PYB350 Advanced Statistical Analysis
The Complete Exam Bible · S2 2026PYB350 Overview
- Queensland University of Technology
- Semester 2 2026
- Statistics
- SPSS and JASP
- Final exam in Central Examination Period
What PYB350 is
PYB350 Advanced Statistical Analysis belongs to the linked series of research-methods units in QUT psychology, and it carries the PYB210 groundwork in analysis of variance, regression and qualitative methods on to more complex designs.
- Most of the grade is the exam The final exam carries 60% and the written assignment 40% of the unit.
- Output, not recall Short answers give a research scenario with statistical output and ask you to answer its research questions.
- Interaction first In every factorial question, read the interaction before the main effects; it decides which follow-up is legitimate.
- Your own data The assignment data set is generated for each student, so the analysis and the data file must be your own.
How PYB350 is assessed
| Component | Weight | Format |
|---|---|---|
| Assessment 1: Written assignment | 40% | Results section and preliminary discussion for an unfinished manuscript, using your own data; abstract 100 to 150 words; 1,800 words; due Week 10; submitted through Turnitin |
| Assessment 2: Final exam | 60% | Central Examination Period; multiple choice and short answer questions on a research scenario with statistical output; tests all work covered in lectures and tutorials |
The two weights sum to 100%. The written assignment is marked out of 100 and converted to 40% of the final grade; the exam carries the other 60%. The unit materials available here do not state a hurdle or minimum mark for either component, nor the exam's date, duration or permitted materials, so confirm those in the unit outline on QUT Virtual and on Canvas.
Current PYB350 dates
| Date | Item | Control |
|---|---|---|
| Fri 2 Oct 2026, 23:59 | Assessment 1: Written assignment due | Week 10; submit the report, the output appendix and the data file on Canvas |
Dates are as published in the current course materials. Confirm exact deadlines and submission settings in the live LMS.
What PYB350 covers
Nine chapters follow the unit from the Week 1 refresher through qualitative methods, follow-up tests and factorial designs to multiple regression and the written results section.
T-Tests, Effect Size and Bootstrapping
Week 1 refresher: single-sample and independent t-tests, Cohen's d, the Central Limit Theorem and bootstrapped intervalsThe unit opens with a Week 1 refresher because every later analysis reuses its logic. This chapter rebuilds the single-sample and independent-groups t-tests, the standard error, Cohen's d and the conventions for calling an effect small, medium or large.
- Notation for scores, group means, the grand mean, n and N
- The single-sample t-test when the population SD is unknown
- Cohen's d and the small, medium and large conventions
- The Central Limit Theorem and the normality of the sampling distribution
- Bootstrapped standard errors and confidence intervals
- Independent groups, pooled variance and Levene's test
One-Way ANOVA and Partitioning Variance
Between and within variance, the F ratio, degrees of freedom, the general linear model and eta-squaredOne-way analysis of variance compares any number of group means by splitting the total variability in scores into a between-groups part, which carries the effect of the independent variable, and a within-groups part, which carries sampling error.
- Why ANOVA is needed for more than two groups
- Between-groups and within-groups variance
- The general linear model: grand mean plus effect plus error
- Degrees of freedom for total, between and within
- The source table and the F ratio
- Eta-squared and the APA reporting line
Checking ANOVA Assumptions
Independence, normality and homogeneity of variance; Levene's test, Welch, Games-Howell, transformations and outliersEvery between-groups F test rests on three assumptions: independent observations, a normal sampling distribution within each cell, and similar variances across cells.
- Independence of observations as a design feature
- Normality of the sampling distribution in each cell
- Skewness and kurtosis z-scores against 1.96 and 3.29
- Homogeneity of variance and the four Levene rows
- Welch's F, Games-Howell and Dunnett T3
- Transformations, outliers and bootstrapped follow-ups
Qualitative Interviews and Thematic Analysis
Interview formats, protocols, question and probe types, transcription, and reporting themes with quotesThe unit's two qualitative weeks prepare you for the thematic analysis inside the written assignment.
