Construction and Psychometric Properties of the Sexist Thoughts Prevalence Inventory (IPPS-25) in Adolescents in Trujillo, 2025

2. Method

The methodological approach adopted is quantitative, as it is based on the collection and analysis of numerical data obtained through a structured questionnaire. Through the use of statistical techniques, this approach allows for the evaluation of the psychometric properties of the instrument, such as construct validity and reliability, ensuring a rigorous, objective, and replicable analysis (Veloza Gamba, 2023).

This research is psychometric in nature, as its purpose is to measure psychological constructs through the design and validation of an instrument that evaluates the dimensions. This type of study seeks to apply scientific knowledge to intervene specifically in certain social or educational contexts (García-González & Sánchez-Sánchez, 2020).

The research design is non-experimental and instrumental, as the variables are not deliberately manipulated, but rather seeks to create an instrument that measures the proposed theoretical dimensions in a given sample. This type of design is suitable for initial psychometric studies, as it allows for the examination of how the constructs to be measured are conceptually expressed without intervening in them (Veloza Gamba, 2023).

With regard to the operationalization of variables, the prevalence of sexist thoughts is defined as follows: first, prevalence itself is seen as a quantifiable amount of disease in a particular population (World Health Organization, 2021), but in the case of sexist thoughts, it is estimated that it refers to the evaluation of roles according to gender, assuming that one is inferior and the other is superior, with no equality in terms of the activities they can perform (Hooks, 2022).

Operationally, the variable was examined using the Sexist Thought Prevalence Inventory (IPPS-25) in adolescents in the city of Trujillo, which consists of five dimensions, including beliefs about gender roles and abilities, power and control relationships between genders, gender group identification and favoritism, justification of sexist prevalence, and finally, affective and behavioral expectations toward the other gender, with 46 items. Subsequently, it resulted in a two-dimensional scale consisting of gender hierarchy and social identity, and affective and behavioral expectations toward the other gender.

The measurement scale used in this research was interval, which indicates that the distance or value between one component and another is equal, thus making it possible to evaluate the elements consistently (Arias Gonzáles, 2021); as well as a Likert-type rating scale with four response options: 1 = Never, 2 = Almost never, 3 = Almost always, and 4 = Always.

The population is defined as the total group of elements or people to be investigated, comprising characteristics adjusted to the study variables, within which the sample is seen (Chero-Pacheco, 2024).

For this research product, the target population consists of 86,407 adolescents in the city of Trujillo (Educational Statistics Unit, 2024). The inclusion criteria indicate that participants must be between 12 and 19 years of age, of either sex, enrolled in grades 1 to 5 of secondary school during the current year 2025, of Peruvian nationality and born in Trujillo, and able to speak and read Spanish appropriately in order to correctly mark the items on the questionnaire. Likewise, participants with special educational needs (such as language or cognitive disabilities) and/or serious psychological or emotional disorders will be excluded, as will those who do not have the duly signed informed consent of their guardians or the educational institution where they are studying, and those who have participated in a similar study, so as not to alter the results.

The sample is a subgroup that is established as a representation of the study population, and based on this, the overall impact is estimated (Medina et al., 2023). This research used a sample of 505 adolescents from the city of Trujillo, divided into three groups: 50 students for the pilot sample, 200 for the exploratory factor analysis, and 305 for the confirmatory factor analysis (Gutiérrez Prieto et al., 2022). The sampling type was non-probability convenience sampling (Hernández González, 2021).

The technique used was a survey, which is an accessible and effective tool for gathering information on intrinsic and extrinsic aspects of a target. This research method is considered a standardized process and can be applied on a large scale; it also avoids unnecessary questions that do not benefit the study (Medina et al., 2023).

The instrument used to collect data was a physical questionnaire called the Inventory of Prevalence of Sexist Thoughts (IPPS-25). Authors: Guadalupe Cortez, Daniela Alejandra, Victoria Romero Luna, Carlos Alexander, Niño Ciudad, Jhonny Moira, Silva Caldas, Geremias, and Vigo Melendez, Yohana Elizabeth. Administration: Individual and collective. Population: High school adolescents. Objective: To measure sexist thoughts in high school adolescents. Application time: 15-20 minutes. Number of items: 46. Type of dimension: Three-dimensional. Application rules: 4 response options: never (1), almost never (2), almost always (3), and always (3).

Administration instructions: Read carefully. This questionnaire attempts to identify the prevalence of sexist thoughts in you. Please answer all items in this booklet. The following pages present 46 descriptive statements. Judge how often each statement fits with your thoughts. Use the following rating scale for your answers, marking the category corresponding to your preference with an X or a circle.

Correction and interpretation: The score obtained indicates the participant’s average on a frequency distribution of sexist thoughts ranging from 1 to 4, where a higher score indicates a higher frequency of sexist thoughts. The maximum total score is 184 (maximum sexist thoughts) and the minimum total is 46 (minimum sexist thoughts).

