**7.3. Formulation of indicators and development of the research tool**

All the literature review summarized in Figure 1 passed through a systematic process of analysis and classification for the construction of the research tool. The values obtained resulted from a process of content analysis which methodology will be described below. From papers, books, national and international documents, this study aimed to collect information provided in each text classifying all existing conceptions about technology as well as which are the challenges of technology in today's global scenario. It was also prioritized the provision of information that could classify the various sectors of society. Based on these categorizations, the indicators of this work were developed. All variables were grouped into categories and transformed into statements (indicators) and the end result, after a refinement based on the methodology of content analysis in [92], is presented in tables 2, 3 and 4.


Source: [10]

**Table 2.** Indicators proposed for Conceptions of Technology.



Source: [10]

44 Multivariate Analysis in Management, Engineering and the Sciences

from their respective inverse relationships.

**DIMENSIONS INDICATORS**

efficient.

public.

effects.

**Table 2.** Indicators proposed for Conceptions of Technology.

different types of tasks.

**CONCEPTIONS** 

**TECHNOLOGY** 

Source: [10]

**OF** 

attitudes in the face of a sustainable technological development.

**7.3. Formulation of indicators and development of the research tool** 

methodology of content analysis in [92], is presented in tables 2, 3 and 4.

CON 03: Technology explains the world around us.

for many, such as cell phones, stereos, computers, etc.

CON 10: Technology can destroy the planet.

CON 06: The technology does not suffer influences from society.

CON 11: Technology increases the socio-economic inequalities. CON 12: Technology threatens the privacy of individuals.

social, political, economic, environmental and cultural relationships.

CON 14: Genetic engineering can help to cure diseases.

In general, this model can be translated into the following hypothesis: the social dimension influences the conceptions of technology of the individuals within it, providing favorable

This initial model formed the basis for other five variations, two of which were obtained by exchanging the places of the constructs in the model, and the remaining three were obtained

All the literature review summarized in Figure 1 passed through a systematic process of analysis and classification for the construction of the research tool. The values obtained resulted from a process of content analysis which methodology will be described below. From papers, books, national and international documents, this study aimed to collect information provided in each text classifying all existing conceptions about technology as well as which are the challenges of technology in today's global scenario. It was also prioritized the provision of information that could classify the various sectors of society. Based on these categorizations, the indicators of this work were developed. All variables were grouped into categories and transformed into statements (indicators) and the end result, after a refinement based on the

> CON 01: Technology is the application of laws, theories and models of science. CON 02: The technology does not need theories; only needs to be practical and

CON 04: Today there are technologies that can be purchased at an affordable price

CON 07: The way we use technology is what determines whether it is good or bad. CON 8: The inventor loses control over the invention since it is available to the

CON 05: Technologies are tools (or artifacts) built to assist humans in solving

CON 9: A new technological discovery can be useful anywhere in the world.

CON 13: The benefits of technological development are greater than its negative

CON 15: Different groups of interests determine the technological production from

**Table 3.** Proposed Indicators for the Social Dimension.



Technology and Society Public Perception: A Structural Equation

Modeling Study of the Brazilian Undergraduate Students' Opinions and Attitudes from Sao Paulo State 47

Following guidelines from [90], at the end of the collect, the data recorded in the questionnaires were entered in an Excel spreadsheet to be later processed by specific statistical software's to aid in the treatment and analysis of quantitative data. The software

package that became popular in social science research, as shown in [95], and has adequate resources to the purposes of this research ([91, 95-102]). A The coding was made with the SIMPLIS command language, available in the system, which made possible the estimation of the parameters of the model through confirmatory factorial analysis, according to different estimation methods, and the verification of the respective measures of adjustment

From the individual evaluation of each construct was then possible to conduct the validation of the models of measures of each of these (DSO, ATI and CON) and this validation was performed by applying the Confirmatory Factorial Analysis (*Confirmatory Factor Analysis - CFA*). This technique has the purpose to test the hypothesis of adjustment of empirical data to a theoretical model, where a relationship structure is imposed and confirmed by analysis. Nevertheless, the variables need not to be related to all common factors. In particular, as is the case of this investigation, each variable is related to only one

The constructs presented earlier had their dimensionalities tested since this action is an premise to the reliability of the construct. The observation of the unidimensionality was made observing if each value of the normalized residue matrix of the construct was lower than 2.58, in modulus, at a level of significance of 1%, indicating if the effect on the overall adjustment of the model was low. In each process the indices of fit were checked, supplemented by information generated by the option "Modification Index" programmed in LISREL ®, which points out how much is expected to decrease the chi-square if a given a reestimation occurred, as in [98]. A detailed analysis of the standardized residuals of all dimensions was made and it was found that the overall quantity of residues which exceeds the value of 2.58 is very low and don't reaches 3% of the total. Thus, the unidimensionality

Reliability is a measure of the internal consistency of the construct indicators and of the adequacy of the scales to measure it. According to the authors, a value commonly used for acceptance of reliability is 0.70, although this is not an absolute standard, and values below

13.0 was used to verify the reliability and constructs unidimensionality, as well as the

 *8.54,* one of the most traditional statistical structural equation modeling

**8. Methodology of data analysis and results** 

**8.1. Individual evaluation of the constructs** 

**8.2. Unidimensionality of the constructs** 

of the constructs is not compromised.

**8.3. Reliability of the constructs** 

*SPSS*

system *LISREL*

of the models.

factor.

Source: [10]

**Table 4.** Proposed Indicators for Attitudes toward technological development.

#### **7.4. Sampling and data collection**

In this research we adopted the technique of cross-section as it brings the advantage of allowing the acquisition of a picture of the variables of interest at a given moment in time and to emphasize the selection of a significant and representative sample of the target population ([93, 94]).

The four institutions that represented the sampling unit were selected considering the criteria of being institutions both public and private. The selected public university, located in Campinas/SP, has students from different regions of Sao Paulo State, as well as the other three private institutions. These private institutions were one university and one faculty of Sao Paulo/SP and one faculty of Campinas/SP. The two private faculties selected receive students from different regions in the state and were also chosen because the researcher had already served for a long period in one of them and is now starting activities in the other one. The diversity of courses that the four institutions have was also a decisive factor in their choices.

The data collection in the public institution was done directly with the students, from different courses, and the questionnaires were, in the most part, passed before the beginning of the classes in the days chosen for the data acquisition. Students were selected from the following courses: Environmental Engineering, Computer Science, Nutrition, Psychology, Business Administration with emphasis in International Business, Electrical Engineering, Production Engineering, Physics, Mathematics, Technology in Environmental Management, Administration and Education.

Initially, around 1006 questionnaires were returned, yielding a proportion of almost 23 interviewed by assertion. However, LISREL software was used in a procedure that made the disposal of questionnaires that were not fully answered. Thus, the amount passed to 600 valid questionnaires, representing a proportion of nearly 14 respondents per statement, which is a significant value considering [90] as basis, and taking into account that the model is not complete and it still gave a good fit in LISREL software.
