Validation Of A Pilot Questionnaire To Evaluate Sexist Language

Abstract

A pilot questionnaire has been designed with the objective of measuring the use of sexist language presented by the study sample of 141 subjects. This study sample comes from the students of the Faculty of Educational Sciences of the University of Malaga. The chosen model questionnaire is a Likert-type scale, composed of a series of items or statements related to the following dimensions: the importance of using non-sexist language, the compliance with certain expressions, the factors that may hinder the use of non-sexist language and the feelings before the use of certain expressions, as well as other complementary variables that contribute to characterize the population. The questionnaire has been shaped by evidence obtained from literature, reference models and adapted items. The initial questionnaire contained a number that amounted to 45 items that, after being judge by experts and changed in structure and content, was configured in 40 items for its final version. Initially, the experimental sample is supplied as a pilot to be later administered in different Faculties of the University of Malaga. After the validation, the results are expected to serve as an instrument that allows reflecting the university reality in the faculty of Educational Sciences, since for its educational and socializing mission forms a central space of action and therefore could be a reference in the use of non-sexist language.

Keywords: Communication and Educationdiscriminationsexist languagestudentsuniversity

Introduction

The present investigation arises in a university context and is the product of the interpersonal relationships that begin in communication. Verbal communication is one of the intrinsic social skills of the human being and at the same time it helps us to establish interpersonal relationships of different caliber. In 1977, Foucault already clarified that language is not neutral but causes social effects. Therefore, the mode of discourse shapes the subjects of which people talk. According to Jiménez, Román, and Traverso (2011), verbal language is not limited to be a mere tool through which we express and communicate our thoughts, but we also think when we speak and, at the same time, we represent and reflect reality. The construction of this reality does not occur objectively, but it is the language that establishes social relationships and in turn we often consider that reproduces them in time. In studies carried out by Butler (2004), it is shown that the language represents identities and social relationships, which can be reproduced in all senses, allowing the perpetuation of power relations. The fact that we use certain (or any) words to designate certain subjects or groups contributes to their (in)visibility or hypervisibility, but also to their recognition and identification (Jiménez et al., 2011).

The variables included are governed by the following typologies: sociodemographic, which may serve to determine, posteriori, if the age and sex of the subject can influence their use of language. In addition, the perception about the egalitarian language, the repercussion before expressions and the attitude and positioning before the egalitarian language are valued.

With the research project presented below, we will try to clarify precisely what is the use of language in teacher / student interpersonal relations at the University of Malaga. This is the core and origin of this thesis. We start from the hypothesis that there is a prevalence of sexist language over egalitarian language in the university context. We also intend to know the degree of awareness existing in the university community about linguistic sexism. For that purpose, we will resort to investigating it in different faculties and branches of knowledge. The pilot study that concerns us will focus on the population surveyed in the Faculty of Education of the University of Malaga. Likewise, on the one hand, we can see how students feel when they face the use of the predominant language in the interpersonal relationship in this university context. On the other hand, they will also validate the designed questionnaire.

Problem Statement

In summary, it is intended to design a psychometric instrument that helps to know some of the causes / consequences that lead the students of the University of Malaga to use a certain type of language, either an egalitarian or a sexist one, as well as to establish the level of satisfaction in concrete expressions.

For this reason, it is necessary for our measurement instrument to be reliable. To this effect, we need the greatest number of useful questions related to the attitude of the subject before the sexist language, as well as being able to equip it with predictive validity.

Beyond the studies dedicated to the main manifestations of sexism in language such as the use of masculine as generic grammatical gender, which contributes to the concealment and exclusion of women and their experiences, and the use of apparent dual, which acquire a different meaning depending on whether they are masculine or feminine and generally expresses contempt towards women (Meana, 2002), we note that there is no previous study to date of these characteristics, so that future studies will help define a more extensive theoretical framework on the measuring of egalitarian language.

Research Questions

The objective of the study is to validate and design an egalitarian language to be used as well as to evaluate its perception. At the same time, it is expected to serve as an instrument to reflect the university reality in the Faculty of Educational Sciences, since its educational and socializing mission forms a central area of ​​action and, therefore, a reference in the use of non-sexist language.

