1 of 45

Complex and Ambiguous: �Understanding Sticker Misinterpretations in Instant Messaging

Yoonjeong Cha Jongwon Kim Sangkeun Park MunYong Yi Uichin Lee

2 of 45

Emoji �& �Stickers!

3 of 45

Motivation�

We Use Stickers A Lot!

2 Billion / per month

  • 2017 KakaoTalk IR

*Global Monthly Average User (MAU) : 50 million

*

4 of 45

Face to Face Communication�

Albert Mehrabian, Silent Messages, 1971

Gesture

Posture

Eye Contact

Facial Expression

7%

38%

55%

5 of 45

Online Communication in Messenger �

:D

(Text Messages)

6 of 45

Motivation�

7 of 45

Motivation�

My Intention :Come here”

?

8 of 45

Motivation�

My Intention :Come here”

Other people interpret :“I’m doomed”

Misinterpreted

9 of 45

Motivation�

Do Stickers Convey

Sender’s Meaning Well to Receiver?

10 of 45

Emoticon vs. Emoji vs. Sticker�

Emoticon

  • Text
  • Can be sent with text

11 of 45

Emoticon vs. Emoji vs. Sticker�

Emoji

  • Unicode mapping of text
  • Can be sent with text

12 of 45

Emoticon vs. Emoji vs. Sticker�

Sticker

  • (Animated) Image
  • Must be sent separately from text

13 of 45

Research Question�

Are stickers misinterpreted?

If so, what factors lead to misinterpretation of stickers?

14 of 45

Related Work�

  • Purpose of using Messengers (Ramirez et al., ‘09 / O’Hara et al., ’15 / Kim et al., ‘15)
  • Comparison to SMS (Church et al., ‘13)
  • Multiple MIM & Diverse Usage (Nouwens et al., ‘17 / Bouhnik et al., ‘14 / Johnston et al., ‘15)
  • Emoticons for interpersonal functions (Lo ‘08 / Dresner et al., ‘10)
  • Emojis for adding info or even repurposing (Kelly et al., ‘15 / Cramer et al., ‘16 / Wiseman et al., ‘18)
  • Stickers for functional, cultural usages (Lee et al., ‘16 / Zhou et al., ‘17)
  • Emoticon’s impact on message interpretation (Walther et al., ’01 / Derks et al., ‘08)
  • Emoji in different platforms (Miller et al., ‘16 & ‘17 / Morstatter et al., ’17)

Role of Instant Messengers

in Everyday Life

Role of Emoticons, Emoji, and Stickers

in CMC

Message Interpretation with Emoticon and Emoji Use

15 of 45

Methodology�

  • Goal : Examine sticker characteristics that lead to emotion and semantic difference of popular stickers

  • Methodology : Online survey on 20 KakaoTalk stickers (n=87)

Study 1 : Sticker characteristics that leads to misinterpretation

Study 2 : In Instant Messenger considering contextual factors

  • Goal : Identify sticker misinterpretation in instant messenger between sender and receiver

  • Methodology : Semi-structured interviews in pairs (n=28)

16 of 45

Study 1�Interpretation Discrepancy of Popular Stickers

Online Survey on 20 Animated Stickers

(Default & complimentary stickers in KakaoTalk)

17 of 45

Questionnaire�

    • Emotion in Valence (Pleasure) & Arousal (Activation) Level
    • Semantic Meaning (open-ended question)

18 of 45

Participants�

  • Recruiting
    • Online campus community & Facebook
    • November 23th to 29th 2017
    • Compensation: approx. 6 USD for randomly chosen 10 people

  • Participants (n=87)
    • Age : Mean=22.43, Median=21.5, SD=3.41
    • Gender : 42 male, 45 female

19 of 45

Data Analysis �

Standard Deviation of Emotion Valence (Pleasure) &

Arousal (Activeness) Level

Plot and Group Stickers

(Four group themes with different reasons of misinterpretation)

20 of 45

Findings in Study 1 :�Stickers with High Variance in Valence (Pleasure) Level

“embarrassed that I made a mistake”

“I’m shy”

Ambiguous and multiple facial/bodily expressions

Pleased

Displeased

21 of 45

Findings in Study 1 :�Stickers with High Variance in Arousal (Activeness) Level

“I want to hit you”

“being pissed off”

“pretending to be angry at a joke”

“hatred”

High intensity and dynamism of a character’s gestures

Activated

Deactivated

22 of 45

Findings in Study 1 :�Stickers with Not much Difference in Emotion

“Cheer up”

“Showing off”

Some stickers were interpreted diversely

without much difference in emotion

“I feel proud”

“Good job”

23 of 45

Summary of Study 1�

  • Ambiguous and multiple facial/bodily expressions
  • Intensity and dynamism of a character’s gestures

Sticker characteristics that lead to misinterpretation :

24 of 45

Study 2�Identifying Sticker Misinterpretation in Real-World

25 of 45

Participants�

  • Recruiting
    • Pairs who had used KakaoTalk to exchange at least 10 stickers in 3 months
    • Online campus community & Facebook
    • November 26th to December 1st 2017
    • Compensation: approx. 9 USD per person

