Science teachers' collaborative innovative activities: the role of professional development and professional experience

The study addresses the challenge of engaging students in science learning and the increasing emphasis on innovation in science education. It examines science teachers’ innovative work behaviour—specifically collaborative activities of trying out new ideas and sharing new ideas—framed by Rogers’ diffusion of innovation (RDI) theory and sociocultural perspectives on collaboration. Using TIMSS 2015 teacher data from Japan, Lithuania, and South Africa (countries differing in science achievement and teacher qualifications), the research investigates how teaching experience and professional development (PD)—both duration and content—predict these collaborative innovative activities. The authors posit three hypotheses: (H1) beginner teachers are more curious to try new ideas, while more experienced teachers contribute by sharing experiences; (H2) longer duration of formal PD is positively associated with innovative activities; and (H3) PD focusing on pedagogical content knowledge predicts stronger collaborative innovative activities than PD focused on disciplinary content knowledge.

Theoretical framing draws on innovative work behaviour literature that delineates stages from idea generation to implementation and sharing (Janssen; Scott & Bruce; Rogers). RDI theory describes five stages (knowledge, persuasion, decision, implementation, confirmation), with collaborative activities especially pertinent in implementation and confirmation. Sociocultural and activity theory perspectives emphasise innovation as a collective process, countering overly individualistic approaches to teacher innovation. Prior research identifies demographic, individual, and organizational factors influencing teacher innovation, including PD and collaboration. Empirical findings on experience are mixed: some suggest positive effects on collaboration and competence (e.g., Ertesvåg; Kini & Podolsky), others indicate declining innovativeness with increasing experience (Thurlings et al.; Loogma et al.). PD generally supports teaching effectiveness, with some literature highlighting the importance of sustained duration; however, little research has directly examined how specific PD content areas relate to teachers’ innovative activities. These gaps motivate the hypotheses focusing on experience, PD duration, and PD content (with an emphasis on pedagogical content knowledge).

Design: Secondary analysis of TIMSS 2015 teacher questionnaire data for Grade 8 science from Japan, Lithuania, and South Africa. Countries were selected to represent high (Japan), average (Lithuania), and low (South Africa) levels of student science achievement. Participants: Japan N=170; Lithuania N=965; South Africa N=331. Teachers’ degree attainment differed across countries (nearly all Japanese and Lithuanian had bachelor’s or higher; many South African teachers did not). Measures: - Teaching experience: Total years teaching (continuous). - PD participation: Duration in the past two years (categories: None; <6 h; 6–15 h; 16–35 h; >35 h). For modeling, recoded into short PD (≤15 h) vs long PD (≥16 h). - PD content (binary yes/no for each): science content; science pedagogy/instruction; science curriculum; integrating ICT; improving students’ critical thinking or enquiry skills; science assessment; addressing individual students’ needs. (The questionnaire also listed integrating science with other subjects, but the analysis focused on the seven listed contents.) - Innovative activities (dependent variables): Frequency of (1) working together to try out new ideas (TRY) and (2) sharing what I have learned about my teaching experiences (SHARE), each on an ordinal scale (very often, often, sometimes, never/almost never). To address small cell counts, TRY was recoded by combining very often + often; SHARE by combining sometimes + never/almost never for OLR. Analysis: - Ordinal logistic regression (cumulative logit with proportional odds) separately by country for TRY and SHARE: TRY = f(DPD, CPD, PE); SHARE = f(DPD, CPD, PE), where DPD=PD duration category, CPD=PD content factors, PE=years of teaching (covariate). Proportional odds (parallel lines) assumption tested and met (p>0.05) for all models. Model fit assessed via likelihood ratio tests (final vs intercept-only) and deviance goodness-of-fit; Pearson GOF used where sparse cells complicated interpretation. Multicollinearity checked via multiple linear regressions; VIFs 1.018–2.300 indicated low collinearity. Descriptive statistics summarized means of teaching experience and distributions of PD participation and innovative activity frequencies across countries.

