Case 2: Comparison of Work Efficiency
Comparison of Work Efficiency in Proofreading Requiring Deep Understanding
When we read a document carefully or one that is difficult to understand, we sometimes point to the text or slide our fingers along the text without noticing it. This tendency is more pronounced when we read a document written in a foreign language or study for an exam. On the other hand, more people are reading electronic books using tablet devices represented by iPad. The operation of these tablet devices is intuitive and a user can enlarge or flip pages by just touching the panel. In order to determine the effects of using tablet devices in proofreading that requires a deep understanding of content, we compared the work efficiency of proofreading a document printed on paper with that of proofreading a document by using a tablet device.
Example: Proofreading a text document
Text documents created by editing newspaper columns are used (where one document consists of about 660 characters). Each document has five errors that cannot be found unless a reader correctly understands the context (such as when "decreased" is used incorrectly instead of "increased"). Twenty-four participants were asked to find as many errors as possible within the time limit (of four minutes). We compared the error detection rate when using paper with that when using a tablet device (Table 1). The results show that the error detection rate when using paper for proofreading is 17.2% higher than that when using a tablet device (Fig. 1).
Table 1: Experimental conditions
|Participant||24 persons (23 to 40 years old, average age of 31.3)|
|Comparison target||A paper document and a tablet device
|Task||Proofreading text documents
The behavior of participants during proofreading is analyzed to find out why using a tablet device for proofreading was less effective than using paper. You can see that the participants pointed to and slid their fingers or pens along the text more in the paper-based task (Fig. 2). And there is a positive correlation between the frequency of these actions and the error detection rate. In other words, the more a participant slid a finger or pen along text and pointed to the text, the higher the error detection rate.
People are apparently less likely to slide a finger or pen along text or point to the text when reading text using tablet devices, because pages can be accidentally flipped when a person touches the screen. They also tend to avoid touching the glossy screen of a tablet device. These may be the reasons why the participants could not detect as many errors as when using paper in proofreading.
In order to confirm this hypothesis, we conducted an experiment to investigate the effects of such actions as pointing to text and sliding a finger or pen along text during reading tasks. We confirmed a lower error detection rate when these actions are not allowed, compared to that when the actions are allowed, even when using paper for proofreading (Fig. 3).