Rather than repeat the same remarks on various of your answers to the Midterm, I've assembled some of the important points here. This should help you understand any problems that were apparent to me and caused you to lose some points.
Question 1 (25 points) - Give an example of how information objects might be shared in synchronous and in asynchronous interactions.
Most missed "information objects", esp. for the synchronous case. The simplest such objects would be addresses, phone numbers, times to meet, etc. When exchanged on the phone or VOIP a person would need to write them down. IM can save the exchange, or by going off record, can keep, say, a password that was sent from appearing in the record. There are old-fashioned telephone conference calls still available.
Most got asynchronous, esp. attachments in email.
Question 2 (25 points) - Discuss two problems of anthropomorphic design.
A few thought that truly human-like agents are possible, but no one knows how to do that yet.
So one problem is failing to be human enough, leading to mistrust.
Another is a design that "apologizes" when it's just a machine.
Another is children that could spend too much time with a toy that is limited and shallow, depriving them of true human interaction and possibly shutting down the imagination they would use if they had a non-agent such as a simple stuffed animal or a Lego character.
The creator of Dungeons and Dragons, Gary Gygax, who passed away on 3/4/2008, had commented on these issues,
"There is no intimacy; it's not live," he said of online games. "It's being translated through a computer, and your imagination is not there the same way it is when you're actually together with a group of people. It reminds me of one time where I saw some children talking about whether they liked radio or television, and I asked one little boy why he preferred radio, and he said, 'Because the pictures are so much better.'"
An aside: A puppet is an interesting mix, sort of a Wizard of Oz agent, a constructed non-living creature that has the intelligence and responses of the puppeter to another person in a conversation.
Question 3 (25 points) - Give an example of an interaction based on an abstract design and an example of the same type of interaction based on a realistic design.
There is a continuum between abstract and realistic, with no sharp dividing line.
File icons arranged in a window, especially if they are represented by thumbnails is on the realistic end. A list of filenames, sizes, etc. is on the abstract end of the spectrum.
Abstract interactions are often more efficient for experienced users. Realistic interactions are difficult when, for example, you are dealing with 1 million files on a machine.
A multi-touch screen (iTouch) is more realistic than an interface that requires you to enter a numerical percentage.
A "hand" that allows scrolling is more realistic than scroll bars.
Wii-based games are quite realistic. Games from twenty-five years ago were sometimes command-line only.
The most imaginative answer I read was a realistic grocery shopping system in which you were presented with a realistic depiction of a store and you controlled your avatar as it moved, let you look at items, select some, and put them in your cart - all simulated. (Later they'd be delivered to your residence.) The amusing part was a scene in which your avatar would get into an argument with another avatar about whether you or them (it?) would get the last container of a particular flavor of yogurt. The other avatar was presumably controlled by another logged in human shopper.
Question 4 (10 points) - Give an example of a Likert scale and semantic differential scale. Briefly discuss how they differ.
In a Likert scale, a statement is made and the answer is a series of numbers or "agree" ... "disagree".
A semantic differential scale has pairs of contrasting words at the two ends of the scales, e.g., "Clear" and "Confusing" as well as "Attractive" ... "Ugly", and so forth. Rather than a statement, the various scales describe different aspects of or attitudes towards some thing or system (textbook pgs. 315 and 367). The better answers had more than one scale.
It can be difficult to come up with good choices for contrasting words in the semantic differential scales.
Question 5 (15 points) - When you attempt to make sense of patterns of responses from subjects, how might they be used to improve your instruments (questionnaires, interviews, and observations).
This was a broadly posed questions. The answers to it had to focus on how and why you might change your instruments on the basis of the types of responses you got. This could be used after a pilot study with a small group to refine the instruments for a larger study.
If your questions revealed very strong opinions about a topic, positive or negative, it might be good to add further questions to probe the possible causes of the strong responses. If you included the same question twice, with different wording and found the answers inconsistent, the question might need more explanation for the subject or related questions added to try to get a better result. If the users, in interviews or comments, brought up issues you had not planned to probe, you might want to add them to future interviews and questionnaires.
Return to ISU570 Fall 2007 homepage. or RPF's Teaching Gateway or homepage