The RSVP Report
In this test, we ran two distinct variance measures (pre-target and post-target) of variance. The two variance measure was appropriate because the pre-target was circumscribed to trials, thus creating uneven numbers of observations in both conditions. The figure below shows a screenshot of the ANOVA test results.
Results Analysis
We performed a two-way group analysis of sample variations to find out the differences in rapid stream visual presentation (RSVP) tasks on pre- and post- target instructions. The ANOVA test results were in three phases; within the subjects’ effects, between the subjects’ outcomes, and pure main effects stream, as shown in the figure above. In an attempt to determine whether the differences in the means were statistically significant, we compared the observable p-values (see the picture above) to the significance level of α=0.001 to assess the null hypothesis of the research. The null hypothesis of the research states as follows:
“If slower engagement encourages all kinds of error, there should be an equal increase in both pre-target and post-target intrusions in the two-stream condition relative to the single-stream condition. If slower engagement is relevant ONLY for the post-target intrusions (and not for pre-target intrusions), then there should be an increase in post-target intrusions in the two-stream condition relative to the single-stream condition. Otherwise, there should be no increase in the pre-target intrusions. A third, even more, last option, is that the pre-target intrusions will be in the two-stream condition.”
The results above show that there were statistically significant differences at the p<0.001 level in rapid stream visual presentation for both within-subject effect and simple main effect streams. There was a significant increase in post target intrusions in the two-stream condition in contrast with one stream condition 0.176 to 2.973. There was no increase in the pre target intrusions in one stream versus two-stream conditions as in 0.035 and 0.176. There was a significant interaction between the effects of streams for pre and post target conditions (F (.012) =9.566e-6, p>.915). To further examine the nature of this interaction, we conducted a simple effects analysis. Simple main effects analysis showed that post-target interaction is significantly more significant (F (40.680) = 0.176, p<.001), but there was little to no difference for the interactions between subjects.
The result also showed that there is a statistically significant difference between error type, streams, and error type*streams since all their p values were less than 0.001 (p<.01). For instance for the error type, F (93.22) = 2.973, p<.001; streams F (43.75) = .087, p<.001; and error type*streams F (28.14) = .089, p<.001. In spite of these two results (for streams and error type*streams) reaching the statistical significance level, it is important to note that there was a small difference in the mean scores. On the other hand, the post-target comparisons showed that the mean score for the simple main effect streams F (40.680) = 0.176, p<.001) was statistically different from the pre-target (F (.012) =9.566e-6, p>.915). Otherwise, there was no statistical difference in the RSVP between-subject effects.
For result interpretation, it would be probable to say that a slower engagement corresponds to all types of errors differently. After manipulating attention speed by including one stream versus two, there was a faster attentional engagement in one stream. So, the null hypothesis holds in the sense that the slower encounter encouraged all types of errors, and there was an equal increase in both the two target level instructions in their two-stream condition compared to one stream condition.