This experiment was to determine whether different group sizes affect foraging success in birds. We wanted to answer the question 'Why do some animals forage in groups?' We also wanted to test whether there was significant relationship between animal group size and foraging success, animal group size and number of predator scans and animal group size and mortality rates.
Trends in the Data Statistical Tests
From looking the data, we can see three clear trends. They are as follows;
The larger the group size, the higher the net number of seeds consumed by the group.
The larger the group size, the lower the number of scans for predators there are. The larger the group size, the lower the percentage of animals that are killed.
Trends in the data can be seen by the figures below. Figure 1 shows foraging success against an increasing group size. The graph shows that as group size increases, foraging success increases. Standard deviation bars show the spread about the mean. For figure 1, only standard deviation values above the mean were included, as there was a large range of value and the standard deviation bars below the mean for group size 1 had negative values. All the standard deviations bars in figure 1 overlap, meaning the spread about the mean for all four groups is quite similar in size, and that they overlap. This could be to do with the size differences of the groups. Figure 2 shows a decrease in predator scans as group size increases. This also has standard deviation included as error bars, so the spread around the mean can be seen. The size of the standard deviation bars in figure 2 are all quite similar, all overlapping Figure 3 shows that as group size increases, forager mortality decreases. Interpreting graphs and tables only by visualisation does not show a significant effect therefore further statistical tests must be carried out in order to determine if the results are significant.
Get your grade
or your money back
using our Essay Writing Service!
For the effect of group size on foraging success, the data was first tested for normality. As the data was shown to be not normal, a non-parametric analysis of variance must be used, so the Kruskal-Wallis test was performed on the data. This test was chosen as the data differ from a normal distribution and there are four independent variables involved in the test for group size.
In order to test to see if tend 2 is significant, I will carry out an ANOVA. I will carry out this test because the data do not differ significantly from a normal distribution and again, there are four independent variables involved in the test. Therefore, a parametric analysis of variance must be used, namely an ANOVA.
For the effect of group size on the number of predator scans, we first tested for normality. The data was not significant so was normally distributed. A parametric analysis of variance, or ANOVA was then carried out because the data did not differ significantly from a normal distribution and as before, there are four independent variables involved in the test. A Bonferroni post hoc test was also carried out, which allows multiple comparisons within groups.
For the effect of group size on forager mortality, a chi-squared test was carried out in order to test for association. This test determines whether the percentage of foragers killed differs significantly between the different group sizes.
Figure 1. Foraging Success in group sizes of 1, 2, 4 and 8 with positive standard deviation values.
Figure 2. Number of Predator Scans in group sizes of 1, 2, 4 and 8 with standard deviation values.
Figure 3. Forager Mortality in group sizes of 1, 2, 4 and 8.
Results and Significance
The Effect of Group Size on Foraging Success
Null Hypothesis; Net seed intake does not differ between the group sizes.
Tests of normality were first carried out. Kolmogorov-Smirnova gave a significance value of 0.006 (P=0.006). This means p<0.05 so is shown to be significant. Shapiro-Wilk gave a significance value of 0.000 (P=0.000). This means p<0.001 so is shown to be very significant. As both of these values are significant, the data does not lie along a normal distribution, so a non-parametric multiple comparison test must then be used. This is the Kruskal-Wallis test.
Always on Time
Marked to Standard
When computing Kruskal-Wallis, an exact significance value could not be computed. The asymptotic significance value is 0.000. (P=0.000) This means that p<0.001, the results of the test are very significant (Kruskal-Wallis, χ2=75.690, df =3, asymptotic p=0.000). As the results are significant, we can reject the null hypothesis in favour of an alternative hypothesis; net seed intake does differ significantly between the different group sizes.
The Effect of Group Size on the Number of Scans Carried Out.
Null Hypothesis; Scan rate does not differ between the four group sizes.
Tests of normality were first carried out. Kolmogorov-Smirnova gave a significance value of 0.200. (P=0.200) This means p>0.05 so is not significant. Shapiro-Wilk gave a significance value of 0.085 (P=0.085). This means p>0.05 so is also not significant. As both of these values are not significant, the data lies along a normal distribution, so a parametric Univariate Analysis of Variance (ANOVA) should be used.
The output for ANOVA gives an output of F= 19.81 with a significance level of p=0.000 at three degrees of freedom. The significance value for Group Size is 0.000. This is significant at the p<0.001 level of significance. The results from this test have been shown to be significant, and so we are able to reject the null hypothesis in favour of an alternative hypothesis, scan rate differs between the four group sizes (ANOVA F =19.811, p= 0.000).
Bonferroni Post Hoc Tests on 'Group Size'
Bonferroni post hoc tests allow multiple comparisons between groups, as it corrects the p-value for multiple comparisons Bonferroni shows that there is a significant difference between a group sizes of 1 and 4 and 1 and 8 at the p<0.001 level of significance (P=0.000). There is also a significant difference between group sizes of 2 and 4 and 2 and 8 at the p<0.001 level of significance (P=0.000)
There is no significant difference between a group size of 1 and of 2 at the p<0.05 level of significance (P=0.506) There is also no significant difference between a group size of 4 and 8 at the p<0.05 level of significance. (P=1.000).
The Effect of Group Size on Mortality, using Chi-Squared.
Null Hypothesis; There is no association between group size and mortality of individuals in different group sizes.
The chi-squared value for test of association is 51.68. There were three degrees of freedom. The critical value for chi-squared test of association at three degrees of freedom is 16.266. Our value of 51.32011466 is much larger than the critical value of 16.266 so our chi-squared value is shown to be significant at the p<0.001 level of significance (χ2=51.68, df=3, p<0.001).
The frequencies show a significant deviation from a uniform distribution.
DF 3 = (4-1)*(2-1)
Critical Values df 3 is 16.266
Factors Influencing Group Size
There are two inter-connected factors that influence group formation. These are predation pressure and foraging efficiency. To minimise energy output, feeding must be efficient. The optimum number of prey depends upon the time it takes to find and pick up food and the time it takes to carry food to the nest. The optimum number of prey maximises the number of prey carried and caught and minimises the time spend finding and transporting food. WEASELY THINGS?!!
Group formation also protects against predators. In the case of pigeon flocks and goshawks, the strike rate of Goshawks decreases with increasing flock size because larger flocks were able to spot the goshawk earlier and so took flight earlier. This is one of the main advantages of foraging and living in a large group.
Group formation also means that the time spent scanning by each individual decreases with increased group size. As the group size increases, there is more time for feeding. In terms of ostrich groups, when group size increases, the percentage time each bird has its head up scanning decreases, and as group size increases, the percentage time the group has one or more heads scanning increases. This is advantageous for each individual ostrich, as they are able to spend more time feeding. Also the time spent feeding is safer, so they are likely to eat more ..............?
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
Group foraging can also make it easier to find ephemeral food and exploit it. Groups of animals that hunt in packs, such as wolves are better adapted to hunting and finding food sources when group sizes are larger. They are able to use teamwork to take down prey that would be too large for them to tackle alone.
Another good example of where it pays to be in a larger group is that of Prairie dogs. Most live in 'towns' and within those towns are groups of related females in close-knit groups with a single male. The larger the group, the more likely that a predator will be spotted and an alarm will be raised. Prairie dogs will take it in turn to dig and forage, and all prey caught is taken back to the burrow, and shared out. This means that even the animals that are not out foraging are fed.