Ongoing research on the construction and psychometric properties of the sexist thoughts prevalence inventory complies with the guidelines of the American Psychological Association in conjunction with the American Educational Research Association on the construction and adaptation of instruments in psychology and education (American Educational Research Association et al., 2018).

Evidence of content validity: Evidence of content validity will be obtained through the expert judgment method, which will involve the participation of nine experts in psychology with master’s or doctoral degrees, who will review each of the items and also evaluate their clarity, consistency, relevance, and appropriateness of each item in order to evaluate the variable of sexist thoughts in adolescents in Trujillo. Items that achieve a value greater than 0.80 on Aiken’s V test will be considered adequate (American Educational Research Association et al., 2018).

Descriptive statistics of the items: To meet this objective, a pilot sample of 50 respondents was used, followed by a target sample of 200. The analysis began with all the responses provided by the participants, which were obtained through the application of the constructed instrument. These responses were extracted into Microsoft Excel, assigning them a numerical value to determine the descriptive statistics of the items by dimensions, such as frequency, mean, mode, standard deviation, skewness, and kurtosis. The corrected homogeneity index and communality were also determined in order to obtain the correlation of the items and whether they measure the dimension and variable under study. JAMOVI (V25) (American Educational Research Association et al., 2018) was used for this analysis.

Evidence of internal structure validity: This procedure was supported by exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). First, in the Exploratory Factor Analysis, the sample population was 250 adolescents, where the Kiser-Meyer Olkin (KMO) contrast discarded items < 0.80 for better consistency. Chi-square was assessed as significant when > .05, and degrees of freedom (df) were also evaluated to estimate empty spaces in comparison to the base model. In addition, the Standardized Residual Mean Square Root (SRMR) was used to analyze whether the data fit the model and to estimate the average difference. The Mean Square Root of Approximation Error (RMSEA) was also evaluated, as it attempts to predict the error of the sample if it were applied to the total population. In both elements, an acceptable fit of ≤ .08 was assessed. Likewise, the Comparative Fit Index (CFI) was used, as it evaluates the improvement of the model after the removal of unnecessary items, and the Tucker-Lewis Index was used, as it examines parsimonious dimensions, rated ≥ .90 with acceptable fit. On the other hand, in the Confirmatory Factor Analysis, a sample of 305 adolescents was used to verify the relationship with the AFE. There, items were also discarded for better consistency, and criteria were evaluated using the processes explained above. This sequence of assessment was achieved using the JAMOVI and JASP platforms.

This research took into account the principles established in the César Vallejo University Code of Ethics in Research, which promotes scientific integrity, respect for participants’ rights, and professional responsibility throughout the research process.

The principle of autonomy was guaranteed by obtaining informed consent from the parents or legal guardians of the participating adolescents, as well as from the educational institutions where they are studying, as stipulated in Article 7 of the aforementioned code.

Clear and accessible information was provided on the study’s objectives, duration, possible risks and benefits, and the option to withdraw at any time without negative consequences. In addition, the principles of beneficence and non-maleficence were respected in order to minimize any possible emotional distress during the application of the data collection instrument.

The items in the questionnaire were evaluated in advance by specialists to prevent content that could compromise the emotional stability of the participants. The principles of equity and human integrity were upheld, ensuring that all participants were treated equally, without distinction based on gender, economic or social status, or any other criterion that could violate their dignity (Article 3, sections h and i). Likewise, the privacy and confidentiality of the personal data collected were protected; the information obtained was coded and stored under the custody of the principal investigator, without linking any identifying information, in accordance with the provisions of subsection k of the same article.

On the other hand, the principles of truthfulness, honesty, and transparency were honored in the collection, analysis, and presentation of results, thus avoiding bias, manipulation, or falsification of data. The originality of the research was also ensured, preventing any form of plagiarism (Articles 10 and 11).

Finally, the ethical review process for the project was managed before the corresponding Research Ethics Committee, in compliance with the provisions of Article 4 of the Code, awaiting favorable approval for its execution. Throughout the process, the research team maintained a conduct based on professional responsibility, scientific commitment, and respect for current institutional regulations.

3. Results

Table 1 presents the content validity of items j4 to j13 of the Inventory of Prevalence of Sexist Thoughts (IPPS-25), evaluated by a group of nine expert judges using Aiken’s V coefficient, considering the criteria of clarity, coherence, and relevance. The results obtained indicate that all items achieved values above .80, which, together with confidence intervals within acceptable ranges, supports the suitability and consistency of the items evaluated. Consequently, it can be stated that the items analyzed in this dimension have adequate content validity, ensuring that they are clear, coherent, and relevant to the variable being measured.

Table 2 presents the content validity of items e1 to e14 of the Inventory of Prevalence of Sexist Thoughts (IPPS-25), which were evaluated by a group of nine expert judges using Aiken’s V coefficient. This evaluation is based on criteria of clarity, consistency, and relevance. The results obtained indicate that all items exceed values of .80, which, added to the confidence intervals within acceptable ranges, supports the suitability and consistency of the items analyzed. Therefore, it can be stated that the items in this dimension have adequate content validity, which guarantees their clarity, consistency, and relevance for the variable to be measured.

