I don't know how to tell if something is significant or not

#1
I don't know how to tell if a value passes the critical region or not. What does it mean if something is significant at the .01 level? How do I tell? What does it mean for a value to be less than the critical?

And what does a .000 significance level mean? That is on a lot of SPSS outputs and I don't know what it means.
 

trinker

ggplot2orBust
#2
I don't know how to tell if a value passes the critical region or not. What does it mean if something is significant at the .01 level? How do I tell? What does it mean for a value to be less than the critical?
Short answer: if your p value is less than your critical it's sig. Long answer: you need to read around this to gain understanding. Here's a link to a wiki articcle that will get you started (LINK)

And what does a .000 significance level mean? That is on a lot of SPSS outputs and I don't know what it means.
There is no such thing as .000 p value (except in theory). .000 in SPSS is due to rounding, say .00001 becomes the .000
 
#3
On the SPSS output, where do I look to find the P Value? It says Sig. .028 is that the P Value? If an item is less than 0.28 (like 0.15 or 0.26) than is that item significant? And if an item is larger than .028 (like .178 or .272) then those items are not significant?
 

trinker

ggplot2orBust
#4
On the SPSS output, where do I look to find the P Value? It says Sig. .028 is that the P Value?
yes

If an item is less than 0.28 (like 0.15 or 0.26) than is that item significant? And if an item is larger than .028 (like .178 or .272)
You're not comparing the p-value to the .28. The .28 is the p value. You compare that to alpha which is set by the research as her/his acceptable chance of making a type I error. This is pretty key to stats and in pretty much very test you run. I strongly suggest you read about this.
 
#5
4. A broker for a local investment firm has been studying the relationship between increases in the price of gold (X) and her customer’s requests to liquidate stocks (Y). From a data set based on 15 observations, the sample slope was found to be 2.9. If the standard error of the regression slope coefficient is 0.18, is there reason to believe (at the 0.05 significance level) that the slope has changed from its past value of 3.2? CAN SOME ONE TELL ME WHAT I AM SUPPOSE TO DO IN IT
 

Dason

Ambassador to the humans
#6
4. A broker for a local investment firm has been studying the relationship between increases in the price of gold (X) and her customer’s requests to liquidate stocks (Y). From a data set based on 15 observations, the sample slope was found to be 2.9. If the standard error of the regression slope coefficient is 0.18, is there reason to believe (at the 0.05 significance level) that the slope has changed from its past value of 3.2? CAN SOME ONE TELL ME WHAT I AM SUPPOSE TO DO IN IT
Why didn't you make your own thread? WHY ARE WE YELLING? DID YOU READ THE FORUM GUIDELINES? AHHHH.
 
Last edited:

noetsi

Fortran must die
#7
And what does a .000 significance level mean?
That p is less than .001. P never equals 0 (it is not possible).

Unless my memory is bad the slope/divided by the standard error will give you a T value. This in turn leads to a given p value, which is essentially tell you how likely the results you got were due to random chance alone. If p < than .05 that means the slope you found was signficant at the .05 level. If its less than .01 than its signficant at that level. Etc.

Generally if p is less than .05 you can reasonably conclude that the slope you found was "real" (and thus reflects what you would find in the actual population). What the slope means is something you have to interpret substantively not based on statistics. It would appear that that an increase in gold price (the independent variable) leads to a 2.9 unit increase in customer's request o liquidate stock and that this is statistically signficant (not likely to be random error).

Note that 15 cases is way to small to be reasonably sure of anything in regression. You normally want a 100 or more cases at least.
 

trinker

ggplot2orBust
#8
100 or more cases at least.
My opinion is that it depends on your field and number of DVs. In educational research 100 cases is a luxury we rarely see (budgets, restrictions etc.). But I agree n=15 is small.