Lecture Notes: Research and Study Design – Science Health Stream (Lecture 1)

Introduction and context

  • Rupert Kobecki introduces himself as the Subject Coordinator for STM 1,001 and the 1st lecture for the Science Health Stream content.
  • Distinction between streams:
    • Data Science stream students should watch the Data Science lecture with Tarek (or attend that session).
    • Science Health Stream content is for Science Health Stream students.
  • Recordings and links are available in the LMS (Zoom links and recordings tile).
  • Week 1 overview focuses on introduction to research and study design, Jamovi, and readings for topic 1B to prepare for the computer lab.
  • Acknowledgement that there is a lot of information; aim is a laid-back overview to build foundational knowledge.
  • Reading content in weeks 1–4 is substantial; recommended to go through readings at your own pace on the LMS.
  • The stream material is tailored to your chosen course: core fundamental statistics components plus stream-specific applications.
  • Week-by-week timetable and activities are in the overview PDF under STM 1,001 Important Information tile.
  • Jamovi is used for learning to code and perform analyses; it is user-friendly and different from older software like SPSS.
  • The course emphasizes using software to conduct analyses and interpret results rather than performing many hand calculations.
  • Jamovi is a streamlined interface built on R; most weeks require little to no R coding, though some weeks may include light R tasks.
  • Jamovi is free and can be installed on desktops (Windows/Mac); cloud access is available if needed.
  • The lecturer demonstrates that Jamovi can produce cleaner, easier-to-interpret results compared to manual calculations (e.g., t-tests).

Jamovi: what it is, how it works, and why we use it

  • Jamovi is a user-friendly statistical software designed for statistical analyses, reducing the need to code everything by hand.
  • It is based on R but provides a more accessible interface; in most weeks you won’t need to write R code.
  • Comparison points:
    • Excel is not ideal for reproducible statistical analyses; it can have bugs and may auto-change entered data, and it doesn’t leave a clear audit trail for analyses.
    • Jamovi (and R/Python) are better for reproducible research and handling larger data sets with higher precision.
  • Why use Jamovi in STM 1,001:
    • Efficient, user-friendly, free, and suitable for the subject’s statistical needs.
    • Helps focus on interpreting results rather than manual calculations.
    • It’s designed for statistical analyses rather than general spreadsheet tasks.
  • Demonstration concept (in the session):
    • Open a data set in Jamovi, view variables, drag variables to explore descriptives, and generate a histogram.
    • You can customize visuals (e.g., color schemes) beyond black-and-white defaults.
  • Quick setup guidance:
    • Install from jamovi.org and download the desktop version (works on Macs; iPads have limited support).
    • If issues arise, use the cloud version as an alternative; desktop recommended for full functionality.
  • Example workflow in Jamovi vs manual calculation:
    • Manual t-test workflow involves calculations, table lookups, and decisions.
    • Jamovi provides a direct interface to run the test and view results, simplifying interpretation.
  • Accessibility and rollout: the lecturer notes a lot of students already installed Jamovi; labs will cover installation and use in depth.

Key concepts: research, study design, and evidence-based practice

  • Why research and statistics matter across disciplines:
    • Regardless of field, you will engage with research and may design or conduct studies.
    • Foundational knowledge helps prevent overwhelm and supports evidence-based decision making.
  • Evidence-based research:
    • Conclusions should be based on data and evidence rather than solely on gut instinct or traditional practice.
    • Real-world decisions often mix data collection, measurement, and counting with informed judgment.
  • The role of data in research:
    • Data: information from various sources, including numbers, labels, text, or even media like a TikTok video.
    • A data set is a structured collection of data prepared for analysis.
    • Excel limitations: may suffer bugs, lack of audit trails, auto-modification of data, and poor support for reproducible analyses.
    • Statistical software (Jamovi, R, SPSS, Python) supports reproducible research and handles larger data sets with greater precision.
  • The importance of starting statistics early in study design:
    • Involve statistics knowledge at the design stage to maximize the quality of results, rather than adding statistics after data collection.
    • Early stat input can influence what data to collect and how to structure measurements.
  • Real-world examples and metaphors used in the discussion:
    • Historical example: the 17th-century belief that leaving a dirty shirt out would spawn a mouse; illustrates the need for data-driven conclusions.
    • Anecdotal example: a project where late data collection caused headaches due to unclear rationale and suboptimal data.
  • Core vs stream content alignment:
    • Core statistics fundamentals are taught upfront.
    • Stream-specific content shows how statistics apply to your discipline (e.g., health science, data science).
  • The six components of research (overview):
    1) Define what you are interested in (formulate the research question).
    2) Design the study (choose design, controls, sampling, etc.).
    3) Collect the appropriate data (data collection planning and execution).
    4) Describe or summarize the data (descriptive statistics, summaries).
    5) Analyze the data (inferential statistics, modeling).
    6) Report the results (presentation, interpretation, conclusions).
  • The relationship between study design and data analysis is bidirectional and iterative; statistics should be integrated from the start.