Table 3 shows the analyses relating to the items corresponding to the Hierarchies and Social Gender Identity dimension of the Sexist Thoughts Prevalence Inventory (IPPS-25). With regard to the frequency distribution, there is significant variability between the items. For example, item j7 accounts for 43.0% of responses in category 2, while item j13 shows a more balanced trend, with percentages distributed between 14.5% and 32.5% in the four response categories (Chávez-Ayala et al., 2024). The means fluctuated between 2.29 (item j7) and 2.71 (item j13), suggesting a spectrum ranging from slight disagreement to moderate agreement. Standard deviations ranged from 0.865 to 1.11, all within the recommended range (≤ 1.5), implying adequate stability in the responses obtained (Calvete et al., 2021). In relation to asymmetry (g1), values ranged between −0.259 and 0.214, while kurtosis (g2) varied between −1.34 and −0.597. These findings remain within the acceptable range of ±1.5, indicating the absence of extreme biases and adequate data distributions (Calvete et al., 2021). Regarding the corrected homogeneity index (IHC), the items presented values ranging from 0.3522 (j8) to 0.671 (j9), generally exceeding the threshold of 0.30, indicating adequate internal consistency. Similarly, the communalities (h2) ranged from 0.418 (j7) to 0.656 (j12), demonstrating an appropriate relationship between the items and the general construct of the dimension (Calvete et al., 2021).

Table 4 presents the analyses of the items related to the dimension of Affective and Behavioral Expectations toward the Other Gender from the Sexist Thoughts Prevalence Inventory (IPPS-25). Regarding the frequency distribution, there is evidence of variability in the responses. For example, item e3 shows a concentration of 43.5% in category 3, indicating a clear tendency toward moderate agreement. In contrast, item e10 exhibits a more dispersed distribution of responses, with a predominance in categories 1 (32.5%) and 2 (32.0%), suggesting higher levels of disagreement (Vargas Vargas & Vásquez Huamán, 2022). The means ranged from 2.14 (item e10) to 2.73 (item e3), showing a range from slight disagreement to moderate agreement. The standard deviation ranged from 0.88 to 1.06, which are appropriate values as they are below the threshold of 1.5, indicating stability in the responses (Cortés Rodríguez, 2023). With regard to asymmetry (g1), the values ranged from −0.352 to 0.39, while kurtosis (g2) ranged from −1.17 to −0.668. These results are in line with the suggested range of ±1.5, suggesting the absence of extreme biases and adequate distributions (Vargas Vargas & Vásquez Huamán, 2022). In relation to the corrected homogeneity index (IHC), the values vary between 0.222 (e3) and 0.5663 (e11), exceeding the established minimum threshold of 0.30 in all cases, which highlights adequate internal consistency. With regard to communalities (h2), these ranged between 0.4 (e2) and 0.764 (e3), confirming an appropriate relationship between the items and the general construct of the dimension (Cortés Rodríguez, 2023).

Table 5 presents the findings of the Exploratory Factor Analysis (EFA) carried out on the Sexist Thoughts Prevalence Inventory (IPPS-25). In an initial phase, a model composed of four factors with a total of 40 items was analyzed, which showed an adequate fit to the proposed theoretical structure, manifested in a χ² = 745.643, df = 626, p < .001. This result was complemented by indices that support the validity of the model, such as SRMR = 0.043, RMSEA = 0.031, CFI = 0.949, and TLI = 0.935. These values are within the recommended ranges, suggesting a good overall fit of the model; however, items with cross-factor loading were identified, which could complicate the clear differentiation of each of the dimensions separately (Chávez-Ayala et al., 2024). Next, a comparison was made of a two-factor model comprising 14 items, whose results also showed satisfactory indicators: χ² = 90.741, df = 64, p = .016, with values of SRMR = 0.042, RMSEA = 0.045, CFI = 0.957, and TLI = 0.939. This model, despite being more simplified, preserves adequate levels of fit, with CFI and TLI indices slightly higher than those of the original model, suggesting a simpler structure without compromising the quality of the fit (Casas-Caruajulca et al., 2021). Together, both models provide strong evidence of factorial validity; however, the two-factor model with 14 items is more efficient, improving the fit and simplifying the structure without compromising consistency.

Table 6 shows the factor loadings associated with the two factors obtained through exploratory factor analysis. Factor 1, called Gender Hierarchies and Social Identity, groups items J4, J5, J7, J12, J9, J13, and J8, with factor loadings ranging from 0.456 to 0.740. These values suggest that the items maintain an appropriate relationship with the factor, with the highest loadings corresponding to items J4 (0.740) and J5 (0.708), implying a strong association with this dimension.

Factor 2, called Affective and behavioral expectations toward the other gender, comprises items E6, E2, E11, E14, E3, E1, and E10, whose factor loadings range from 0.367 to 0.633. On this occasion, the item with the highest weight is E6, with a loading of 0.633, while the item that contributes the least is E10, with a loading of 0.367. This finding indicates that, although the relationships are moderate, the items exhibit adequate internal consistency with the dimension being evaluated.

