WAHLEN AND SMITH: OBSERVER EFFECT ON INDICENTAL DOLPHIN MORTALITY 



not differ significantly in their percent of zero-kill 

 sets before or after March 1981. 



Same Time Period 



Before March 1981 the percent of zero-kill sets was 

 higher for NMFS-observed trips than for lATTC- 

 observed trips (Fig. 4), and within all area-period 

 strata with complete data, the percent of zero-kill 

 sets was larger for the NMFS observers (Ikble 5). 

 The one-sided test (Comparison lb, Ikble 2) was 

 significant within four of the five area-period strata 

 which had complete data (Tkble 5). Thus, for trips 

 making dolphin sets from 1979 through March 1981, 

 NMFS observers recorded a significantly higher per- 

 cent of zero-kill sets than did lATTC observers. 



According to our hypothesis, the difference in per- 

 cent of zero-kill sets between NMFS- and lATTC- 

 observed trips before March 1981 should have disap- 

 peared after March 1981 when the purposes for data 

 collection of NMFS observers became nearly the 

 same as for lATTC observers. After March 1981 the 



percent of zero-kill sets was higher for lATTC- 

 observed trips than for NMFS-observed trips (Fig. 

 5), but within the six area-period strata the dif- 

 ferences were not consistent (Table 6). The two-sided 

 test (Comparison 2b, Tab\e 2) was significant within 

 only one of the six area-period strata (Period 1, 

 North Inside), yet this one chi-square statistic was 

 so large that the overall conditional test for all six 

 strata was also significant (Ihble 6). It is difficult to 

 interpret the overall result in this situation because 

 of the extraordinary influence of one stratum. 

 However, after March 1981 the percent of zero-kill 

 sets on NMFS-observed trips was clearly not higher 

 than on lATTC-observed trips. 



While one would expect the mean number of 

 dolphins killed to decrease when the percent of zero- 

 kill sets increases, this is not necessarily so because 

 of the sensitivity of the mean of a sample to the max- 

 imum value in the sample For instance, in Figure 

 4 the NMFS maximum is nearly twice that of the 

 lATTC, resulting in a larger NMFS mean despite the 

 higher percent of zero-kill sets in the NMFS sample 



>■ 

 O 



z 



Ui 



o 



lU 



e 

 11. 



lU 



> 



UI 



flC 



I ^ . — p^ 



T 



=p 



=f 



1-9 10-19 20-29 30-39 40-49 50-99 ^100 



NUMBER OF DOLPHINS KILLED 



Figure 4.— Relative frequency distributions of number of dolphins killed incidentally during sets 

 made from 1979 through March 1981, by observer type 



527 



