2002-2003: Incidence of Foul-hooking in FMRI* Boca Grande Pass Tarpon Catch and Release Mortality Study (REVISED AND UPDATED SEPTEMBER 11, 2013)

The purpose of this document is to clarify our results to date relating to the frequency of foul-hooking associated with tarpon caught on break-away jigs and live bait in Boca Grande. The results from 2002-2003 observations relative to hook position and hook orientation are summarized in Table 1. For the live bait method, all fish were hooked in the mouth region, 88% were hooked in the jaw with hook orientation inside/out. Three live bait fish were hooked inside the mouth. One live bait fish was hooked in the clipper.

For jig caught fish, 50% (13 out of 26) were hooked in the clipper. Hook orientation is available only for 7 of the 13 fish hooked in the clipper. For these 7 fish, only one was hooked outside/in, the remaining 6 were hooked inside/out. A total of 4 jig caught fish were hooked in a position other than the jaw or the clipper. One fish was hooked inside the mouth, one fish was hooked in the gills, and two fish were hooked elsewhere on the head.

Table 1: Summary of Tarpon Catch and Release Study Boca Grande Pass, 2002-2003 Landed Tarpon (tagged and untagged)

Hook orientation was not recorded for clipper hooked fish in 2002.
*5 of these 6 fish were hooked from the inside corner of the mouth membrane out to the clipper

Definition of Foul Hooking
Our definition of foul hooking for this study was based on definitions used for other areas, published scientific literature, and input from experts in tarpon feeding behavior and fish functional morphology. The most common and widely accepted (AL, CO, MA, TX, Canada, UK, the International Gamefish Association, and virtually all popular glossaries and "Dictionary of Ichthyology") definition of foul hooking is:

More restrictive definitions exist. For example in New Zealand, foul hooking is defined as:

At least one state (MS) has adopted a less restrictive definition:

Based on the information we have in hand, we are adopting the following definition of foul-hooking for tarpon:

"A foul hooked fish is hooked in a part of the body other than the mouth"

Based on this definition, only 3 fish listed in Table 1 qualify as foul-hooked fish. All three of these fish were jig-caught (1 in the gills and 2 on the head). These results indicate that 11.5% (3 out of 26) jig caught fish were foul hooked. No live bait caught fish were foul-hooked based on this definition.

Statistical Analysis of Foul Hooking Data
Based on the definition of foul-hooking above, results from the first two years of sampling may be summarized as follows:

The analyses of these data are complicated by the relatively small sample size. The small sample size results in a zero cell for livebait foul hooked fish - theoretically, even if the frequency of foul hooking for live baiters is extremely low we would get at least one instance of foul hooking as sample size increases. Statistically, it is appropriate to view these count data as generated by a Poisson process and the use of chi-squared test for a two way contingency table is an appropriate analysis. The chi-squared test tests the null hypothesis of independence, i.e. that fishing method and number of foul hooked fish are completely independent. Rejection of the chi-squared test indicates dependence, i.e. the tendency to foul hook fish depends on fishing type.

Expected frequencies under the assumption of independence are easily computed:

Under the assumption of independence, and with our sample size of 58, we should have observed approximately 2 foul hooked fish from the livebait method and one foul hooked fish from the jig method. Again, these small expected values are an artifact of the small sample size.

The chi-square test yields the following results:

Phi for 2 x 2 tables: .2591001

The Pearson Chi-square and the Maximum-Likelihood Chi-square test the hypothesis of independence, the only difference is that the ML chi square is based on Maximum-Likelihood theory. In practice, the M-L Chi-square is usually very close in magnitude to the Pearson Chi-square statistic. Both the Pearson Chi Square and the M-L Chi Square result in rejection of the null hypothesis at the 5% significance level - i.e. the number of foul hooked fish depends on fishing technique. It must be noted that the only assumption underlying the use of the Chi-square (other than random selection of the sample) is that the expected frequencies are not very small. The reason is that the Chi-square inherently tests the underlying probabilities in each cell; and when the expected cell frequencies fall, for example, below 5, those probabilities cannot be estimated with sufficient precision.

The approximation of the Chi-square statistic in small 2x2 tables can be improved by reducing the absolute value of differences between expected and observed frequencies by 0.5 before squaring (Yates' correction). This correction, which makes the estimation more conservative, is usually applied when the table contains only small observed frequencies, so that some expected frequencies become less than 10 - exactly the case we have with tarpon. This correction results in the test statistic labeled Yates Chi-square in the table above. The Yates corrected test fails to reject the hypothesis of independence at the 5% level, it does, however, reject the null at the 10% level. Again, we have uncertainty generated by our relatively low sample size.

The Phi-square statistic in the table above is a measure of correlation between two categorical variables in a 2x2 table. Its value can range from 0 (no relation between factors; Chi-square=0.0) to 1 (perfect relation between the two factors in the table). The calculated correlation of 0.26 indicates a low degree of correlation between fishing method and frequency of foul hooked fish. As with the other results, the interpretation of this correlation is hampered by the relatively small sample size.

Power Analysis
Statistical power in this case is simply the probability that we reject the null hypothesis (independence of fishing type and frequency of foul hooked fish) when it is in fact false (i.e. the probability of detecting a statistical effect if one exists). The most effective way to increase statistical power is to increase sample size. A statistical methodology has been developed to determine the power of the Pearson chi squared test for 2x2 contingency tables. The plot below presents statistical power on the y axis and sample size (number of fish) on the x axis. The current dataset results in a statistical power of about 0.5 based on a sample size of 58 fish. One commonly accepted baseline for a powerful experiment is Power = 0.7 or greater. For this study, a sample size of approximately 90 fish will be required to achieve a power of 0.7. It should be noted that 1) sample sizes in this study were set relative to the catch and release mortality assessment portion of the project and 2) this power analysis may change as another year of data on foul hooking are collected. The total number of fish necessary to achieve Power=0.7 may decline, but will likely not increase with additional data collection.