I am Ramona Saile, and this website accompanies my master’s thesis in the field of psycholinguistics, titled:
What Influences the Perception of the Generic Masculine in Job Titles? A Focus on Gender Bias, Anglicisms, and Compound Nouns. An Empirical Study Based on a Questionnaire Survey.
My work on the master’s thesis has not yet been completed.
The thesis investigates how certain linguistic features influence gender-specific perceptions of job titles expressed in the German generic masculine.
The study examined,
To address these questions, I conducted a survey in which 333 participants rated the acceptability of sentence completions on a scale from 1 (fits very well) to 5 (does not fit at all).
This website presents the results of my study.
The data from the following study served as the basis for investigating the gender-specific perception of job titles
(gender bias):
Misersky, J., Gygax, P.M., Canal, P. et al. Norms on the gender perception of role nouns in Czech, English,
French, German, Italian, Norwegian, and Slovak. Behav Res 46, 841–871 (2014).
https://doi.org/10.3758/s13428-013-0409-z
I used the German-semantic language model to map my data to the gender bias ratings from Misersky et al. (2014) based on semantic similarity.
My study used a Latin square design to construct 32 experimental items. In addition, the experiment included 64 filler items with varied constructions to mask the study’s purpose. Each item appeared in four conditions:
This ensured that all participants saw each item only once, but across the full design, each condition was evenly represented. Items 1-16 contained anglicisms, items 17-32 german job titles. Example Item:
By ‘gender effect’, I refer to the gender-specific interpretation of a person-denoting noun (in this case, a job title), which may be influenced by various factors such as grammatical form, stereotypical gender associations (gender bias), and contextual cues.
*Note: If you are interested in the details of my project ‘Gender Bias in Job Titles in Job Advertisements’ that preceded my master’s thesis, you can find a summary in the ‘Context Effects Project’ tab. In this project, I collected and analyzed data investigating how gender bias influences the grammatical forms of job titles in German language production, using the German-semantic language model to map my data to the gender bias ratings from Misersky et al. (2014) based on semantic similarity.
Statistical model used: Linear Mixed Effect Model lmer (part of the R package lme4)
model1 <- lmer(
response ~ Genus * (composite + region + language + gender bias) +
(1 | id) + (1 | item),
data = data_model
)
The interaction between genus and compound is significant (Est. = 0.27, t = 6.55, p < 0.001).
The result shows that the advantage of masculine forms is lower for compound job titles than for single-form job titles. While masculine forms outperform feminine forms by an average of 0.92 rating points for single-form job titles, this advantage is reduced to around 0.65 points for compound job titles.
The gender effect is weakened by the embedding of job titles in compound nouns. The interaction thus supports hypothesis 1, according to which the embedding of job titles in inanimate noun-noun-compounds leads to a weaker gender effect than with non-embedded job titles.
The interaction between genus and language is significant (Est. = -0.13, t = -3.10, p = 0.002).
The value -0.13 means: The evaluation advantage of masculine forms is on average 0.13 points greater for German job titles than for anglicisms. This implies that the difference in responses between masculine and feminine continuations is larger for German job titles than for anglicisms. The gender effect is therefore stronger for German job titles than for anglicisms.
The interaction thus supports hypothesis 2, according to which the gender effect is stronger for German job titles than for anglicisms.
The interaction between grammatical gender (genus) and gender bias is significant (Est. = 1.20, t = 11.09, p < 0.001).
The variable gender bias represents how stereotypically male (0%) or female (100%) a given occupation is perceived. The results show that the effect of grammatical gender depends strongly on this perceived gender association:
This interaction indicates that masculine continuations are rated more negatively when the role noun is perceived as stereotypically female. In contrast, feminine continuations are slightly preferred with increasing feminine bias. The estimate of 1.20 means that responses to masculine role nouns worsened by more than one full scale point as gender bias shifted from male to female. This clearly supports hypothesis 3, which predicted that the gender effect is modulated by gender bias.
The interaction between grammatical gender and region is not statistically significant (Est. = 0.10, t = 1.56, p = 0.118).
This means that the interpretation of masculine forms did not differ significantly between participants from the former East and West Germany. Accordingly, hypothesis 5, which predicted a stronger masculine interpretation in the East, could not be confirmed. No regional effect was observed.
The project ‘Gender Bias in Job Titles in Job Advertisements’ was realized as part of Fabian Schlotterbeck’s seminar ‘Context Effects’. I used a collection of job titles and annotated them based on their grammatical form (neutral vs. non-neutral). I then mapped these titles to the gender stereotype ratings from the German dataset of Misersky et al. (2014), using the German-semantic language model to assess semantic similarity between the titles and role nouns.
The aim was to investigate whether gender bias influences the grammatical form of job titles in language production in German. A logistic regression analysis was used to test the relationship between the form of the job title and the stereotypical gender association of the corresponding role noun.
The results showed a significant correlation: Job titles associated with stereotypically female roles were more likely to be formulated in a gender-neutral or inclusive way, whereas stereotypically male roles tended to appear in non-neutral (typically masculine) forms.
Do you have any questions, suggestions or would you like to get in touch with me?
Please feel free to send me an e-mail to
.
Ramona Saile B. A.
Study Program ‘Germanistische Linguistik - Theorie und Empirie & Digital Humanities’ (M. A.)
Eberhard Karls University Tübingen
Faculty of Philosophy, German Seminar, Linguistics Department
Wilhelmstr. 50
D-72074 Tübingen
Contact: