P value results

momo

New Member
#1
Dear Colleagues

I did a bivariate regression using 10 independent variables on each dependent variable.
all the results are significant at 95% confidence level, but I have two p values as follows 0.170 and 0.490 for two variables.

so what it could be considered, I mean can I interpret their coefficients or they have no meaning.

Thank you
 

CB

Super Moderator
#2
all the results are significant at 95% confidence level, but I have two p values as follows 0.170 and 0.490 for two variables.
I'm assuming you mean that all the coefficients except for these two IV's are significant.

I wouldn't say that the non-significant coefficients have "no meaning". For these two coefficients you cannot reject the null hypothesis that their values are zero in the population from which your sample is drawn. You can always interpret a coefficient - just make sure your interpretation is correct! (You should not be claiming that there is evidence that these parameters are of a positive value in the population).
 

momo

New Member
#3
I'm assuming you mean that all the coefficients except for these two IV's are significant.

I wouldn't say that the non-significant coefficients have "no meaning". For these two coefficients you cannot reject the null hypothesis that their values are zero in the population from which your sample is drawn. You can always interpret a coefficient - just make sure your interpretation is correct! (You should not be claiming that there is evidence that these parameters are of a positive value in the population).
Thank you very much Cowboybear,

Actually yes, I did the biprobit for three samples, so I have 60 variable coefficient and 60 p value which all are significant, but I have these p values which aren't significant, meanwhile their coefficient convenient economically.

so can I consider that the p values below 0.5 could be interpreted or how can I statistically justify that they are not significant. meanwhile their economic interpretation is OK.

what if their coefficients are positive, (and you said that I can't consider their value equal zero which my sample is drown, what is this mean)

Please more explanation.

In general is the p values 0.108, 0.180, 0.455 and 0.790 are equal because they aren't significant.

All my best
 

CB

Super Moderator
#4
Hi momo,

I'm sorry but I am finding it a bit hard to understand your questions. I think English is not your first language, yes? :)

If you could try to re-write your questions carefully to make them as clear as possible this would be helpful. The grammar check in Word might be helpful. The phrase "meanwhile their coefficient convenient economically" is especially confusing - I have no idea what you're saying here!

Providing an example of the type of interpretation you are wanting to give to a specific coefficient might also help.

what if their coefficients are positive, (and you said that I can't consider their value equal zero which my sample is drown, what is this mean)
Careful - I think this is the opposite of what I said. Maybe re-read my post.

In general is the p values 0.108, 0.180, 0.455 and 0.790 are equal because they aren't significant.
Not really. These p values are all non-significant by the conventional .05 alpha level, so you might treat them in a fairly similar way - but their values and meanings are still different. It might be worth checking your understanding of what a p value means.

All the best :)
CB.
 

momo

New Member
#5
Hi momo,

I'm sorry but I am finding it a bit hard to understand your questions. I think English is not your first language, yes? :)

If you could try to re-write your questions carefully to make them as clear as possible this would be helpful. The grammar check in Word might be helpful. The phrase "meanwhile their coefficient convenient economically" is especially confusing - I have no idea what you're saying here!

Providing an example of the type of interpretation you are wanting to give to a specific coefficient might also help.



Careful - I think this is the opposite of what I said. Maybe re-read my post.



Not really. These p values are all non-significant by the conventional .05 alpha level, so you might treat them in a fairly similar way - but their values and meanings are still different. It might be worth checking your understanding of what a p value means.

All the best :)
CB.
Hi CB,
Firstly thank you for your interesting

Simply I got a final results from the regression that I did, some variables are not significant. But the parameters in line (agree) with the economic theory pre-expectations (theoretical predictions), i.e the theory expect the variable has a positive relationship with the independent variable and the empirical results give the same result but again, unfortunately its not significant, so what should I do:

Interpret the variable although its not significant. Or it has no meaning.

The confusing thing for me that you said:

You can always interpret a coefficient (its OK) - just make sure your interpretation is correct! (You should not be claiming that there is evidence that these parameters are of a positive value in the population).[/B]

if I haven't evidence that the parameter of the positive values in the population and you wouldn't say the parameter has no meaning, on what (statistically) can I relay in the interpretation of the parameter.

or what do you mean of the bold paragraph, I am sorry I didn't understand, could you please explain more.

Thanks
 

CB

Super Moderator
#6
The confusing thing for me that you said:

You can always interpret a coefficient (its OK) - just make sure your interpretation is correct! (You should not be claiming that there is evidence that these parameters are of a positive value in the population).[/B]

if I haven't evidence that the parameter of the positive values in the population and you wouldn't say the parameter has no meaning, on what (statistically) can I relay in the interpretation of the parameter.
Hi momo, thanks for providing more detail.

I think the confusion here may be because we may be thinking of different things when we say "interpret". To me, an "interpretation" is any written description of what a particular statistical finding means. So, to me, it is quite ok to interpret a non-significant coefficient - as long as the interpretation that I give is an accurate description of what the statistics actually mean. Both significant and non-significant results mean something, just different things.

In your case you might say that the coefficient for variable X was positive, as predicted by theory, but that the coefficient was not significant, meaning that you do not have evidence to reject the null hypothesis that this parameter is zero in the populaton.
 

pff

New Member
#7
Hello,
I am new to the forum but since I was looking for some information on the interpretation of regression coefficients on internet, I found this page so I thought it might have been worth trying asking something related to the topic discussed here.
I saw that you talked about the problem of the interpretation of nonsignificant coefficients.
I was interested in the answer the question posted before: "how can I statistically justify that they are not significant"?
I would like to know whether there might be a valid statistical explanation for a coefficient not being significant.
I mean, in my case I have a dummy indipendent variable (coded 0/1). Might it because the number observations coded 1 for that specific explanatory variable are not many in the sample? or it still doesn't depend on this?
Is there any other explanation attributable to the structure of the sample maybe?
or is it just something else?

Ps: I apologize if I am writing here. Shall I start a new thread?
Many thanks