No cross-factor loadings were identified in either factor, reinforcing the clarity in the distinction between the two dimensions and the absence of ambiguity in the connection between the items and their respective constructs.

Figure 1. Exploratory factor model of the Sexist Thoughts Prevalence Inventory (IPPS-25).

Figure 1. Exploratory factor model of the Sexist Thoughts Prevalence Inventory (IPPS-25).

Table 7. Confirmatory Factor Analysis of the Inventory of Prevalence of Sexist Thoughts (IPPS).

Model	X2	gl	CFI	TLI	SRMR	RMSEA
One-dimensional (14 items)	292.760	77	0.680	0.622	0.090	0.096
2 factors (14 items)	110.295	79	0.949	0.939	0.048	0.038

Note: X2 = Chi-square gl = degrees of freedom, p = significance, SRMR = Standardized Residual Mean Square Root, RMSEA = Root Mean Square Error of Approximation, CFI = Comparative Fit Index, and TLI = Tucker-Lewis Index.

Table 7. Confirmatory Factor Analysis of the Inventory of Prevalence of Sexist Thoughts (IPPS).

Model	X2	gl	CFI	TLI	SRMR	RMSEA
One-dimensional (14 items)	292.760	77	0.680	0.622	0.090	0.096
2 factors (14 items)	110.295	79	0.949	0.939	0.048	0.038

Note: X2 = Chi-square gl = degrees of freedom, p = significance, SRMR = Standardized Residual Mean Square Root, RMSEA = Root Mean Square Error of Approximation, CFI = Comparative Fit Index, and TLI = Tucker-Lewis Index.

Table 7 shows the confirmatory factor analysis (CFA) of the Sexist Thoughts Prevalence Inventory (IPPS-25): one with a single factor and another with two additional dimensions. The unidimensional model did not show significance; however, the two-dimensional model showed statistical significance in the chi-square test. Nevertheless, additional fit indicators favored the two-dimensional model, thus exhibiting a better fit (X2/gl < 0.06, CFI = 0.949, TLI = 0.939, SRMR = 0.048, RMSEA = 0.038). In comparison, the one-dimensional model showed a lower fit (X2/df < .001, CFI = 0.680, TLI = 0.622, SRMR = 0.090, RMSEA = 0.096). Based on the data, it is confirmed that the exploratory factor analysis model of the Sexist Thoughts Prevalence Inventory (IPPS-25) has two factors and 14 items.

a. Diagram of the factors that make up the instrument

Figure 2 shows the structure of the Inventory of Prevalence of Sexist Thoughts (IPPS-25), which consists of two dimensions: gender hierarchies and social identity, and affective and behavioral expectations toward the opposite gender. These two dimensions represent the two factors that make up the instrument.

Figure 2. Confirmatory factor model of the Sexist Thoughts Prevalence Inventory (IPPS-25).

Figure 2. Confirmatory factor model of the Sexist Thoughts Prevalence Inventory (IPPS-25).

Table 8 shows the results of the convergent validity analysis of the Sexist Thoughts Prevalence Inventory (IPPS-25), using the Adolescent Ambivalent Sexism Inventory (ASI) as an external variable. Spearman’s correlation coefficient indicated a value of ρ = 0.535 (p < .001) based on a sample of 305 participants, indicating a positive and statistically significant correlation between the two measures.

Similarly, the value of r² = 0.345 suggests that approximately 34.5% of the variance in the Sexist Thoughts Prevalence Inventory (IPPS-25) is justified by its relationship with the external scale, indicating a moderate level of association.

Overall, the results obtained support the convergent validity of the Sexist Thoughts Prevalence Inventory (IPPS-25), showing that the instrument has an adequate capacity to assess the construct of sexist thoughts prevalence, in line with other measures that have been established previously.

Table 9 shows the results relating to the internal consistency reliability of the Sexist Thoughts Prevalence Inventory (IPPS-25), calculated using the omega coefficient (ω). The dimension of Gender Hierarchies and Social Identity, composed of 7 items, achieved a value of ω = .728, while the dimension of Affective and Behavioral Expectations toward the Other Gender, also with 7 items, obtained a value of ω = .716. Both results exceed the minimum recommended criterion of 0.70, which indicates acceptable internal consistency in both dimensions.

Overall, the 14-item inventory showed a coefficient of ω = .749, a value that confirms the adequacy of the reliability of the instrument as a whole. These results support the stability and cohesion of the two dimensions, as well as the inventory as a whole, showing that the items are consistently linked when evaluating sexist thoughts.

