"Inference" under Complete Census

Char

New Member
#1
Is there such a thing as inference under a complete census, or is this a logical nonsequitur?

Follow-up question...
Say I wish to model a censused dependent variable in a general linear model (GLM) with a set of explanatory variables. So, I create a GLM and compute least-square means for a classification variable of interest. What would then be the interpretation of the parameter p-values and the confidence interval around each least square mean?

Any insights are appreciated!

C.
 

TheEcologist

Global Moderator
#2
Is there such a thing as inference under a complete census, or is this a logical nonsequitur?

Follow-up question...
Say I wish to model a censused dependent variable in a general linear model (GLM) with a set of explanatory variables. So, I create a GLM and compute least-square means for a classification variable of interest. What would then be the interpretation of the parameter p-values and the confidence interval around each least square mean?

Any insights are appreciated!

C.
Seeing as no else is “biting”, I think I’ll give your problem a try:

There is no need for statistical inference when the entire population is known. There are no sampling errors in a census because the calculations of values of interest are based on the entire population. Why treat the results of a vote as a poll?

Concerning your follow-up question I think I know why nobody is “biting”, it’s too vague. You might know what you are doing but we don’t, I can guess but then I risk rambling on and wasting time. So please elaborate.