Green Team

Kristin Celentano

Jaime Elving

Dustin Kirk

John Puccino

The project that we have decided to pursue and hopefully conduct involves determining whether written messages, such as instant messaging and e-mail, convey the tones and emotions of the message as effectively as verbal messages. We will also try to determine whether marking up the written message with the use of punctuation, bold or italics, and emoticons have an effect on whether or not people understand the intended tone of the message. Our hypothesis for this experiment is that it is easier to determine the tone of verbal messages than written messages. Also, we feel that it will be easier to determine the tone of the written messages that have been marked up as opposed to the plain text messages. We further hypothesize that sarcasm will be the hardest emotion to determine through the written messages.

With e-mail and instant messaging becoming more and more common in our culture, we feel that this is an important study to conduct. In the past, the telephone was the dominant method of communication. A telephone allows for verbal cues that can hint at the intended tone of the message. E-mail and instant messaging are lacking these hints. With e-mail and instant messaging taking over as the dominant forms of communication today, it is essential to determine whether messages are being falsely interpreted.

Our pilot test involved asking individuals to determine the emotions expressed by two messages. Some of the participants had to evaluate written responses and determine what emotions they thought were being conveyed, and some participants were asked to determine the emotions being conveyed in the same messages in a verbal form. The first of the two messages that were used is:

Wasn't class today so interesting? I don't think I have ever heard a more
fascinating lecture. The examples he went over really helped to explain the
concepts. I always understand the material so much better after attending
one of his lectures.

The intended interpretation of this message was the expression of sarcasm. The actual recorded interpretation of this message is that those who received it the written form had more difficulty identifying it as sarcastic and some even thought it could be a happy message. Those who heard the message verbally consistently recognized the message as obviously sarcastic. Hence, as we hypothesized, sarcasm is less likely to be interpreted correctly in a written form than a verbal form.

The second of the two messages was:

I cannot believe that I wasted like 12 hours of my life studying for a test
that I ended up failing! I don't ever want to take another course with this
professor ever again. I should just drop it now. I don't know why I am
even wasting my time. I worked so hard and have absolutely nothing to show
for it.

The intended interpretation of this message was the expression of anger. All of those who heard this message in a verbal form correctly identified it as expressing anger. Most of the people who received the message in a written form also correctly interpreted the message as expressing anger, however, a small number of the participants (all female) indicated that they thought that the tone of the message could be sad or upset. It seems that anger is more easily interpreted correctly in the written form than sarcasm, but in both cases the verbal form is more frequently interpreted correctly.

In the actual study, the participants will take the survey online via a CGI script. The script is already set up to give the written messages, with the corresponding surveys, and the verbal messages (in the form of a sound file) with the same surveys. Each person who agrees to participate will choose a username and give that username and their e-mail address to the person who asks them to be in the study. The person asking them to be in the study will record only that information as well as the gender of the participant. The participants will be e-mailed the username, a password, and the URL to the webpage with the survey. All of the users will be placed in the appropriate groups before they are sent the e-mail, and they will have to log in by entering the user name and password given in the e-mail. When they log in, they will receive the messages as specified by which group they are in (written, verbal) and the surveys. All results will be written to a file by the CGI script in an organized fashion for easy retrieval. (Note: this script has already been written and is performing as outlined above.)

The group received feedback from the professor and class about improvements to make to the survey. As a result we will be making a few adjustments to the design and depth of our survey. The first of which is to expand the number of messages to 7. Each will convey a different emotional tone. They include emotions ranging from anger to embarrassment. The second adjustment to our study is to include not only a voice and written paragraph, but also a marked up paragraph that will include net slang to get the emotion across. The net slang includes uses of italics, capitals, underlined text, and emoticons in the message. We hypothesize that this marked up paragraph will convey the tone of the message better than the paragraph with proper grammar, however neither written message will convey the tone as well as the voice message does. To implement this in our survey, individuals will now take one of three surveys, all voice, all standard paragraphs, or all marked up paragraphs.

Currently the group is choosing from a variety of paragraphs to use in the survey and is fine tuning the online survey which we will use when actual data collection occurs.