Table 10. Preliminary statistical analysis of items from the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

	Items	Frequency				M	DE	g1	g2	h2	IHC
		1	2	3	4
Emotional and behavioral expectations toward the opposite gender	e4	15.5 %	34.5 %	38.0 %	12.0 %	2.46	0.896	-0.0421	-0.752	0.674	0.397
	e8	12.0 %	24.0 %	43.5 %	20.5 %	2.73	0.924	-0.352	-0.668	0.764	0.222
	e13	32.0 %	28.0 %	24.5 %	15.5 %	2.23	1.07	0.298	-1.17	0.794	0.208
	e16	57.0 %	25.5 %	10.0 %	7.5 %	1.68	0.934	1.24	0.498	0.747	0.258
	e19	11.0 %	31.0 %	37.0 %	21.0 %	2.68	0.928	-0.157	-0.837	0.87	0.166
	e25	32.5 %	32.0 %	24.5 %	11.0 %	2.14	0.998	0.39	-0.955	0.699	0.392
	e33	34.5 %	24.5 %	27.0 %	14.0 %	2.21	1.07	0.284	-1.22	0.546	0.457
	e38	16.5 %	31.5 %	37.0 %	14.5 %	0.5	3.28	12.4	168	0.923	0.153
	e43	34.5 %	26.5 %	28.0 %	11.0 %	2.15	1.02	0.31	-1.12	0.743	0.253
	e45	16.0 %	23.5 %	35.5 %	25.0 %	2.69	1.02	-0.281	-1.02	0.666	0.39

Note: M = mean; SD = standard deviation; g1 = Fisher’s coefficient of asymmetry; g2 = Fisher’s coefficient of kurtosis; IHC = corrected homogeneity index; h2 = communality.

Table 10. Preliminary statistical analysis of items from the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

	Items	Frequency				M	DE	g1	g2	h2	IHC
		1	2	3	4
Emotional and behavioral expectations toward the opposite gender	e4	15.5 %	34.5 %	38.0 %	12.0 %	2.46	0.896	-0.0421	-0.752	0.674	0.397
	e8	12.0 %	24.0 %	43.5 %	20.5 %	2.73	0.924	-0.352	-0.668	0.764	0.222
	e13	32.0 %	28.0 %	24.5 %	15.5 %	2.23	1.07	0.298	-1.17	0.794	0.208
	e16	57.0 %	25.5 %	10.0 %	7.5 %	1.68	0.934	1.24	0.498	0.747	0.258
	e19	11.0 %	31.0 %	37.0 %	21.0 %	2.68	0.928	-0.157	-0.837	0.87	0.166
	e25	32.5 %	32.0 %	24.5 %	11.0 %	2.14	0.998	0.39	-0.955	0.699	0.392
	e33	34.5 %	24.5 %	27.0 %	14.0 %	2.21	1.07	0.284	-1.22	0.546	0.457
	e38	16.5 %	31.5 %	37.0 %	14.5 %	0.5	3.28	12.4	168	0.923	0.153
	e43	34.5 %	26.5 %	28.0 %	11.0 %	2.15	1.02	0.31	-1.12	0.743	0.253
	e45	16.0 %	23.5 %	35.5 %	25.0 %	2.69	1.02	-0.281	-1.02	0.666	0.39

Note: M = mean; SD = standard deviation; g1 = Fisher’s coefficient of asymmetry; g2 = Fisher’s coefficient of kurtosis; IHC = corrected homogeneity index; h2 = communality.

Table 9. Reliability based on internal consistency of the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

Dimensions	Coefficient ω	items
Hierarchies and social gender identity	.728	7
Emotional and behavioral expectations toward the opposite gender	.716	7
Total	.749	14

Note: Reliability; ω: Omega coefficient.

Table 9. Reliability based on internal consistency of the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

Dimensions	Coefficient ω	items
Hierarchies and social gender identity	.728	7
Emotional and behavioral expectations toward the opposite gender	.716	7
Total	.749	14

Note: Reliability; ω: Omega coefficient.

Table 10. Reliability based on internal consistency of the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

Dimensions	Coefficient α	Items
Hierarchies and social gender identity	.727	7
Emotional and behavioral expectations toward the opposite gender	.719	7

Note: Reliability; α: Alpha coefficient.

Table 10. Reliability based on internal consistency of the Inventory of Prevalence of Sexist Thoughts (IPPS-25).

Dimensions	Coefficient α	Items
Hierarchies and social gender identity	.727	7
Emotional and behavioral expectations toward the opposite gender	.719	7

Note: Reliability; α: Alpha coefficient.

Table 11. Percentile norms for the Inventory of Prevalence of Sexist Thoughts (IPPS-25) in adolescents in the city of Trujillo.

PC	HIERARCHY	EXPECTATION	GRAND TOTAL
100	28.00	27.00	55.00
95	24.70	24.00	46.00
90	23.00	23.00	43.00
85	22.00	22.00	42.00
80	21.00	21.00	41.00
75	20.00	21.00	39.00
70	20.00	20.00	38.00
65	19.00	19.00	37.00
60	19.00	19.00	37.00
55	18.00	18.00	36.00
50	18.00	17.00	35.00
45	17.00	17.00	34.00
40	17.00	16.00	34.00
35	16.00	16.00	33.00
30	15.00	15.00	32.00
25	15.00	15.00	31.00
20	14.00	14.00	30.00
15	13.00	13.00	29.00
10	12.00	12.00	26.00
5	11.00	10.30	23.30
N	305	305	305
Media	17.57	17.50	37.07
Fashion	19	15	33
DS	4.141	4.092	6.658
Minimum	7	7	16
Maximum	28	27	55

Note: M = mean; SD = standard deviation; g1 = Fisher’s coefficient of asymmetry; g2 = Fisher’s coefficient of kurtosis; IHC = corrected homogeneity index; h2 = communality.