- Structured, semi-structured and unstructured interviews
- When interviews suit a research question and when they do not
- Building and piloting an interview protocol
- Primary questions, follow-up questions and probes
- Interview sequence, rapport, reflection and transcription
- Key themes, sub-themes and quotes in a results section
Planned Contrasts, Post Hoc Tests and Type I Error
Per-comparison and experiment-wise error, Bonferroni, contrast weights, Tukey, Games-Howell and Dunnett T3A significant omnibus F says a difference exists somewhere; follow-up tests find it. This chapter separates planned contrasts, chosen before the data to test predictions, from post hoc tests, which compare every pair when there are no hypotheses.
- Planned contrasts against post hoc comparisons
- Per-comparison, experiment-wise and family-wise error
- Bonferroni and the a minus one rule
- Writing contrast weights that sum to zero
- Contrast values, t tests and the SPSS LSD option
- Choosing a post hoc test by variance and group size
Factorial ANOVA: Main Effects and Interactions
Fully crossed designs, marginal and cell means, sums of squares for A, B and the interaction, and the pooled error termA factorial design crosses two or more independent variables so every combination forms a cell.
- Fully crossed factors and counting cells
- Marginal means and main effects
- Interactions and non-parallel lines
- Sums of squares for A, B and A by B
- One error term for three F tests
- Setting up the data file with a participant ID
Simple Effects and Follow-Up Tests
Following up main effects, simple effects with the pooled error, simple comparisons, omega-squared and the decision treeOnce the omnibus two-way ANOVA is run, the interaction decides what happens next.
- Interpreting main effects when there is no interaction
- Contrasts and Tukey tests on marginal means
- Omega-squared beside eta-squared
- Simple effects with the pooled error term
- Simple comparisons within a significant simple effect
- Type I error control and bootstrapped follow-ups
Correlation and Multiple Regression
Pearson's r, r squared, the least-squares line, residuals, R squared, B and beta weights, sr squared and assumptionsRegression handles continuous predictors. This chapter starts with Pearson's r, the coefficient of determination and the unexplained variance, then builds the bivariate least-squares equation, residuals and the standard error of estimate.
- Categorical and continuous variables
- Pearson's r, r squared and unexplained variance
- The least-squares line, residuals and standard error of estimate
- The model F test, R squared and adjusted R squared
- B, beta and squared semi-partial correlations
- Residual plots and the limits of causal claims
Reporting Results in APA Style
The written assignment brief, its marking criteria, results-writing habits and the preliminary discussionThe written assignment carries 40% of the unit, and this chapter turns its brief and marking criteria into a working plan.
- What the unfinished manuscript gives you and what you add
- Word limit, penalty, extension and the three submissions
- The four marking criteria and their weights
- Results-writing habits: justify, back up, give units, give direction
- A model omnibus paragraph for a two-way design
- The abstract and the preliminary discussion
The unit describes these skills as core to psychology: essential tools for researchers, integral to the scientist-practitioner model of practice, and what makes you a critical consumer of published research.
It is both theoretical and practical, so you learn the principle behind each analysis, run it in SPSS or JASP, and interpret and write up the output.
What the weeks cover
Week 1 is a refresher of t-tests, bootstrapping and one-way ANOVA. Weeks 2 and 3 introduce qualitative research methods, including thematic analysis and interviewing.
Week 5 revisits ANOVA with planned and post hoc comparisons, and Weeks 6 and 7 build between-groups factorial designs, main effects, interactions, simple effects and bootstrapped follow-up tests.
Week 8 turns to repeated measures and mixed designs, Week 9 to multiple regression, and Week 12 to ANCOVA and blocking designs, with revision in Week 13. Lectures carry all the key content, and tutorials, run online and face to face in the computer labs, give hands-on practice with each technique.
How it is assessed
There are two summative pieces.
The written assignment, worth 40%, hands you an unfinished manuscript and your own data set; you enter the data, run the analyses and write the abstract, the results, including a short thematic analysis, and a preliminary discussion.
The final exam, worth 60%, sits in the Central Examination Period and combines multiple-choice and short-answer questions, with the short answers built on a research scenario and its statistical output. This guide follows the same path, with every worked example on new numbers so you can rebuild each calculation by hand.
