Factors influencing interest in STEM careers: An exploratory factor analysis

In this study, the following four steps were employed in developing the instrument on factors that influence interest in STEM careers:

Stage 1: Establishing content validity; literature and content validity by experts in STEM fields

Stage 2: Conducting a pre-test;

Stage 3: Conducting exploratory factor analysis;

Stage 4: Determining construct reliability and interpretation of total score mean.

Research Context

Data used in the validation process were collected from 354 secondary school students (14 years of age) in one of the 13 states in Malaysia. The respondents in this study represented students in three types of schools in Malaysia, namely daily schools, boarding schools, and Junior Science Colleges. For all students, it was the first time that they had seen the items. Table 1 shows the demographics of the respondents from the participating schools.

Table 1. Demographics of Respondents

Variables	Descriptions	N (Respondents)
Type of Schools	Daily school Boarding school Junior science college	101 119 134
Gender	Male Female	174 180

Instrument Development

Stage 1: Establishing Content Validity

The literature review consisted of a search for studies addressing students' interest in STEM and STEM careers, factors affecting interest in STEM careers and social cognitive career theory. The search included the use of ERIC, JSTOR, and Google Scholar and searching under the terms students' STEM interest, instruments measuring STEM, students' perception of STEM and social cognitive career theory and STEM, beginning from the year 2010 until 2016. The literature review and theoretical framework guided the development of our initial pool of survey items as well as other instruments that measure STEM courses and careers, for example from Kier et al. (2014) and Faber et al. (2013).

Expert validity was conducted, and the experts included science educators and ministry officials in the STEM fields. Two experts from the Ministry of Education provided input related to STEM education in Malaysia, namely an expert in STEM education and an expert in the area of counseling. These experts validated each item in terms of content. Based on literature review and expert validity, four main constructs were identified in this study i.e., environmental factors, STEM self-efficacy, perception of STEM careers, and interest in STEM careers. Constructs involved in this study are summarized in Table 2. The complete questionnaire is available in Appendix 1.

Table 2. Constructs and Subconstructs of the Questionnaire

No	Construct	Adaptation Sources	Subconstruct	Examples of Items
1	Environmental factors (34 items)	Constructed by the researchers Nugent et al. (2015), Kier et al. (2013), and Buday et al. (2012)	Activities in the classroom Activities outside the classroom Social influences Media influences	I learned to evaluate the results of experiments.
1	Environmental factors (34 items)			I attended STEM related carnivals. My parents encouraged me to pursue a career in STEM. I like reading books about STEM.
4	Self-efficacy (20 items)	Nugent et al. (2015), Kier et al. (2013) and Buday et al. (2012)	Science	I can obtain good grades in science subjects.
			Technology	I can use the computer properly.
			Engineering	I am sure that I can build a robot from Lego.
			Mathematics	I can solve mathematical problems properly.
5	Perception of STEM careers (14 items)	Constructed by the researchers	Prospects in STEM careers	The income of workers in STEM fields is high.
5	Perception of STEM careers (14 items)	Constructed by the researchers	Skilled needed in STEM careers	Workers in STEM fields require creative problem-solving skills.
6	Interest in STEM careers (12 items)	Adapted from Faber et al. (2013)	Physical Sciences Life Sciences	Aviation engineer, alternative energy technician, lab technician, physicist, astronomer. Pollution control analyst, environmental engineer or scientist, erosion control specialist, energy systems engineer and maintenance technician.

Stage 3: Conducting a pre-test

A pre-test was conducted in one of the neighboring schools. A total of 36 respondents (14 years of age) were involved. The main aim of the pre-test was to identify respondents' understanding of the items used in the instrument. Students were briefed on the nature of the study and how to answer the questionnaire. Students were able to understand all the items in the instrument. The time taken to answer all the items was between 15 to 20 minutes.

Stage 4: Conducting exploratory factor analysis

Exploratory factor analysis is a statistical method used to explore the dimensionality of an instrument by finding the smallest number of interpretable factors needed to explain the correlations among the set of items (McCoach, Gable & Madura, 2013). Exploratory factor analysis takes a large set of variables and looks for a way in which the data may be reduced or summarized using a smaller set of factors or components. It does this by looking for clumps or groups among the inter correlations of a set of variables (Pallant, 2011). In this study, exploratory factor analysis was performed to examine the internal structure of the set of 80 items and to validate the sub-constructs underlying the four main constructs i.e., environmental factors, self-efficacy, perception of STEM careers, and interest in STEM careers. Environmental factors consist of four sub-constructs: activities in the classroom, activities outside the classroom, social influences and media influences. The construct in this study was developed based on SCCT theory, literature review on the factors affecting interest in STEM careers and content validity by experts in STEM fields. This study initially did not extend the analysis to the level of confirmatory factor analysis as this study only aimed to explore the sub-constructs underlying the identified construct - a process of developing an instrument. However, the study has since then extended the analysis to include confirmatory factor analysis (CFA) that aimed to test the pattern of relationship among the factors and confirm the CFA model. The results of this analysis, however, is not reported in this paper.

Stage 5: Determining reliability and interpretation of total score mean

The reliability for each construct was determined based on Cronbach's alpha values. The interpretation of total score mean was also presented based on adaptation of the interpretation from Nunnally (1997).