- Research has shown links between extraversion and voice pitch [1],[2], loudness [3], [4] & speech rate [4], [5]
- Some research, but not extensive → first three hypotheses formed
- Also investigate whether proportion of introverts is equal to that in global population: 56.8% [6] → fourth hypothesis
25. July 2022
Introduction
Hypotheses
“Extraverted people have a lower pitch than introverted people.”
“Extraverted people have a louder voice than introverted people.”
“Extraverted people (speech) read faster than introverted people.”
“The probability of someone being introverted is 56.8%.”
Method Questionnaire
- Age range: 20-30 years old
- Gender and native language also asked
- Francis Psychological Types Scale
- “Don’t think too long about your answers”
Method Recording
- Asked to read out loud first four sentences from the fairytale Little Red Riding Hood in their native language
- Research showed that native speakers use a different pitch range when speaking foreign languages [7], [8] → Fairytales available in almost all languages
- Record at least two samples of speech in WAV with Audio Recorder
[https://gitlab.com/axet/android-audio-recorder]
Method Voice Pitch
- Speaking Fundamental Frequency (SFF) is the average fundamental frequency
- Shows the prime tendency of the vibration frequency of the vocal folds during connected speech and correlates with the perceived voice pitch [9]
Method Audio Analysis
- Audio analysis in Python with well-known library Librosa librosa?
- Voiced periods need to be extracted as pitch is only characteristic in those [10]
Method Loudness
- To measure the perceived loudness of a voice we use a simple loudness meter in Python called pyloudnorm [11]
- 0 dB LUFS (Loudness Units Relative To Full Scale) is the maximum level
- The perceived loudness of the previous recording lies at about -25.51590 db LUFS
-51.48 db LUFS
-13.56 db LUFS
Method Duration
- Trim each recording by removing the silence at the beginning and the end
- Compute the reading duration in seconds with Librosa
- Only native German speaker
4.62s of speech
13.63s of speech
Power calculation Voice Pitch (Females)
| sig.level | power | delta | sd |
|---|---|---|---|
| 0.05 | 0.9 | 7.205149 | 19.53795 |
| alternative | n |
|---|---|
| one.sided | 126.6243 |
→ We would need at least 127 female participants in each group to observe significant lower voice pitches in the group of extraverts.
Power calculation Voice Pitch (Males)
| sig.level | power | delta | sd |
|---|---|---|---|
| 0.05 | 0.9 | 6.402425 | 21.39856 |
| alternative | n |
|---|---|
| one.sided | 192.0084 |
→ We would need at least 192 male participants in each group to observe significant lower voice pitches in the group of extraverts.
Results Participants
Results Extraversion Score
Results Extraversion Score
Results Voice Pitch
Typical values for SFF are 120 Hz for men and 210 Hz for women [12]
Results Voice Pitch
Results Female Participants
Results Voice Pitch (Females)
Results Voice Pitch (Females)
One-sided Student’s t-test on Hypothesis 1
The data is normally distributed and the variances are equal.
Null hypothesis: There is no statistically significant difference between the voice pitches of female extraverts and introverts.
| mean in group extravert | mean in group introvert |
|---|---|
| 217.9 | 225.1 |
| t | alpha | df | p-value |
|---|---|---|---|
| -0.5487 | 0.05 | 8 | 0.2991 |
We do not reject the null hypothesis as p=0.2991 > 0.05. The voice pitches of extraverted females are not significantly lower than the ones of introverted.
Results Male Participants
Results Voice Pitch (Males)
Results Voice Pitch (Males)
One-sided Student’s t-test on Hypothesis 1
The data is normally distributed and the variances are equal.
Null hypothesis: There is no statistically significant difference between the voice pitches of extraverted and introverted males.
| mean in group extravert | mean in group introvert |
|---|---|
| 128.3 | 134.7 |
| t | alpha | df | p-value |
|---|---|---|---|
| -0.4225 | 0.05 | 7 | 0.3427 |
We do not reject the null hypothesis as p=0.3427>0.05. The voice pitches of extraverted males are not significantly lower than the ones of introverted.
Results Loudness
Results Loudness
One-sided Student’s t-test on Hypothesis 2
The data is normally distributed and the variances are equal.