Table 11. Percentile norms for the Inventory of Prevalence of Sexist Thoughts (IPPS-25) in adolescents in the city of Trujillo.

PC	HIERARCHY	EXPECTATION	GRAND TOTAL
100	28.00	27.00	55.00
95	24.70	24.00	46.00
90	23.00	23.00	43.00
85	22.00	22.00	42.00
80	21.00	21.00	41.00
75	20.00	21.00	39.00
70	20.00	20.00	38.00
65	19.00	19.00	37.00
60	19.00	19.00	37.00
55	18.00	18.00	36.00
50	18.00	17.00	35.00
45	17.00	17.00	34.00
40	17.00	16.00	34.00
35	16.00	16.00	33.00
30	15.00	15.00	32.00
25	15.00	15.00	31.00
20	14.00	14.00	30.00
15	13.00	13.00	29.00
10	12.00	12.00	26.00
5	11.00	10.30	23.30
N	305	305	305
Media	17.57	17.50	37.07
Fashion	19	15	33
DS	4.141	4.092	6.658
Minimum	7	7	16
Maximum	28	27	55

Note: M = mean; SD = standard deviation; g1 = Fisher’s coefficient of asymmetry; g2 = Fisher’s coefficient of kurtosis; IHC = corrected homogeneity index; h2 = communality.

Table 10 shows the reliability indices obtained using the alpha coefficient (α). The dimension Hierarchies and social gender identity, comprising seven items, achieved a value of α = .727. On the other hand, the Affective and Behavioral Expectations Toward the Other Gender dimension, which also includes seven items, reported a coefficient of α = .719.

Both coefficients exceed the threshold of 0.70, allowing us to conclude that the two dimensions exhibit appropriate internal consistency. This suggests that the items in each factor are adequately interrelated and consistently assess the construct to which they belong.

Both coefficients exceed the benchmark of 0.70, allowing us to affirm that the two dimensions exhibit adequate internal consistency. This indicates that the items in each factor are sufficiently related and consistently measure the construct to which they belong.

4.1. Percentile Norms

Table 11 presents the percentile norms for the Sexist Thoughts Prevalence Inventory (IPPS-25) in a sample of 305 adolescents. In the Hierarchy dimension, a mean of 17.57 was obtained, with a standard deviation of 4.141, a minimum of 7, a maximum of 28, and a mode of 19. In the Expectation dimension, the mean was 17.50, with a standard deviation of 4.092, a minimum of 7, a maximum of 27, and a mode of 15. Finally, in the overall total, the mean reached 37.07, with a standard deviation of 6.658, a minimum of 16, a maximum of 55, and a mode of 33.

In relation to the percentiles, the highest score in the 100th percentile was 55 in the overall total, while the 50th percentile had a value of 35 points. At the lower end, the 5th percentile obtained a total of 23.30 points. This information facilitates the observation of the distribution of scores in each of the dimensions, as well as in the overall inventory score.

5. Discussion

The prevalence of sexist thoughts refers to the frequency with which cognitive expressions of gender discrimination manifest themselves in a population. This phenomenon may be linked to the perpetuation of notions of male superiority, which foster inequality in contemporary society. For this reason, it is essential to have an objective and validated instrument within the sociocultural context of the Peruvian population that can be used to support various research projects related to the variable in question. Consequently, a study was conducted to determine the construction and psychometric properties of the Sexist Thoughts Prevalence Inventory (IPPS-25) in adolescents in Trujillo, 2025. The results of this research provided the following data:

The first objective achieved was the development of a table of specifications, which included both the consistency matrix and the operationalization of the variable, allowing the prevalence of sexist thoughts to be established as the main variable of the study. Its dimensions and indicators were also identified. Fifty items were created, distributed across five dimensions: Beliefs about gender roles and capabilities, Power and control relationships between genders, gender group identification and favoritism, justification of sexist prevalence, and affective and behavioral expectations toward the other gender, resulting from the interaction between Ambivalent Sexism Theory (Glick & Fiske, 1996) and Social Identity Theory (Hogg, 2021). An interval scale was used as a measurement tool, complying with the statistical requirements necessary to ensure the validity and reliability of the tests, in accordance with the theoretical model of the variable in question (Rodríguez et al., 2021). Likewise, with the aim of reducing bias and the influence of social desirability in the participants, an instrument was developed with items distributed randomly and organized by dimensions. It was ensured that the items were clear, simple, and easily understandable, in addition to including an approximate number of items for each dimension in order to ensure the relevance of the items on the scale.

