search

pbs-assess / herringsr

:paperclip: :fish: Herring Science Response document
0 stars 1 forks source link

Show change in weight-at-age over time #94

Open grinnellm opened 3 years ago

grinnellm commented 3 years ago

Add a figure to show change in weight-at-age over time. Could show percent change by year (as in the data summary) or decade, or weight-at-age relative to the average of the time series, or weight-at-age compared to the first year.

Low priority; not for this year.

jaclyncleary commented 3 years ago

What statistically makes sense for "showing" the changes? Personally I find "percent change" hard to explain to stakeholders, my eye seems to focus on the zero line (like with rec_devs).

Sages11 commented 3 years ago

I was also interested in how strong age classes affect weight at age. - I imagine whatever helps create a strong age class might also increase their weight at age, and this trickles up the age classes. I feel like there are already hints of that in the weight at age, and proportion at age graphs. Maybe there is a way to show this by combining the two graphs. Maybe a Breslow-Day test on brood years vs weight at age could test the null hypothesis weight at age is independent of brood year.

grinnellm commented 3 years ago

For the assessment (or data summary) I think something simpler would be better for the audience. Perhaps a time-series of the difference between the weight in each year and the mean weight of the entire time series (or the weight in 1951). That way it would be easy to see, for example (and I am making these values up), that age-3 herring in 2021 are 5 g lighter than average, but they were 12 g lighter 10 years ago.