A text analytics evaluation of a first-year engineering project-based unit

In the undergraduate engineering program at Griffith University in Australia, the unit 1006ENG Design and Professional Skills aims to provide an introduction to engineering design and professional practice through a project-based learning (PBL) approach to problem solving. It provides students with an experience of PBL in the first-year of their programme. The unit comprises an underpinning lecture series, design work including group project activities, an individual computer-aided drawing exercise/s and an oral presentation. Griffith University employs a ‘Student Experience of Course’ (SEC) online survey as part of its student evaluation of teaching, quality improvement and staff performance management processes. As well as numerical response scale items, it includes the following two questions inviting open-ended text responses from students: i) What did you find particularly good about this course? and ii) How could this course be improved? The collection of textual data in in student surveys is commonplace, due to the rich descriptions of respondent experiences they can provide at relatively low cost. However, historically these data have been underutilised because they are time consuming to analyse manually, and there has been a lack of automated tools to exploit such data efficiently. Text analytics approaches offer analysis methods that result in visual representations of comment data that highlight key individual themes in these data and the relationships between those themes. We present a text analytics-based evaluation of the SEC open-ended comments received in the first two years of offer of the PBL unit 1006ENG. We discuss the results obtained in detail. The method developed and documented here is a practical and useful approach to analysing/visualising open-ended comment data that could be applied by others with similar comment data sets.

Effective self-management strategies for bipolar disorder: a community-engaged Delphi consensus consultation study

BACKGROUND: Self-management represents an important complement to psychosocial treatments for bipolar disorder (BD), but research is limited. Specifically, little is known about self-management approaches for elevated mood states; this study investigated self-management strategies for: (1) maintaining balance in mood, and (2) stopping progression into hypomania/mania. METHODS: To identify the common components of BD self-management, Delphi Consensus Consultation methods were combined with a Community-Based Participatory Research (CBPR) approach across five study phases: (1) Qualitative dataset content analysis; (2) Academic/grey literature reviews; (3) Content analysis; (4) Two Delphi rounds (rating strategies on a 5-point Likert scale, Very Unhelpful-Very Helpful), and; (5) Quantitative analysis and interpretation. Participants were people with BD and healthcare providers. RESULTS: Phases 1 and 2 identified 262 and 3940 candidate strategies, respectively; 3709 were discarded as duplicates/unintelligible. The remaining 493 were assessed via Delphi methods in Phase 4: 101 people with BD and 52 healthcare providers participated in Round 1; 83 of the BD panel (82%) and 43 of the healthcare provider panel (83%) participated in Round 2-exploratory factor analysis (EFA) was conducted on Round 2 results. LIMITATIONS: EFA was underpowered and sample was not ethnically diverse, limiting generalizability. DISCUSSION: High concordance was observed in ratings of strategy effectiveness between the two panels. Future research could usefully investigate the provisional discovery here of underlying factors which link individual strategies. For example, 'maintaining hope' underpinned strategies for maintaining balance, and 'decreasing use of stimulants' underpinned strategies to interrupt hypo/manic ascent. There is merit in combining CBPR and Delphi methods.

Ecosystem responses to fire : identifying cross-taxa contrasts and complementarities to inform management strategies

Changes in fire frequency, extent, and intensity mean that understanding the effects of fire on plants and animals is a primary concern for ecologists and land managers. Given the potentially conflicting fire responses of species both within and across taxonomic groups, prescribing fire regimes based on the response of one or only a few species may have negative consequences for other species. Here, we integrate data collected from a series of independent but complementary studies spanning a 75 + year chronosequence in a semi-arid shrubland ecosystem in south-western Australia to consider how fire management can best promote biodiversity both within and across taxonomic groups (plants, birds, small mammals, and reptiles). Younger fire ages (6–14 years) contained sparse shrubs, large areas of bare ground, and lacked a distinct litter layer and canopy. The oldest vegetation (60–85 years) had a distinct canopy, a well-developed litter layer and cryptogamic crust, higher variability in patch width, and more woody debris. Plant species richness and diversity decreased with time since fire, whereas bird species richness and diversity increased with time since fire, and mammal and reptile species richness and diversity showed no trend. The composition of all four taxonomic groups varied according to time since fire and the presence of 11 species was confined above or below certain fire-age thresholds. Our results support the need to maintain a mix of both younger and older fire ages across the landscape to maximise species diversity, and highlight the particular importance of older fire ages for many species. Future fire management for biodiversity conservation will benefit from identifying and reconciling cross-taxa contrasts and complementarities.