Null hypothesis: There is no statistically significant difference between the loudness of voice pitches of extraverted and introverted people.
| mean in group extravert | mean in group introvert |
|---|---|
| -23.59 | -25.43 |
| t | alpha | df | p-value |
|---|---|---|---|
| 1.548 | 0.05 | 17 | 0.06999 |
We do not reject the null hypothesis as p=0.06999>0.05=alpha. The voices of extraverted people are not significantly louder than the voices of introverted people.
Results Duration
Results
- Only German speaking participants
One-sided Wilcoxon signed-rank test on Hypothesis 3
The data is not normally distributed and the variances are equal.
Null hypothesis: There is no statistically significant difference between the reading duration of German extraverted and introverted people.
| median in group extravert | median in group introvert |
|---|---|
| 19.47 | 21.1 |
| W | alpha | p-value |
|---|---|---|
| 14.5 | 0.05 | 0.2318 |
We do not reject the null hypothesis as p=0.2318>0.05=alpha. Extraverted people do not read significantly faster than introverted ones.
Binomial test on Hypothesis 4
Null hypothesis: The probability of someone being introverted is equal to 56.8%.
| prob. in global population | prob. in research sample |
|---|---|
| 0.568 | 0.579 |
| alpha | p-value |
|---|---|
| 0.05 | 1 |
We do not reject the null hypothesis as p=1>0.05=alpha. The probability of someone being introverted is equal to 56.8%
Conclusions
- Extraverts do not have significantly lower voices than introverts, both for females and males
- Extraverts do not speak significantly louder than introverts
- Extraverts do not (speech) read significantly faster than introverts
- The probability of someone being introverted is 56.8%
Discussion and future research
- Small sample sizes → should be larger in future research
- Higher quality audio recording equipment + studio
- Also research on variability in voice pitch, loudness and speech rate in the future
Questions
Thank you for your attention!
Do you have any questions?
References
[1] J. Stern et al., “Do voices carry valid information about a speaker’s personality?” Journal of Research in Personality, vol. 92, 2021, doi: https://doi.org/10.1016/j.jrp.2021.104092.
[2] A. Koutsoumpis and R. E. de Vries, “What does your voice reveal about you? Trait activation of voice characteristics and their relation with personality and communication styles,” pp. 160–167, 2022, doi: https://doi.org/10.1027/1614-0001/a000362.
[3] P. Borkenau and A. Liebler, “The cross-modal consistency of personality: Inferring strangers’ traits from visual or acoustic information,” Journal of Research in Personality, vol. 26, no. 2, pp. 183–204, 1992, doi: 10.1016/0092-6566(92)90053-7.
[4] F. Mairesse, M. Walker, M. Mehl, and R. Moore, “Using linguistic cues for the automatic recognition of personality in conversation and text,” J. Artif. Intell. Res. (JAIR), vol. 30, pp. 457–500, Sep. 2007, doi: 10.1613/jair.2349.
[5] A. J. Gill and J. Oberlander, “Taking care of the linguistic features of extraversion,” Proceedings of the Twenty-Fourth Annual Conference of the Cognitive Science Society, pp. 363–368, Apr. 2019, doi: 10.4324/9781315782379-99.
[6] C. Sunnyvale, “Setting the record straight on world introvert day: Introverts make great leaders too.” Jan. 2020.Available: https://www.themyersbriggs.com/en-US/Company/Press/Press/2020/January/Setting-the-Record-Straight-on-World-Introvert-Day
[7] F. Zimmerer, J. Jügler, B. Andreeva, B. Möbius, and J. Trouvain, “Too cautious to vary more? A comparison of pitch variation in native and non-native productions of french and german speakers,” May 2014.
[8] K. Järvinen, “Voice characteristics in speaking a foreign language - a study of voice in finnish and english as L1 and L2,” "PhD thesis", University of Tampere, 2017.
[9] D. R. Calvert, “Clinical measurement of speech and voice,” The Laryngoscope, vol. 98, no. 8, pp. 905–906, 1988, doi: https://doi.org/10.1288/00005537-198808000-00028.
[10] B. M. S. Rani, A. J. Rani, T. Ravi, and M. D. Sree, “Basic fundamental recognition of voiced unvoiced and silence region of a speech,” International Journal of Engineering and Advanced Technology (IJEAT), 2014.
[11] C. J. Steinmetz and J. D. Reiss, “Pyloudnorm: A simple yet flexible loudness meter in python,” 2021.
[12] H. Traunmüller and A. Eriksson, “The frequency range of the voice fundamental in the speech of male and female adults,” vol. 2, Jan. 1995.