The second objective was to evaluate content validity using Aiken’s V coefficient based on the criteria of nine judges. Four proposed items were eliminated, and in each case aspects such as clarity, relevance, consistency, and significance were verified. The values obtained ranged from a minimum of 0.89 to a maximum of 1, indicating that acceptable and valid values were met (Chávez-Ayala et al., 2024). Consequently, our research shows that all items are within an acceptable range. Even so, one item has been modified to use more accessible language that is understandable to adolescents and more closely related to the variable that is intended to be measured theoretically according to the standards proposed by the American Psychological Association (APA) (Rodríguez et al., 2021).

Likewise, the third objective consisted of conducting a descriptive analysis of the elements of the instrument, following the information provided by the theories mentioned above. Two dimensions were established: the first referred to gender hierarchies and social identity (with items 4, 5, 7, 8, 9, 12, and 13), and the second related to affective and behavioral expectations toward the other gender (with items 1, 2, 3, 6, 10, 11, and 14). The results showed maximum frequency values of up to 43.0%, without exceeding 80%, suggesting that these findings can be considered adequate (Vargas Vargas & Vásquez Huamán, 2022). With regard to asymmetry and kurtosis, results were observed that fell within the ranges of +/- 1.5, indicating that the responses are normally distributed (Cortés Rodríguez, 2023). On the other hand, the homogeneity index exceeded the minimum threshold of 0.30 in all cases, which is considered acceptable, except for item e3 (Calvete et al., 2021). Likewise, the communality showed scores above .40, which shows a relevant relationship between the items and their respective factors (Calvete et al., 2021).

In accordance with the fourth objective, which proposes exploratory factor analysis to corroborate the coherent organization of the dimensions presented in the construct, it was found that the items measure what is expected according to the theoretical model of ambivalent sexism and social identity theory. To carry out the proposed study, the population consisted of 200 adolescents, and most of the items initially projected were deleted to maintain validity in the internal structure. According to Ondé & Alvarado (2022), the refinement of items allows for greater precision in the measurement of variables, since the variance between factors is controlled. Through this process, the internal structure was consolidated from 50 items in 5 dimensions, 10 items in each dimension, to 14 items in 2 dimensions with 7 items in each dimension. The exploratory factor analysis found an adequate overall fit with 14 items, indicating: χ² = 90.741, gl = 64, p = .016, with values of SRMR = 0.042, RMSEA = 0.045, CFI = 0.957, and TLI = 0.939. It highlights parsimony and simplicity without losing validity. With regard to factor loading, there is no evidence of cross-loading, and most items are > 0.50, showing a favorable association of dimensions. This finding contrasts with that of Galván-Cabello (2021), who conducted a similar study based on the theory of ambivalent sexism, initially proposing 20 items, all of which were retained; that is, it was not necessary to delete any items. They considered two dimensions: “benevolent sexism” and “hostile sexism.” The fit indices showed viability, reflecting: p < .001, indicating significance, KMO (.91) exceeding .80, and GFI = .99 very close to 1, showing a very good fit. Thus, it is estimated that the contrasted model has greater psychometric properties for evaluation.

Regarding confirmatory factor analysis, significant chi-square data were observed; however, the fit indices are adequate, and therefore acceptable (X2/df < 0.06, CFI = 0.949, TLI = 0.939, SRMR = 0.048, RMSEA = 0.038). In relation to this, (Sideridis & Alghamdi, 2025) comment that it is common for the chi-square test to be highly sensitive to the sample population (in this case, 350 adolescents) in large samples, which is why the model is often rejected, even though the indices show an adequate fit. In their research, the sample is large, so they consider it more appropriate to interpret the fit indices, as they provide more realistic and stable information about the model. Similarly, (Groskurth et al., 2023) indicate that goodness-of-fit indices behave differently under different situations, such as variety in the number of items, sample size, and others; but they particularly note that when the sample is very large, the models may be rejected, but this does not determine that the indices do not have an appropriate fit. Along the same lines, Goyal & Aleem (2023) have a large sample in their study, which gives it significance; however, the fit indices show a good fit (CFI close to .96-.97, similar TLI, low SRMR, and small RMSEA), so according to the indices, their model is acceptable.

Regarding the convergent validity of the Sexist Thoughts Prevalence Inventory (IPPS-25), which was externally tested using the Ambivalent Sexism Inventory (ASI) for adolescents, a positive correlation was confirmed based on Spearman’s coefficient analysis of ρ = 0.535 (p < .001), and statistically significant between both measures. In turn, the association is considered moderate with 34.5% of the variance of the Sexist Thoughts Prevalence Inventory (IPPS-25). In relation to the findings, a previous study conducted with adolescents in Spain corroborated the usefulness of the ASI test for measuring sexist behaviors and related it to romantic myths with double standards through the Romantic Love Myths Scale (EMA) test (p >.01). This result corroborates the methodological validity of the convergence between sexist attitudes and related constructs (Guerra-Marmolejo et al., 2021).