Quantitative vs. Qualitative research: definitions, differences, and examples

  • Fundamental distinction:
    • Quantitative research: numerical data, focuses on averages, percentages, medians, and ranges; aims for generalizability to a larger population.
    • Qualitative research: descriptive data from small samples, focusing on opinions, experiences, and context; not always generalizable.
  • Mixed-method approaches: you can combine qualitative and quantitative methods to gain both breadth and depth.
  • Examples used in the session:
    • Quantitative example: wombat weights collected to compute average weight; large sample size; potential generalizability to other wombat populations.
    • Qualitative example: interviewing people about reasons for buying electric cars; small group sizes; deeper understanding of motivations, with potential segmentation by age, gender, or location.
  • Distinctions in data handling:
    • Quantitative: numerical data (weights, percentages, scores).
    • Qualitative: descriptive data (interviews, open-ended responses).
  • Practical implications for study design:
    • Which type of data you collect influences the study design, data collection methods, and analysis approach.
    • Some studies benefit from a mixed-methods approach to capture both numerical trends and contextual understanding.
  • Data quality and objectivity:
    • Quantitative analyses emphasize objectivity and generalizability.
    • Qualitative analyses emphasize depth of understanding, context, and nuance.

Data, data sets, and the workflow of analysis

  • Data as information:
    • Data include numbers, labels, text, and other forms of information; anything that can be measured or described.
  • Data sets:
    • A data set is a structured collection of data prepared for analysis, not a random assortment of sources.
  • Why we don’t rely solely on Excel:
    • Excel can introduce bugs, may not preserve an explicit analysis log, and can alter data entries automatically.
    • Statistical software enables reproducible analyses and robust handling of larger data sets.
  • Reproducible research:
    • Jamovi, R, or Python support reproducible workflows, making it easier to replicate analyses and share methods.
  • Practical use of Jamovi in the labs:
    • Load a dataset, inspect variables, perform descriptive statistics, and generate visuals like histograms.
    • Lab sessions reinforce how these steps translate into research workflows.

The t-test and interpretation basics (illustrative example from the session)

  • A quick intuition for a t-test: compare means from two groups and assess whether the observed difference is likely due to chance.
  • A standard t-test formula (illustrative): t = rac{ar{X}1 - ar{X}2}{sp \, \sqrt{\frac{1}{n1} + \frac{1}{n_2}}}
    • Here, \bar{X}1, \bar{X}2 are group means; (sp) is the pooled standard deviation; (n1, n_2) are group sample sizes.
  • In practice, Jamovi can compute the t-statistic and the p-value directly, avoiding manual table lookups.
  • Emphasis from the lecturer:
    • The goal is to learn to conduct analyses and interpret results, not to perform extensive hand calculations.
    • Some weeks may involve light R coding, but most tasks rely on Jamovi’s interface.

Readings, labs, and LMS resources

  • Readings for week 1–4 are substantial and designed to prepare you for the labs and topic 1B content.
  • The LMS hosts: overview PDFs, the timetable, and links to recordings and session tiles.
  • Kahoots (interactive quizzes) are used to check understanding and engagement during sessions:
    • Access kahoot.it with the PIN provided on screen.
    • Quizzes serve as quick checks and memory aids; they are not stressful but informative.
  • Quiz 1 is open; details are in the LMS Quiz tile. Most students will choose options based on their developing understanding of terms.
  • Communication and help channels:
    • If Jamovi installation issues persist after lab checks, email the lecturer for assistance.
    • Questions can be posted in the LMS discussion forums (Core Science Health, Data Science, etc.).
    • FAQs on the LMS provide answers to common administrative questions.
  • Labs and staff:
    • Bendigo: Tina and Tone; Aubrey Wodonga: Michael; ~35 labs with different staff.
  • This session ends with an invitation to reach out with questions and highlights about starting the semester.

Practical implications and takeaways for exam preparation

  • Core idea: integrate statistics early in research planning; design studies with statistical considerations in mind.
  • Distinguish clearly between quantitative and qualitative methods and know when mixed methods are appropriate.
  • Understand data concepts: data vs data set, types of data (ordinal vs nominal), and the issues with spreadsheet software for rigorous analysis.
  • Be comfortable with Jamovi as a primary tool for analysis and with a basic understanding of why R underpins Jamovi.
  • Know how to interpret results from basic descriptive statistics and simple inferential tests (e.g., t-test) and how to communicate findings.
  • Be prepared to discuss study design steps and provide examples from real-world research (e.g., wombat weight study, consumer behavior on electric cars).
  • Familiarize yourself with LMS resources, forum discussions, and how to access labs remotely if needed.