t" distributions with degrees of freedom greater than 30 are rarely used as they are extremely well approximated with the normal distribution.

"r" is a corelation coefficient, and "Beta" is yet another distribution function.

[Flying Penguin]

Thanks guys. That was indeed from a psychological study (http://www.apa.org/journals/psp/psp7761121.html - a most interesting read!), but I've seen similar notation in numerous other places. thanks again.