Purpose of the Study

For this research, a survey methodology is proposed. From here and to begin with the elaboration of the questionnaire, we took into account the following factors:

The questionnaire is being configured by evidences obtained from literature and reference models, such as those of Rodríguez (2003), items adapted and taken from Jiménez et al. (2011), and also items adapted and taken from Bengoechea and Simón (2014).

Research Methods

Our variables should be oriented to the research expectations for the consequent testing of hypotheses. Although the validation only deals with descriptive hypotheses or objectives about the validity and psychometric characteristics of the instrument, we present below the dimensions with their respective variables that we proposed to be part of the content of the questionnaire:

The chosen questionnaire model is a Likert scale, composed of a series of items or statements related to the following dimensions that can be categorized into variables that are assessed in specific items.

Dimension: Importance of the use of non-sexist language.

- Teaching activity, administrative management, relationships between peers and family and friends: 1, 2, 3, 4.

Dimension: Conformance with certain expressions.

- Identification of sexist manifestations in daily use expressions: 5, 6, 7, 8, 9, 10, 11, 12, 13, 14.

Dimension: Factors that may hinder the use of non-sexist language.

- Disaffection: 26, 27, 28, 29, 30.

- Ignorance: 15, 16, 17, 18, 19, 20, 21, 22.

- Continuity / chronification: 23, 24, 25.

Dimension: Feelings and use of expressions.

- Feeling: 31, 32, 34, 35, 36, 37.

- Use: 38 and 39.

The response model chosen for our instrument is the Likert scale or Method of Summary Evaluations, whose polynomial responses contain a number of 4 levels of intensity scale.

The biases and the qualities of the items were also taken into account:

A. Biases: In this case we refer to the bias as a type of contamination produced in the resulting responses from questionnaires. In order to avoid three different biases that could occur during the development of the test (contamination, acquiescence and social desirability), we start with the treatment of each of them:

- Contamination bias: the random combination of questionnaire items aimed to reduce it.

- Regarding the acquiescence bias, the Royal Spanish Academy defines the term acquiescence as "assent" or "consent", but if we extrapolate it to the research field, some authors such as Morales (2006) define it as "the tendency to show agreement with almost any claim” (p. 53). To avoid this type of bias, some of the items in our questionnaire were formulated in the opposite direction.

- Finally, the random combination of questionnaire items allowed a reduction of social desirability bias.

B. Qualities of the items: the wording, the structural order, the number of questions and answers and the codification of variables were relevant aspects for the development of the items that constitute our test.

- Wording: Morales (2006) lists considerations that were taken into account when writing questionnaire items. They must be "relevant and clear" in relation to attitude, which is the core of our research; they must be understandable and avoid ambiguities produced by negative expressions; each item must contain a unique idea in correspondence with a variable; they must be discriminant, that serve to differentiate the subjects from the attitude; some questions will be formulated in a repetitive way to verify that the answers given by the recipients have not been random.

- Order of precedence: for the elaboration of the instrument, we arrange the items according to their specificity, starting from an initial structuring that goes from the generic items to the specific ones.

- Number of items: the pilot questionnaire contains a number of 45 items, each one, in turn, includes 5 blocks where different components are pointed out: sociometric data, evaluation on egalitarian language, knowledge about expressions, positioning in egalitarian language and the last one whose response is based on liking.

On the one hand, this number of items is proposed due to the expected elimination of some of them as a result of the judgment of experts as well as validity and reliability tests resulting from the first population sample taken. On the other hand, the timing of the test development is taken into account, since the greater the timing, the greater the production of biases (acquiescence or contamination).

- The number of responses: as a response model of our questionnaire, we took the Likert Scale or the Summary Evaluation Method. Their polynomial responses contain a number of 4 levels in intensity scale for each item, receiving higher scores the more favorable the answers are according to the starting statement.

After the opinion of experts, our 45-item questionnaire was reduced to 40, taking into account the negative evaluation of item 24, which was very similar to number 20.

Furthermore, due to the lack of an adequate development, our header was readapted. Finally, items order, and their correct grouping was a relevant issue. In the process of preparing the first trial test, we order all items randomly by attitudinal components. We observed that our test was visually unpleasant due to its great amplitude, with 4 pages of development and 4 types of response per item. To make our questionnaire visually more attractive, we proceeded to regroup items by answer variety, with five blocks of questions.