Worked example · free
Reading a two-way interaction from cell means
- 2Marginal means for difficulty: easy (8.0 + 7.6) / 2 = 7.8 and hard (6.8 + 4.0) / 2 = 5.4. For feedback: immediate (8.0 + 6.8) / 2 = 7.4 and delayed (7.6 + 4.0) / 2 = 5.8. The grand mean is 6.6.
- 1Effect of delay at each level of difficulty: easy 8.0 minus 7.6 = 0.4 points; hard 6.8 minus 4.0 = 2.8 points.
- 2Interaction contrast with weights 1, minus 1, minus 1, 1: 8.0 minus 7.6 minus 6.8 plus 4.0 = minus 2.4. The non-zero value says the effect of delay is not the same at both difficulty levels.
- 1Decision: if the interaction test is significant, the main effect of feedback (7.4 against 5.8) averages over two very different effects, so it is reported but not interpreted alone; simple effects of feedback at each difficulty level come next.
Key terms
- Analysis of variance
- A test that compares group means by setting the variance between groups against the variance within groups in an F ratio.
- Grand mean
- The mean of every score in a study, ignoring group membership, which every effect in an ANOVA is measured from.
- Eta-squared
- The share of total variability explained by an effect, computed as the effect's sum of squares divided by the total sum of squares.
- Omega-squared
- An effect size that corrects the tendency of eta-squared to over-estimate the effect in the population, so it is usually slightly smaller.
- Homogeneity of variance
- The assumption that the spread of scores is similar in every group, needed because ANOVA pools the group variances into one error term.
- Levene's test
- A test of homogeneity of variance whose non-significant result supports pooling; the row based on the mean is the one the unit reports.
- Bootstrapping
- A resampling method that draws many samples with replacement from your own data to estimate a standard error or confidence interval without assuming normality.
- Planned contrast
- A comparison among group means, defined by weights that sum to zero, chosen before the data are seen to test a specific prediction.
- Bonferroni adjustment
- A correction that divides the .05 error rate among the comparisons, so each of three planned tests is judged at .0167.
- Marginal mean
- The average score at one level of a factor, collapsed over the levels of the other factor, used to test a main effect.
- Interaction effect
- A pattern in a factorial design where the influence of one variable changes from one level of the other variable to the next.
- Simple effect
- The effect of one factor at a single level of the other factor, tested on cell means with the pooled within-cells error term.
- Semi-partial correlation
- The correlation between the outcome and the part of one predictor not shared with the others; squared, it gives that predictor's unique variance.
PYB350 FAQ
How is this unit assessed?
There are two summative pieces. The written assignment is worth 40%: you complete an unfinished manuscript by analysing your own data and writing the abstract, results and a preliminary discussion. The final exam is worth 60% and sits in the Central Examination Period. The assignment is marked out of 100 and scaled to 40, and both pieces are submitted or sat individually.
What does the final exam look like?
The unit says the exam combines multiple-choice and short-answer questions and tests all work covered in lectures and tutorials. In the short answers you are given a research scenario with its statistical output and must respond to the research questions it poses. A practice exam is released in the overview and exam preparation folder at the end of semester, and the date and conditions are confirmed on Canvas.
What do I hand in for the written assignment?
Three things through separate Canvas links: your report section, meaning the abstract, the results and the preliminary discussion; an appendix with your key SPSS or JASP output and a qualitative summary table; and your data file with value labels and a participant ID column. The limit is 1,800 words, up to 10% over is accepted, and the automatic 48-hour extension applies.
Which statistics software does the unit use?
SPSS is used from the first tutorial, through both the One-Way and the General Linear Model procedures, and JASP is introduced for factorial follow-up tests because it makes simple effects and custom contrasts easier to request. JASP is free and installs on your own computer, while SPSS runs through the QUT virtual service, which needs VPN access arranged a few days ahead.
What topics does the unit cover?
The unit moves from a refresher of t-tests, bootstrapping and one-way ANOVA to qualitative interviewing and thematic analysis, then planned and post hoc comparisons, between-groups factorial designs with main effects, interactions and simple effects, repeated measures and mixed designs, multiple regression, and ANCOVA with blocking designs. Each analysis is taught with its assumptions and its write-up.
Why does the interaction come before the main effects?