  • Participants (n=28 in 14 pairs)
    • Age : Mean=22.71, Median=21.5, SD=4.15
    • Gender : 10 male, 18 female
    • Relationship: 7 pairs of friends, 6 couples, and 1 pair of coworkers
    • Avg. duration of a user pair having known each other : 22.71 months
    • Avg. number of days per week of sending each other KakaoTalk : 5.64 days

26 of 45

Interview Procedure�

Submit screenshots of

10 sticker usages & surrounding texts

sent between a pair of participants

(within recent 3 months)

1) Before the Interview

27 of 45

Interview Procedure�

2) During the Interview

Answer Survey Questions

Separately

  • Description on the situation
  • Sender’s intention
  • Sticker’s meaning (in a sentence)
  • Valence & Arousal Level (7-point scale)

28 of 45

Interview Procedure�

2) During the Interview

Answer Survey Questions

Semi-structured Interview

  • Detailed Intention & Interpretation
  • Sticker Characteristics (animation, texts)

29 of 45

Data Analysis�

140 stickers

(10 stickers per pair x 14 pairs)

Answers from

the survey & interviews

Qualitative Content Analysis

(Two Authors categorize matched or mismatched)

Thematic Analysis

(Derive common themes of mismatched stickers)

Matched

Mismatched

30 of 45

Findings in Study 2 : �Match vs. Mismatch

22 % of stickers were misinterpreted

31 of 45

Findings in Study 2 : �Match vs. Mismatch

They were unaware that

they understood the sticker differently

before the interview

32 of 45

Findings in Study 2 : �Two Themes from Mismatched Stickers

Visual Representation and Reference

(n=25)

Message Correspondence

(n=11)

Both

(n=5)

THEME 1

THEME 2

: Sticker’s supporting text is differently interpreted

: Sticker’s face/body expressions are ambiguous

33 of 45

Theme 1 : �Visual representation and reference mismatch

Sender is asking receiver

on Sunday night

34 of 45

Theme 1 : �Visual representation and reference mismatch

“I did not want to bother the receiver. But I was in an urgent situation of having to ask questions.” (P15)

Sender

Is it okay to ask you something?

Sender is asking receiver

on Sunday night

35 of 45

Theme 1 : �Visual representation and reference mismatch

Is it okay to ask you something?

Where are you?

“I did not want to bother the receiver. But I was in an urgent situation of having to ask questions.” (P15)

“I thought there is nothing for the sender to feel sorry about. As his mentor, it is okay for him to ask me anything. I didn’t know he thought like that” (P16)

Sender

Receiver

Sender is asking receiver

on Sunday night

36 of 45

Theme 2 : �Message correspondence mismatch

Sender drank Korean vodka (Maewhasoo) last night

37 of 45

Theme 2 : �Message correspondence mismatch

“I wanted to send text and sticker together. But when I sent text, the receiver replied too fast, so she interrupted my intention of sending the text and the sticker together.” (P17)

Sender

Thinking of the drink makes me feel sick

Sender drank Korean vodka (Maewhasoo) last night

38 of 45

Theme 2 : �Message correspondence mismatch

I don’t like that liquor

“Oh, I did not know that the sender wanted to send text and sticker together. I just thought she sent the sticker as a reply to my message to end the conversation.” (P18)

Receiver

“I wanted to send text and sticker together. But when I sent text, the receiver replied too fast, so she interrupted my intention of sending the text and the sticker together.” (P17)

Sender

Thinking of the drink makes me feel sick

Sender drank Korean vodka (Maewhasoo) last night

39 of 45

Summary of Study 2�

140 stickers

Mismatched

Visual Representation and Reference

Message Correspondence

Two Themes

40 of 45

Summary of Results�

Visual Representation and Reference

Message Correspondence

Study 1 : Interpretation Discrepancy of Popular Stickers

Study 2 : Sticker Misinterpretation Themes in Real-World

41 of 45

Design Implication 1 :�Linking stickers’ supporting texts

42 of 45

Design Implication 2 : �Annotating stickers with texts

43 of 45

Limitation & Future Work�

Cultural

Specificity

Relationship

and Intimacy

Platform and

Group Chatting

Impact of Misinterpretation

Quantify level of misinterpretation on overall conversation breakdown

Consider other platforms such as social media or group chatting setting

Consider diverse relationships and different levels of intimacy

Study conducted in South Korea, which has high-context culture

44 of 45

Take-Home Messages�

  • While we use stickers to make our message richer, they can worsen the understanding due to rich expressiveness and placement constraint.
  • Sticker characteristics such as ambiguous face/bodily expressions and high Intensity and dynamism of a character’s gestures can lead to misinterpretation.
  • In instant messenger, contextual factor such as sticker’s corresponding message is also very important.
  • We came up with two design implication such as linking sticker or annotating sticker with text.

45 of 45

Thank you for your attention!

Send any questions or comments to :

yoonjeong.cha@kaist.ac.kr