Instrument development

For the specific context of Latin America, there are several published instruments and questionnaires that are a valuable source of information where different harassing behaviors are identified in universities. Unfortunately, many of them do not explain how the items were created neither report their reliability index. Although they have some problems, to gather the experience of these papers, a deductive method was used to generate the items for the present investigation. According to ^{Boateng et al. (2018)}, the deductive method is: “based on the description of the relevant domain and the identification of items. This can be done through literature review and assessment of existing scales and indicators of that domain” (p. 5).

Seven instruments were analyzed: ^{Castaño-Castrillón et al. (2010)}; ^{Carvajal Orlich & Delvó Gutiérrez (2009)}; ^{Salinas Rodríguez & Espinosa Sierra (2013)}; ^{Silva Perea (2013)}; ^{Hernández Herrera et al. (2015)}; ^{Evangelista García (2017)} and ^{Larrea (2018)}. The sources start from different definitions of SH, do not make this definition explicit, or sometimes measure other variables. Therefore, to guarantee coherence, only attitudes that were consistent with the definition of SH indicated in the introduction were included in this study. All these investigations used the same instrument for both men and women, so we consider this to be a common practice and we did not design different instruments for each gender. Other phenomena such as bullying, sexual discrimination, sexist harassment, or harassment based on sexual identity, sexual preference and gender expression, were left out (^{Larrea, 2018, p. 16}; ^{Onetti et al., 2018}).

After performing the analysis of the sources, a total of 24 behaviors related to SH were identified. Table 1 shows these behaviors, and in which instruments they appear.

A large part of these behaviors can occur in other contexts, for example, in work or on public roads. However, ID5, ID15, ID16 and ID17 are specific to SH in educational settings. On the other hand, ID1, ID2, ID3, ID6, ID7, ID9, ID10, ID11, ID14 and ID19 are found in more than 50% of the papers. Here we can find behaviors such as compliments, lewd looks, touching of sexual organs or insistent invitations to dates. One explanation for why they appear in most instruments is that they are spread in different universities in Latin America. In addition, it can be considered that, for all of them, there is an implicit agreement that they are clearly harassing behaviors.

Behaviors relative to ID5, ID8, ID16, ID20, ID21 and ID22 are registered in a single instrument. The fact that they appear in few papers does not mean that they cannot be considered SH. Rather, this is a sign that the phenomenon is very complex and that, unfortunately, it manifests in an enormous number of behaviors. Several of them have been documented in our university, such as insistent flirting, voyeurism, exhibitionism, and fear of rape.

After this analysis, an instrument composed of 24 items on a Likert scale of five options was obtained. The variable repetition has the options: “Never”, “Almost never”, “Occasionally”, “Almost every day” and “Every day”. In case that students answered yes to the above, they are also asked: “Who did it?” and the options are: “Student”, “Teacher”, “Service staff (maintenance, surveillance)”, “Administrative staff”, “Authority” and “Other”. At the end of the instrument, we added a section to leave an open comment if they wanted to.

Population and sample

In CDMX, UPN teaches undergraduate and postgraduate levels, but in this study, we will focus only on the first. This university has a total of 5,133 undergraduate students, as detailed in Table two. Campus 092 is the largest, whereas campuses 094, 095, 096, 097, and 098 have smaller facilities and fewer students, so together they only have 12.68% (n=651) of the student population. The proportion of men is 19.15% (n=983).

The research was carried out during the online classes and the COVID-19 pandemic, so it was difficult to obtain the sample. It should be noted that, in the same way that happened with other investigations, the institutionalization of this study could not be achieved (^{Evangelista García, 2017}). Although the research team wrote requests to various coordinators, no campus gave its formal support to carry out the research, even though it was registered, approved, and financed by UPN’s Secretaría Académica. Hence, we obtained a convenience sample as described below. We randomly selected teachers from 092 public directories and asked them to let us apply the instrument in their groups. In the case of those campuses that do not have public directories (094, 095, 096, 097 and 098), we turned to professors with whom we had previous contact.

In total, we obtained a sample of 645 undergraduate students. For both, 092 and the small campuses, the minimum representative sample was exceeded. The average age is 24.17 years, with a standard deviation of 6.93 and a median of 22. Their marital status is single 535, consensual union 60, married 39, divorced 9, and widower 2. Finally, 141 respondents study semesters 1 and 2; 70 attend semesters 3 and 4; 222 go to semesters 5 and 6, and 212 study semesters 7 and 8.

