The cross-cultural measurement of teacher motivation in developing countries

• In Event: Border crossings and hybridity in education for sustainable development

Proposal

In the pursuit of better outcomes for students (LMICs), effective teachers matter and especially in low- and middle-income countries. The measurement of teacher motivation and related issues has been minimal in developing countries. In the discourse around teacher development, the role of teacher motivation is increasingly being recognized as an important construct in understanding teacher effectiveness (World Bank, 2018). As Watt and colleagues (2017) point out:
“Teacher motivation is increasingly being foregrounded as an incredibly important dimension in what sustains teachers in working in social and educational contexts that do not necessarily support teachers in their endeavours, to engage students in learning in environments characterised by constrained autonomy and increasing levels of accountability.” (Watt et al., 2017; p. 3)
Extrinsic factors of motivation play a part, and teachers can be demotivated due to low pay, inadequate working conditions, amongst other infrastructural issues (Bashir et al., 2018). Equally valuable, intrinsic components, such as self-efficacy (Bandura and Schunk, 1981) and self-determination (Deci and Ryan, 1985) matter, as do teachers’ goals for their present and future (Miller and Brickman, 2004). The confluence of these factors determines the teacher’s motivation in that particular context. The measurement of this complex intersectionality of the sub-constructs of motivation dictated by the context is an overarching goal for this study.
Comparability across different countries, contexts, and groups of importance (such as gender, socio-economic status) allows for a deeper look at the many components at play in each context that make it different or similar to other context. In response, there has been an increased interest in both theoretical and psychometric development of measures to ensure equivalence across contexts. The domain of measurement invariance is experiencing both growth and recognition for its crucial role in cross-context comparisons. This study presents the theoretical and psychometric properties of a teacher motivation instrument in four different, LMIC, contexts.

This study will employ a three-step approach to furthering our understanding teacher motivation across different contexts. First, it uses a narrowed a list of constructs that encompass both state and trait teaching- related motivation have already been generated from a broader review of the literature on teacher motivation. This is to emphasize both school and classroom level impacts on a teacher’s motivation. As a result, sub-scales for emotional exhaustion, self-efficacy, collegial support, and growth mindset, in the context of being a teacher in a low-resource and low-support environment are included in motivation measure.
The sample sizes for analyses were 164, 455, 539 and 671 for Uganda, Uttar Pradesh, Karnataka and Delhi, respectively. It is evident that Uganda and India are different countries, on different continents and were thus very different. Within India, the three sites were deemed diverse contexts, even though Uttar Pradesh, Karnataka and Delhi are states of India. This is because the samples from these states are from rural remote, peri-urban areas and urban areas in Uttar Pradesh, Karnataka and Delhi, respectively. Further, in cultural contextualization of the measure to these contexts, the difference in language and culture, as indicated by field colleagues and back translation efforts provided further substantiation of the differences between the three contexts within India.
Next, using data collected from these 4 different cultural contexts, the study will investigate the underlying factor structure of the measure as well as the measurement equivalence. At noted earlier, little is known about the interplay of the forces that impact teacher motivation. As a first step to understanding how teacher motivation manifests across these contexts, we employ exploratory factor analyses to examine the underlying factor structure of the motivation measure in each context.
Preliminary results of the Scree plots in the exploratory factor analyses indicate that 2-5 factor structures should be evaluated for model fit and factor loadings. Within these models, the 4-factor model indicated the best fit (Hu & Bentler, 1999) across all four contexts.
Table 1.
Comparison of the four-factor model fit across the different contexts
Contexts
Model Fit Indices Chi Square Value Df RMSEA CFI TLI SRMR
Uganda 442.523* 321 0.05 0.88 0.83 0.07
Karnataka 596.730* 321 0.04 0.93 0.90 0.04
Delhi 729.532* 321 0.05 0.94 0.92 0.04
Uttar Pradesh 1224.529* 321 0.065* 0.95 0.93 0.04
*Note *p < .05; **p < .01; ***p < .001
Interestingly, the four-factor structure also resembles the hypothesized motivation sub-constructs of emotional exhaustion, self-efficacy, collegial support, and growth mindset. The emergent framework for motivation will be confirmed through confirmatory factor analysis, when feasible within, and across contexts (Marsh et al., 2014).
Finally, measurement invariance (MI) analyses will be conducted to ascertain the degree to which this motivation measure can support cross-context comparison. The study will employ multi-group, partial and full measurement invariance testing (Millsap, 2012) to distinguish and evaluate both full measure and sub-scale MI. This is to provide a holistic MI picture of the measure, as a whole or the included sub-scales individually.
In conclusion, the paper will provide a discussion on the importance of establishing psychometric properties of cross-contextually utilized measures that are both complex and multidimensional in cross comparison studies. It will provide a perspective on the utility of such measures in informing policy and research, especially when measurement invariance may be fully or partially met (Davidov et al., 2014) across different contexts.

Authors

• Mahjabeen Raza, New York University

• Maidah Mir, New York University

• Sharon Kim, New York University

• Edward Seidman, New York University