Furthermore, it is argued that the Ambivalent Sexism Inventory (ASI) can be applied to the adolescent population with a consistent and reliable structure, reinforcing the fact that it is possible to create similar constructs in young people (Juarros-Basterretxea et al., 2023).

In relation to the fifth objective, we sought to establish evidence of reliability through internal consistency of the inventory using the Omega coefficient, which yielded an overall result of .749. Therefore, reliability is adequate, and it is estimated that the dimensions measure what is expected and are consistent. Likewise, the overall Alpha coefficient is .719, reaffirming that internal consistency is favorable. As stated by (Colorado Romero et al., 2025), the interpretation of the results can be considered acceptable if the score is greater than .70 and less than .80. In contrast to the data found, in his similar research on sexism, he obtained an omega coefficient of .73, which is adequate, but it is estimated that his finding is slightly lower than that of the present research. Similarly, (Bonilla-Algovia et al., 2024), in contrast to the findings, the alpha coefficient is .92, which is very good for demonstrating reliability through internal consistency, slightly exceeding the Sexist Thoughts Prevalence Inventory (IPPS-25).

The final objective sought to establish the norms and cut-off points for the scale. The average range lies between the 25th and 75th percentiles, indicating normative or typical scores for sexist thoughts, while scores above 75 show high levels of sexist thoughts and scores below 25 show low levels of sexist thoughts. Current literature suggests that percentile approaches are useful for establishing solid ranges within the usual standardization (Urban et al., 2025). In contrast to other studies, no ranges have been established to categorize levels of sexist thoughts. However, it is estimated that a high level of sexism involves more frequent hostile behavior toward women through negative attitudes and rejection, while a low level encompasses egalitarian thoughts between men and women, even if cultural stereotypes persist (Ayala et al., 2021).

6. Conclusions

The psychometric properties of the Sexist Thoughts Prevalence Inventory (IPPS-25) were evaluated in a sample of 555 adolescents from the city of Trujillo. Solid evidence of validity and reliability was found, indicating that the instrument is suitable for the selected population. A specification table for the variable was then constructed, consisting of a 5-dimensional (later 2-dimensional) factor structure and 50 items. The content of the instrument was also evaluated by a group of 9 experts. This analysis, using Aiken’s V coefficient, in terms of clarity, coherence, and relevance, showed adequate scores V>.80 and adequate confidence intervals, considering it valid and acceptable, showing that all items were within the acceptable range. However, four items that did not correspond to this score were eliminated, leaving a total of 46 items. Likewise, a descriptive analysis of the items was established, showing appropriate adjustment indices. With regard to exploratory factor analysis, it allowed the instrument to be optimized by reducing its items from 46 to 14, distributed across two dimensions. It also showed acceptable factor loadings, which improved the internal consistency of the questionnaire. Likewise, confirmatory factor analysis revealed a better fit for the two-dimensional model, confirming the model proposed by AFE. On the other hand, it was concluded that the anger expression and control inventory shows evidence of convergent validity due to its correlation with a recognized external measure: the Ambivalent Sexism Inventory for Adolescents (ISA). Spearman’s Rho coefficient was used, given that the data did not show a normal distribution. The result shows a positive and significant correlation ρ = 0.535, (p < .001), suggesting that both instruments assess related constructs. Likewise, the value of r², which indicates the percentage of variance shared between both scales, is 0.345. This means that 35.5% of the variability in the scores on the Inventory of Prevalence of Sexist Thoughts (IPPS-25) can be explained by the scores on the ISA, which provides acceptable evidence of convergent validity, albeit with a moderate effect size. Finally, interpretation norms were created using percentiles for the inventory.

7. Recommendations

It is recommended that future research strengthen the reliability of the questionnaire, ensuring its stability and equivalence through more rigorous psychometric analyses. In this regard, it is important to involve a larger number of expert judges in the validation of the items, which will improve the clarity and relevance of the instrument. It is also recommended that the study population be expanded in order to obtain more representative and generalizable data. Similarly, it would be valuable to construct an instrument that covers a wider age range, which would facilitate the analysis of the evolution of sexist thoughts at different stages of adolescent and youth development.

Furthermore, it is necessary to strengthen the methodological component of the questionnaire, ensuring that the instrument has a representative sample and rigorous validation processes, such as factor analysis, reliability analysis, and item analysis. This will allow the questionnaire to measure sexist thoughts in a clear, accurate, and contextualized manner, also considering the gender perspective and intersectionality with variables such as age, socioeconomic status, and geographic context.

Similarly, it is recommended to broaden the scope and impact of research through longitudinal and comparative studies in different populations, with the aim of analyzing the evolution of sexist thoughts and their relationship with everyday behaviors. In this way, the results can be interpreted within a broader cultural and gender framework, guiding the design of effective interventions and programs that promote the reduction of sexist attitudes in adolescents.

Finally, it is recommended that the analysis of results be deepened by considering the cultural and social context and gender differences. This will allow for a more accurate understanding of how sexist thoughts manifest themselves in different groups. The findings should not be limited to describing their prevalence, but should guide educational and preventive actions.