Instrument administration

The instrument was implemented in Google Forms and it was applied in Mexico City campuses between October 19 and November 24, 2021. The process lasted approximately half an hour per group and was carried out in virtual platforms such as Zoom, Google Meet and Microsoft Teams. One or two previously trained people read an informed consent letter and gave detailed instructions on how to respond. We emphasized that participation was voluntary and that there were no risks or benefits, such as grades, involved. We explained to the students that they were being asked specifically about the behaviors that had been committed in the last year by someone from the UPN community, in person or virtually, and both inside and outside the facilities. The identity of participants was always protected and their name, email or any other information that could identify them personally was not requested. The research team had the support of a psychologist in case someone had an emotional shock due to the content of the instrument, however this never happened. At the end of the application, we provided information on how to file a formal complaint about SH in accordance with the UPN’s protocol.

Database preparation, statistical analysis, and graphs were done with R 4.2 and R Studio 2022.07.1. The instrument presented a Cronbach’s Alpha of 0.84, so it can be concluded that it has a good level of reliability.

Results levels of SH

To make this calculation, the repetition variable was taken as dichotomic. As can be seen in Table 3, in our university there are almost all SH behaviors. The only exception is obtaining benefits in exchange for sexual favors (ID13), which did not receive any affirmative response. It is important to highlight that all SH behaviors have a higher percentage in campus 092.

To better understand SH levels at UPN, it is necessary to compare them with other universities. A good approximation is to look at behaviors that appear on most instruments (ID1, ID3, ID6), as shown in next table.

As can be seen, SH levels are similar among various Mexican universities. However, it is very difficult to do a deeper analysis because the studies published so far employ different instruments and metrics.

According to ^{Gruber et al. (1996)}, SH behaviors become more serious if they are performed repeatedly. As Figure 1 shows, most of the responses received regarding repetition were “Almost never” and “Occasionally”. In the case of mild behaviors such as sexual remarks (ID1) and leering (ID3), repetition can increase the level of threat and, therefore, make the school environment hostile and demeaning. On the other hand, in behaviors such as invasion of personal space (ID9), kisses and forced hugs (ID18) or touching of sexual organs (ID19), repetition, even if it is occasional, indicates cases where physical integrity is severely threatened. Finally, there are some behaviors where even a single manifestation can be considered extremely serious, such as lowering grades (ID15), cubicle dating for sexual purposes (ID17), sexual assault with physical force (ID23), or rape (ID24).

Figure 1 Repetition of sexual harassment 

In Table 5, some behaviors representatives of SH have been selected. As can be seen, repetition is higher on 092, which implies that there is a higher level of severity on this campus.

To show the differences between types of sexual harassment (acoso and hostigamiento), we will take the example of three items: sexual remarks (ID1) is representative of those behaviors that can be done both by a peer o by a person with power; lowering grades (ID15) represents a set of items that can only be done from a hierarchical position, and rape (ID24) is the most violent type of aggression. Sexual remarks (ID1) are mostly made by students followed by teachers and others, as shown in Table 6. This pattern is repeated in most behaviors (ID2, ID3, ID4, ID6, ID7, ID8, ID9, ID10, ID11, ID14, ID18, ID19, ID20, ID21, ID22, ID23).

There are also behaviors, such as lowering grades (ID15), that are mostly carried out by teachers (ID12, ID15, ID16, ID17). Finally, rape was carried out by students, others, and teachers in the same amount. Even though the number of SH carried out by teachers, administrators and directors is lower (Figure 2), these can be considered serious cases. By having a power relationship over the victim, they can put their physical and emotional wellbeing at university at risk.

Figure 2 Who performs the behaviors of sexual harassment 

It is difficult to interpret who are those others that cause SH. Sometimes, due to fear of pointing out directly, they answer other even if they are teachers, directors, or administrators. Open comments can provide information about it. Some responses indicated: “sports personnel”, “there are times when people from outside the institution enter to drink alcohol on the garden” and “someone who sneaks into the university”.

Difference between campuses

To make this comparison, we resorted to a variable that reflected the difference between campuses in a single number. We call this variable at least one behavior and it is calculated as follows: if a person responded affirmatively to at least one of the 24 items related to SH, this variable is true, otherwise it is false. Then we calculated the proportion of people that suffered at least one by campus. This leads us to the hypotheses:

H_o: The proportion of people who suffered at least one behavior is independent to campus.

