Forums

Back

Sample size and weighting

RSS (Opens New Window)(Opens New Window)
MC
Massimiliano Cardinale, modified 5 Years ago.

Sample size and weighting

Youngling Posts: 40 Join Date: 2/23/17 Recent Posts

I have a question regarding the use of weighting in SS. My understanding is that when using Ianelli or Francis methods, the weight of a given component cannot be increased over 1, as it would be basically saying that we have more samples than those reported, which is illogical. However, I heard different opinions regarding, especially when using hauls instead of the number of fish as sample size (which to me should be always the case, although I have seen models using number of fishes instead). What are the opinion of the forum on this aspect of SS models?

Cheers

Max

PS
Paz Sampedro, modified 5 Years ago.

RE: Sample size and weighting

Youngling Posts: 285 Join Date: 1/9/18 Recent Posts
Hola. Estaré fuera del laboratorio del 29 de agosto al 12 de septiembre con acceso limitado a mi correo electrónico. Paz Hi, Thanks for your email. I’ll be out of lab with limited access to e-mail until 12/09/2019. Kind regards. Paz
CW
Chantel Wetzel, modified 5 Years ago.

RE: Sample size and weighting

Youngling Posts: 32 Join Date: 10/7/15 Recent Posts
Hi Max,

It depends on how you calculated your input sample size for the length- or age-composition data.  If you used the number of samples as an input sample size, then no you would never want to go above 1 for the reasons you stated.  However, if you were using something lower than the number of samples (i.e., the number of trips), weighting with values greater than 1, as long as that would not exceed the numbers of samples.  Similar logic applies to the Francis weighting approach.  In the appendix section of the manual, there is some additional information on the alternative data weighting approaches.  If there is no text regarding weighting above 1 by method, I will add it for clarity.

Chantel Wetzel, PhD
Fishery Resource Analysis and Monitoring Division
Northwest Fisheries Science Center
National Marine Fisheries Service
National Oceanic and Atmospheric Administration
Phone: 206-302-1753


On Mon, Sep 2, 2019 at 8:05 AM Massimiliano Cardinale <VLab.Notifications@noaa.gov> wrote:

I have a question regarding the use of weighting in SS. My understanding is that when using Ianelli or Francis methods, the weight of a given component cannot be increased over 1, as it would be basically saying that we have more samples than those reported, which is illogical. However, I heard different opinions regarding, especially when using hauls instead of the number of fish as sample size (which to me should be always the case, although I have seen models using number of fishes instead). What are the opinion of the forum on this aspect of SS models?

Cheers

Max


--
Massimiliano Cardinale Stock Synthesis Virtual Lab Forum https://vlab.noaa.gov/group/stock-synthesis/discussions-forums-/-/message_boards/view_message/7492497 VLab.Notifications@noaa.gov
MC
Massimiliano Cardinale, modified 5 Years ago.

RE: Sample size and weighting

Youngling Posts: 40 Join Date: 2/23/17 Recent Posts

Dear Chantel,

 

Thanks for your answer but I am convinced that the number of fish should be the limit of the effective sample size for age and length compositions. Imagine the situation where you have 10000 fish in one year from one single haul and 10000 fish from 100 hauls, each with 100 fish. As life trait histories are generally correlated within the space, the effective sample size is not the same in the two cases, with the one coming from the single haul being lower than the other. I still think that the primary sample unit should be the haul and that sample size can be increased only in case we use trip, as they might contain fish taken from different hauls, hopefully not too close in space. However, as you say, I have seen people using all sorts of sample unit and mixing the up in the same model.

Cheers

 

Max