Aug 17, 2014 14:04
9 yrs ago
3 viewers *
Spanish term
igualar la probabilidad de equilibrio de error con la de acierto de error
Spanish to English
Medical
Mathematics & Statistics
Sample size
Hi,
I am unsure as to what this means. I think it's to do with type I and type II error in determining simple size but I am not enough of an expert to really understand what it means. Are there any experts out there?
The heading is "Cálculo de tamaño muestral" for a clinical research study. The text under the heading reads:
Para determinar el tamaño de la muestra que representa la población en estudio, es necesario emplear la ecuación h = n / (P2 (n - 1) +1), la cual resulta de igualar la probabilidad de equilibrio de error con la probabilidad de acierto de error.
h = tamaño muestral
n= número de sujetos que constituye la población de estudio
P= probabilidad de error de la muestra
I am unsure as to what this means. I think it's to do with type I and type II error in determining simple size but I am not enough of an expert to really understand what it means. Are there any experts out there?
The heading is "Cálculo de tamaño muestral" for a clinical research study. The text under the heading reads:
Para determinar el tamaño de la muestra que representa la población en estudio, es necesario emplear la ecuación h = n / (P2 (n - 1) +1), la cual resulta de igualar la probabilidad de equilibrio de error con la probabilidad de acierto de error.
h = tamaño muestral
n= número de sujetos que constituye la población de estudio
P= probabilidad de error de la muestra
Change log
Aug 17, 2014 15:27: Alison Imms changed "Field (specific)" from "Medical (general)" to "Mathematics & Statistics" , "Field (write-in)" from "Sampling / probability" to "Sample size "
Proposed translations
1 hr
by balancing the margin of error and the confidence level
Declined
It'll probably be accompanied by some version of the formula in the link. Quite possibly a simplified one, since this has the feel of being aimed at a fairly general audience.
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:24:02 GMT)
--------------------------------------------------
Just noticed the formula! Yes, that's a very special case of the general formula given in the link.
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:30:38 GMT)
--------------------------------------------------
Assumes equal proportions in the population, 95% confidence interval. Rounds off a bit.
Fine if the assumptions actually hold!
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:55:55 GMT)
--------------------------------------------------
Perhaps a better way of putting it is "by trading-off the required level of confidence and the precision (as measured by the width of the confidence interval)" See e.g. discussion (especially fourth paragraph from the end) in https://www.utdallas.edu/~ammann/stat3355/node28.html
There's a trade-off between level of confidence and precision, increase one and you decrease the other (for a fixed sample size).
The formula derives from an error bound calculation, I think "igualar" is being used in a general way here rather than the more usual "equate".
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:24:02 GMT)
--------------------------------------------------
Just noticed the formula! Yes, that's a very special case of the general formula given in the link.
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:30:38 GMT)
--------------------------------------------------
Assumes equal proportions in the population, 95% confidence interval. Rounds off a bit.
Fine if the assumptions actually hold!
--------------------------------------------------
Note added at 1 hr (2014-08-17 15:55:55 GMT)
--------------------------------------------------
Perhaps a better way of putting it is "by trading-off the required level of confidence and the precision (as measured by the width of the confidence interval)" See e.g. discussion (especially fourth paragraph from the end) in https://www.utdallas.edu/~ammann/stat3355/node28.html
There's a trade-off between level of confidence and precision, increase one and you decrease the other (for a fixed sample size).
The formula derives from an error bound calculation, I think "igualar" is being used in a general way here rather than the more usual "equate".
19 hrs
equating the probability of balance of error with the probability of certainty of error
Declined
I think it should be translated literally. 'Balance of error' is a statistical term. See the following article:
https://blogs.oracle.com/darcy/entry/balance_of_error
--------------------------------------------------
Note added at 1 day6 hrs (2014-08-18 20:18:32 GMT)
--------------------------------------------------
Given all the doubts, it might be better to use looser language such as 'corresponds to' / 'is considered to be the same as'
Thus:
la cual resulta de igualar la probabilidad de equilibrio de error con la probabilidad de acierto de error
translates as:
'where the probability of balance of error corresponds to/is considered the same as the probability of certainty of error'
https://blogs.oracle.com/darcy/entry/balance_of_error
--------------------------------------------------
Note added at 1 day6 hrs (2014-08-18 20:18:32 GMT)
--------------------------------------------------
Given all the doubts, it might be better to use looser language such as 'corresponds to' / 'is considered to be the same as'
Thus:
la cual resulta de igualar la probabilidad de equilibrio de error con la probabilidad de acierto de error
translates as:
'where the probability of balance of error corresponds to/is considered the same as the probability of certainty of error'
Peer comment(s):
neutral |
DLyons
: Both terms are very occasionally used in Statistics and I've never seen them together. This would mean nothing to a reader. And "probability of certainty" is definitely not a possible term.
8 mins
|
Thanks for your comment.
|
Discussion
I'm not aware of any version of this standard equation which actually equates two probabilities in its derivation (although some might certainly exist, there often are alternative ways of deriving results). If you read your reference carefully you'll notice that they are talking about SETTING both Type I and Type II error to be 0.05. "If one desired to set the probability of a Type II error equivalent to the 0.05 level used for the Type I error ...". They only use "equating" in the Abstract and this usage is similar to what's happening in your source.
http://www.saylor.org/site/wp-content/uploads/2011/06/MA121-... is a typical discussion.