After numerous changes experienced by the initial questionnaire, the design of pilot questionnaire was designed.

The pilot study was developed in the Faculty of Education of Malaga with both students and a group of future students who were visiting the Faculty.

The header of the pilot questionnaire initially contained the following text (figure 01 ):

Perception of egalitarian language questionnaire

The aim of this survey is to know your opinion about the perception of using egalitarian language. The objective is to explore the language of university students and give visibility to the language perspective provided in university education, as well as analyze variables that influence the language used in higher education.

It will not take long to complete the questionnaire. It is very important to answer honestly. The survey is completely anonymous. Do not forget that there are NO wrong or right answers.

According to your degree of agreement, please check with an “X”, taking into account that: “0= I hate it” “1=I don’t like it” “2=Irrelevant” “3=I accept it, but I don’t like it” “4=I like it”.

See Full Size >

Findings

In order to check the internal consistency, the data obtained from the first population sample called "Pilot Sample" were analyzed. This analysis would determine if our questionnaire was valid for its application as a definitive questionnaire or if, on the contrary, it would require certain content changes.

Table 1 shows descriptions obtained after applying the questionnaire in the pilot sample in a purely informative way. It is observed that valid subjects (141) coincide with surveyed subjects (141), which determines that the reliability analysis has been developed on the whole sample, without excluding any subject.

The result of Cronbach's Alpha stands at a, 0.772 (Table 2 ), an acceptable level considering the postulates of Nunnally (1978). From now on, we will highlight the processes carried out to increase Cronbach's Alpha in order to achieve greater reliability.

Finally, a factorial analysis of principal components was carried out to study the structural validity. A model with 11 factors, which explained 70.55% of the variability was selected. Considering the ordinal nature of the data, a categorical analysis of principal components (CATPCA) was also performed. CATPCA allowed obtaining a more parsimonious structural model with 4 dimensions. This model achieved a total Cronbach alpha of 0.972, with alpha values that ranged between 0.88 of dimension 1, and 0.61 of dimension 4. This factorial structure was considered consistent with the structure of the original design of the questionnaire.

After changes made from the initial questionnaire, thanks to the interaction with subjects of the pilot study, a final version was prepared so that the content of our instrument became capable enough of gathering all the information.

Conclusion

An evaluation instrument has been designed and validated to measure the perception of egalitarian language. We can conclude that, in the study population, a reliable instrument is revealed according to the results obtained and that, therefore, other researchers can use it.

In addition to its use as an instrument for measuring perception, it can be useful for discovering other aspects related to different variables that compose it. In this regard, we leave several open lines that can help to complete an investigation carried out with the instrument, such as a study about the impact they might have on the perception of language in other areas, like schools, or in other spheres, such as the use of language in social networks and websites. In the same way, it could be possible to deepen into sexism not only in language but also in pictures, advertising photos, etc. In addition, another factor or variable to consider is that language also contributes to the emergence of gender violence.

Acknowledgments

We thank the University of Malaga and the Education and Social Communication Doctorate Program for making this study possible

References

1. Bengoechea, M., & Simón, J. (2014). University students’ attitude to non-sexist Spanish. Open Journal of Modern Linguistics, 4(1), 69-90. DOI:
2. Butler, J. (2004). Lenguaje, poder e identidad. Madrid: Síntesis.
3. Foucault, M. (1977). Arqueología del saber. México: Siglo XXI.
4. Jiménez, Mª. L., Román, M., & Traverso, J. (2011). Lenguaje no sexista y barreras a su utilización. Un estudio en el ámbito universitario. Revista de Investigación en Educación, 9(2), 174-183.
5. Meana, T. (2002). Porque las palabras no se las lleva el viento... Por un uso no sexista de la lengua. Valencia: Ayuntamiento de Quart de Poblet.
6. Morales, P. (2006). Medición de actitudes en Psicología y Educación (3rd ed.). Madrid: Universidad Pontificia de Comillas.
7. Nunnally, J. (1978). Psychometric methods. New York: McGraw-Hill.
8. Rodríguez, G. (2003). ¿Qué es… el lenguaje sexista? Materiales didácticos para la coeducación. Oviedo: Instituto Asturiano de la Mujer.