Descriptive highlights: - Mean years of teaching: Japan 17.38 (SD 12.09), Lithuania 24.54 (SD 10.83), South Africa 14.86 (SD 10.06). - PD duration: Longer PD participation more common in Lithuania and South Africa than Japan (e.g., >35 h over past 2 years: Japan 7.0%, Lithuania 28.0%, South Africa 26.0%). - Frequency of TRY (Very often/Often): Japan 24.7%; Lithuania 32.4%; South Africa 55.6%. Frequency of SHARE (Very often/Often): Japan 49.7%; Lithuania 54.9%; South Africa 68.3%. Modeling results (ordinal logistic regressions by country): - Japan: Teaching experience significantly predicted higher frequency of both TRY and SHARE. TRY: OR per year = 1.036 (95% CI 1.009–1.064), χ²(1)=6.647, p=0.010. SHARE: OR per year = 1.038 (95% CI 1.010–1.067), χ²(1)=6.881, p=0.009. No PD content or duration effects were significant. - Lithuania: PD content predicted both TRY and SHARE; teaching experience and PD duration were not significant. TRY: science curriculum PD OR=1.501 (95% CI 1.119–2.014), χ²(1)=7.361, p=0.007; critical thinking/enquiry PD OR=1.578 (95% CI 1.190–2.092), χ²(1)=10.051, p=0.002. SHARE: science curriculum PD OR=1.493 (95% CI 1.136–1.962), χ²(1)=8.287, p=0.004; critical thinking/enquiry PD OR=1.404 (95% CI 1.078–1.827), χ²(1)=6.356, p=0.012. - South Africa: Addressing individual students’ needs PD predicted both TRY and SHARE. TRY: OR=2.216 (95% CI 1.173–4.189), χ²(1)=6.002, p=0.014. SHARE: OR=2.107 (95% CI 1.162–3.822), χ²(1)=6.024, p=0.014. PD duration did not predict TRY and was negatively related to SHARE frequency in additional analyses noted in the discussion. Across countries, PD duration generally did not predict innovative activities (contrary to H2). Pedagogical content knowledge–oriented PD content (curriculum, critical thinking/enquiry, addressing individual needs) showed positive associations in Lithuania and South Africa (partially supporting H3). Teaching experience predicted innovative collaboration only in Japan (not supporting H1 overall).

The findings address the research question by demonstrating that predictors of science teachers’ collaborative innovative activities vary by country and dimension of innovation. In Japan, increasing teaching experience is associated with greater engagement in trying and sharing new ideas, suggesting experience may facilitate collaboration and innovation within that context. In contrast, in Lithuania and South Africa, specific PD content areas oriented toward pedagogical content knowledge (curriculum, critical thinking/enquiry, addressing individual needs) are more salient predictors than experience or duration. This underscores the importance of PD content over mere time invested in PD. The non-significant or negative role of PD duration suggests that seat time alone does not translate into collaborative innovation; how learning is structured and what it targets matters more. The results are consistent with literature emphasizing pedagogical content knowledge as a key lever for effective instruction and align with sociocultural perspectives that emphasize collaborative processes. Variability across countries points to contextual, cultural, and systemic differences influencing how experience and PD translate into innovative behaviours, and cautions against one-size-fits-all PD or policy prescriptions.

The study contributes by comparing predictors of two collaborative innovative activities (trying new ideas and sharing experiences) among science teachers across three contrasting national contexts using TIMSS 2015 data. It shows that PD content focused on pedagogical content knowledge is a stronger and more consistent predictor than PD duration, and that teaching experience predicts innovative collaboration only in Japan. Educationally, the results suggest prioritizing PD that develops pedagogical content knowledge (e.g., curriculum design, fostering critical thinking/enquiry, addressing individual student needs) and structuring PD in ways that support ongoing, collaborative learning rather than emphasizing hours alone. Future research should: (1) develop and use more comprehensive, reliable measures of teacher innovative behaviour; (2) examine how different PD forms and sustained learning trajectories relate to innovation; (3) explore cultural and systemic moderators across broader sets of countries; and (4) investigate heterogeneous professional learning trajectories and how they connect to innovative practices.

- Measurement limitations: Innovative behaviour was captured by two single ordinal self-report items embedded in broader collaboration questions; no established reliability/validity available. Items reflect only part of the innovation construct. - Self-report and potential cultural response biases in interpreting items and rating scales. - Country selection limits generalizability; three countries do not represent all contexts at similar achievement levels. - Observational, correlational design precludes causal inference. - PD duration measurement may not capture sustained, embedded learning beyond seat time; form and quality of PD not specified. - For modeling, PD duration was dichotomized (≤15 h vs ≥16 h), which may obscure nonlinear or threshold effects; sparse cells required category collapsing for outcomes. - Potential differences in PD availability/choice across countries may confound associations between PD content and innovation.