Fish in the same pen are more similar to each other than fish in different pens, but how similar they are differs at different mean gill scores and which score is used. This app takes this into account and shows the relationship between number of pens per site to sample, and number of fish per pen to sample.
This app can be used to develop sampling strategies that work for you, or to evaluate an existing one, based on the mean values you expect. Mean values might change throughout the season, and it may be possible to adjust the sampling regime to make optimal use of resources, based on this app. For example, in the winter when low gill scores are anticipated, it may be possible to sample less and still have the same statistical power.
This app is designed to allow you to either develop, or evaluate a strategy. Therefore, try different anticipated mean values, and if your practical situation requires more or less pens per sites or fish per pen, you can make an estimate of what your allowable error of the mean will be based on the chart.
Anticipated Mean Value: The value that you think the average score at the moment of sampling is. For example, develop a summer and a winter sampling regime, you can use an average winter score and an average summer score. Play with different scores to see how much it matters is a score is slightly different.
Allowable Error of the Mean (Tolerance): This is the variation in the mean that you are okay with. For example, if you have chosen an anticipated mean value of 1.5, then an allowable error of the mean 1.25-1.75 means that you are comfortable finding a mean that is somewhere in that range. If you choose 0-3.5, then your comfortable with a much wider range, as a result your sample size needed will be much smaller because it is not necessary for you to be very accurate.
Each boll represents a number of pens on the x-axis and a number of fish per pen on the y-axis for the corresponding allowable error of the mean. It indicates that based on the data behind this app, when you follow this sampling regime, your will be able to have enough samples to be 95% sure the mean falls within the allowable interval.
If you have any questions, please get in touch with the CEPH team or Annette Boerlage
This shiny app has been produced as part of the project SL_2017_07