Hancock, J.T., Landrigan, C., & Silver, C (2007)
Expressing Emotion in Text-based Communication
CHI 2007 Proceedings Emotion and Empathy. P 929-932
Theoretical Framework
SIP – strategic relational adaptation in mediated communication. (Parkinson relational alignment). P929 users employ the verbal cues present in CMC to convey relational information that may normally be tramsmitted via non-verbal cues
Method
This study influenced by Walther et al (2007) when confederate in a dyad asked express affinity ( or not)
Using IM Dyads ( undergraduates getting module accreditation) confederate asked to express emotion ( positive or negative) to naïve partner.
What strategies are used to express emotion? Confederate view. Confederates asked to respond to the questionnaire below
Positive and negative descriptives | P values |
Punctuation 4.22 .45 2.95 .43 | <0.05 |
Typed More 4.56 .40 3.95 .38 . | .28 |
Emoticons 4.00 .53 3.15 .50 . | .25 |
Explicit Emotion Statements 4.06 .41 3.00 .39 . | .07 |
Encourage Partner 3.28 .36 3.40 .34 . | .80 |
Respond Quickly 4.72 .37 3.20 .36 | <0.01 |
Self-disclosure 4.50 .37 4.30 .36 | .70 |
Agreement 5.00 .36 2.85 .34 | <0.01 |
What is the textual evidence?
Linguistic analysis using LIWC
For the purposes of the study only the linguistic features shown in the following table were used ie relevant to the purpose of the study ( no further justification given!)
% measure – divides the frequency of the observed variable by the total number of words in the sample
| Linguistic Category positive then negative M SE M SE | |
Word Count | 590.40 44.08 458.70 42.94 < .05 | <0.05 |
Affect | 4.44 .24 5.27 .33 < .05 | <0.05 |
| Positive feeling .80 .12 .63 .11 .30 | .30 |
| Negative feeling .11 .14 .55 .53 < .01 | <0.01 |
| Emoticons .26 .09 .15 .07 .34 | .34 |
Pronoun | | . |
| 1st-Person 7.31 .38 6.72 .45 .33 | .33 |
| 3rd Person 1.38 .22 1.70 .28 .37 | .37 |
Agrrement | | |
| Negations 1.58 .15 2.29 .20 < .05 | <0.05 |
| Assents 1.69 .20 2.51 .35 = .05 | =0.05 |
ExclamationPoints | 7.45 2.04 1.20 .69 < .05 | <0.05 |
Msgs Per Minute | 2.64 .22 2.37 .26 .43 | .43 |
Note p931 ‘ Linguistic analysis of the texts revealed that several verbal dimensions differed significantly across the two emotion conditions although not necessarily in the manner the expressers reported in the questionnaire response
(KRO ? evidence for the involuntary)
Naives view
Naïves given questionnaire with 11 items to assess their ability to detect the expresser’s emotion.
FA gave 2 factors, perceived mood of expressor, and perceived relationship. Naïve partners were able to detect the expressor’s emotional state. Analysis suggest that negations and exclamation points were relied on most heavily by assessors when judging their partner’s affective state…
Discussion
Supports SIP .
Data suggests that four strategies ( ? use of this term) were used for differentiating between positive and negative emotions. Note that expressors reposnses to the questionnaires was not always born out by the linguistic analysis.
Walther et al agreement led to more liking.
- However for this data ( emotion) positive affect users disgreed significantly less.
- Negative used x5 times more negative affect terms, although they seemed to be unaware of this on the questionnaire ( KRO involuntary again)
- More x6 punctuation by positive expressors
- Positive reported responding quickly although analysis showed that they did not.