Thumbnails are a convenient way to store pictures, use as small icons, or as pictorial models of larger pictures. They can benefit the web by cutting down the download times necessary to view pictures that would otherwise take minutes to download. Some organizations out there such as The Bandwidth Society are attempting to mobilize research to find solutions to such problems.
We tried a simple approach to finding an optimal thumbnail. Testing size and reaction speed we hoped that a correlation could be made and statistical evidence might be found to assert that there is a thumbnail size small enough to maintain essential descriptive characteristics. Our experiment showed that size does affect the reaction speed and subjects' recognition of the thumbnail.
It is possible, however, that the whole experiment was within the range of "the optimal thumbnail", and our results would have been affected accordingly. We did find any statistically significant differences in mean recognition times, but some suggestions can still be made to webpage designers. Using thumbnail sizes between the range of 2730 pixels2 and 8580 pixels2 provided us the highest level of recognition with the least time to recognition. There was certainly no significant advantage in using sizes 5 over size 4, nor size 1 over size 2. These groupings we found to be significant steps towards finding such an optimal size.
Unfortunately we did not find that much research on thumbnail sizes has been performed. Until there is more research performed, our hypothesis cannot be confirmed. We believe that with an increased number of subjects and possibly a larger size range of thumbnails, the experiment could be improved upon. Other avenues, such as color saturation and gradation can also be explored. Larger test groups with stringent vision and color-blind testing should be performed beforehand. Using pictures which the subjects can identify, eliminating societal and gender biases, is of import. Many variables are difficult to isolate and eliminate in a dormatory.
Next Section: Acknowledgements