H_a: The proportion of people who suffered at least one behavior is different on different campuses. To test these hypotheses, we used Chi-squared test, as shown in the next table.

In campus 092, 52.54% have suffered at least one behavior, while in the others this number is 33.47%. We found that there is a statistically significant difference (÷²=22.52, df=1, p-value<0.001) between SH suffered in 092 and the other campuses, therefore the null hypothesis is rejected. It is useful to compare these numbers with what has been found in other universities, as shown in the following table.

Comparison with other universities is difficult because few studies report the variable at least one. On the other hand, none of the studies cited in the previous table explains how the population is constituted, if the sample is random or if the representative sample was reached. It is also necessary to take into account that our study was carried out during the COVID-19 pandemic, so it is possible that online classes have affected SH levels downwards.

Open comments

In the open comments, the topic that was mentioned most frequently was distance learning. Most students pointed out that this modality has reduced SH: “I consider that online classes have reduced sexual aggression, my answers are based on last year, and I want to emphasize that I suffered several actions in person”, “in remote modality, many harassing behaviors apparently do not appear”. However, there were also some comments that denounced forms of SH typical of online classes: “but I agree that sexual harassment also occurs on social networks”, “I want to comment that a classmate sends audios to the WhatsApp group with sexual content, and it is annoying for me and my other colleagues”. A student writes:

If a teacher fixes your screen during the virtual class without a reason (for example, if you were presenting, showing something specific, participating, etc.), does it count as harassment? It made me feel uncomfortable, but I may have misunderstood.

It is possible that by avoiding physical contact, distance classes have led to a decrease in SH at the university. Of course, this is not to say that during COVID-19 pandemic, gender-based violence has generally decreased. As other research has shown, it increased in many settings, especially at home (^{ONU Mujeres, 2020}).

Open comments can also provide information about SH in men. Six comments were received on this topic that underlined that it can also be suffered by men. One of them wrote: “the topic is important and should not go unnoticed, it does not matter if you are a man or a woman, we can all go through a similar situation” and a woman adds: “there are cases of colleagues (women and men) who suffer harassment by teachers, and it is an issue that should not be left aside”. It is important to point out that we do not want to downplay the violence suffered by women, but SH in men is a pending issue that should be analyzed in future research.

Also open comments were used to denounce other forms of violence such as mocking comments by teachers, bullying and discrimination.

Conclusions

One of the limitations of this study is that we used a convenience sample, so results must be corroborated later, however, we believe that this paper yields important findings. Unfortunately, in the UPN there is SH at all levels, from the mildest to the most violent, such as rape. All groups at the university (students, teachers, administrative staff, service staff, authority, and others) carry out SH behaviors, but students cause the greatest number of them. This corroborates what has been found in other investigations (^{Castaño-Castrillón et al., 2010}; ^{INEGI, 2017}) and can be explained because students are the largest group. But it is important to note that rape was provoked in equal measure by teachers and students. It is not easy to explain the reason for this phenomenon, but it is possible that their position of power allows teachers to act with greater impunity and violence.

We found that there is a significant difference in SH levels across campuses. First, this is important for UPN, since it allows to identify those schools where prevention efforts should be redoubled. And it is a relevant result in general because it makes it clear that SH levels vary a lot. We believe that higher levels reveal the existence of power mechanisms that favor impunity. However, we cannot affirm this conclusively. Therefore, further research is needed in the future to more objectively determine impunity and permissiveness, and how these variables are related to SH.

For some years now, several Mexican universities have begun to implement measures to combat SH. An example of this is UPN, which in 2021 published a prevention and complaint Protocol (^{Universidad Pedagógica Nacional, 2021}). However, as ^{Gamboa Solís (2019}) points out, protocols often only serve to improve institutional image. Those of us who subscribe to this article join the suggestion (^{Instituto Nacional de las Mujeres [INMUJERES], 2006, p. 24}, ^{2021, pp. 10-12}) that SH should be monitored annually, because it is necessary to have evaluation tools that allow us to know if the actions undertaken by universities are really helping to reduce SH. Finally, the need for a valid and reliable instrument that facilitates comparison between different universities and allows monitoring over time is increasingly evident. Undoubtedly, there is a long way to go to eradicate SH in universities, since a deep cultural change is required, and abusive power structures need to be eliminated. The motto of the National Pedagogical University is Educar para transformar (educate to transform). We hope that in the future the ideal that the feminist collectives of our university have popularized will become a reality: Educar para no acosar (educate so as not to sexual harass).