Personal tools
You are here: Home 5. Interpret 5.1 Introduction: Interpret data

5.1 Introduction: Interpret data

Overview  Introduction Statistical Assessment  Mechanics of Interpretation   Finding Meaning in the Data   Results and Next Steps

 

Letters have to pass two tests before they can be classed as good: they must express the personality both of the writer and of the recipient.
 —E.M. Forster

Once the analyst moves on to the actual interpretation step, he or she should see two parts to it -- assessment and interpretation. The statistical assessment is where field measurements are turned into what we are calling metrics, and then these metrics are turned into indicators. The statistical assessment part of this step is essentially mechanical. In other words, the first part requires a straight-forward application of algorithms to some form of data. As we saw previously, the response and inference designs are the plans for using statistical algorithms to develop the estimates needed to reach the goals of the analysis. Although there is a lot of complexity and specialized knowledge that goes into conducting field and statistical operations, the hardest part is the interpretation step that comes after the statistical assessment.

old style Bendix sonar in Alaska.JPG
Old-style Bendix sonar used to count sockeye salmon in Alaska starting in the 1970s. These machines were starting to be phased out in the late 1990s. Credit Hal Geiger/ADF&G.

Interpreting the data so as to find meaning is what separates fishery science from accounting or arithmetic. The interpretation has not been completed until the analyst returns to the original goals of the study and then finds answers to the questions implied by these goals. To continue with the example of escapement monitoring, the final estimate of escapement for this one year is an indicator for the stock's status in that year. The larger goals of an escapement monitoring project usually require looking at that one-year indicator in an historic or spatial context so that decision makers can be given information about whether the escapement has been trending up or down, or to answer questions about whether or not some expensive management or industry action was effective. 

 

Next, let's look at some more information and examples of statistical assessment.

NEXT: Statistical AssessmentGo Back

Document Actions

Addition of concept and examples of "universe of inference"

Posted by peterman at Sep 16, 2009 04:55 PM
We should add to this section 5 the concept of "universe of inference" and examples of doing this properly as well as improperly. The Schwarz (1998) paper that I put on our group's wiki early in the project gives an excellent overview of both proper and improper extrapolations of findings.
 -- Randall