A significant interaction means the effect of one factor changes across the levels of the other, so a main effect averages over effects that differ and can mislead. The tutorials therefore check the interaction first. If it is significant, main effects are reported but not followed up, and simple effects of the focal factor are tested on cell means using the pooled error term.
Is advanced statistical analysis hard?
The unit builds directly on PYB210 and asks you to revise that material where topics depend on it, so gaps from earlier units show up quickly. The exam holds 60% of the grade and is built on reading real output and choosing the right next test, which takes practice rather than memorising. Working the tutorial answer sheets and the end-of-semester practice exam is the most direct preparation.
Questions students ask in each PYB350 chapter
- Why does the normality assumption not apply to the raw scores?
- The assumption concerns the sampling distribution of the mean, the distribution you would get from many samples of the same size. The raw scores are only evidence about it: if your sample is roughly symmetrical and reasonably large, about thirty scores, the sampling distribution can be treated as normal even if individual scores are not perfectly bell shaped. Chapter 01 · T-Tests, Effect Size and Bootstrapping
- What does a significant F actually tell you?
- It tells you that the group means are not all equal in the population, nothing more. It does not identify which groups differ or in which direction, so a significant one-way ANOVA is always followed by planned contrasts or post hoc comparisons. The F test is the gate; the follow-up tests answer the research question. Chapter 02 · One-Way ANOVA and Partitioning Variance
- How large do cells need to be for ANOVA to be robust?
- The factorial lecture lists three conditions under which departures from normality matter little: roughly equal cell sizes, at least ten people per cell, and departures from normality that run in the same direction in every condition. Outliers are the bigger risk, because one extreme score can inflate the variance and distort the F ratio. Chapter 03 · Checking ANOVA Assumptions
- What is the main strength of a qualitative interview?
- Access to rich and nuanced data. The lecture's own revision question makes the point: interviews are not chosen because they are cheap, quick or generalisable, but because they capture lived experience, subtlety and sensitive topics that a questionnaire cannot reach. If the research aim is to generalise to a population, another method is usually better. Chapter 04 · Qualitative Interviews and Thematic Analysis
- Do planned contrasts need a significant omnibus F?
- No. The lecture allows planned comparisons whether or not the overall F is significant, because they test predictions made before the data were seen. Post hoc tests are different: they explore all pairs after a significant omnibus result, which is why they need stronger control of the Type I error rate. Chapter 05 · Planned Contrasts, Post Hoc Tests and Type I Error
- How many cells does a 3 by 4 design have?
- Twelve. A two-way A by B design has a times b cells, so a 2 by 2 has four and a 3 by 4 has twelve, while adding a third two-level factor to a 2 by 2 gives a 2 by 2 by 2 design with eight cells. The cell count also sets the error degrees of freedom, N minus ab. Chapter 06 · Factorial ANOVA: Main Effects and Interactions
- Do I follow up a significant main effect when the interaction is significant?
- No. The tutorial reports the main effects but does not follow them up, because a significant interaction means each main effect averages over effects that differ across the levels of the other factor. The follow-up work moves to simple effects of the focal factor, tested on cell means with the pooled error term. Chapter 07 · Simple Effects and Follow-Up Tests
- What should be reported first in a multiple regression?
- The test of the whole model: the F test of R squared. If it is not significant, the predictors are not a viable set for predicting the outcome and nothing else is interpreted. Once it is significant, report R squared and adjusted R squared, then each predictor's B with its confidence interval, its beta weight and its unique contribution. Chapter 08 · Correlation and Multiple Regression
- What happens if the report goes over the word limit?
- Up to 10% over the 1,800 words is accepted. Beyond that, 10% of the total available mark is deducted, so a report over 1,980 words loses a tenth of the marks before anything else is judged. There is no penalty for being under the limit, but the brief asks you to check that every part of the task is covered. Chapter 09 · Reporting Results in APA Style
How to study for the exam
Work in the order the analyses depend on each other. Start by rebuilding the one-way source table by hand until you can produce any missing cell, because the factorial and regression tables reuse the same logic of sums of squares, degrees of freedom and mean squares.
Then learn the follow-up decision tree as a sequence of yes or no questions: is the interaction significant, is the effect significant, does it have more than two levels, did the researchers have hypotheses. Most short-answer questions are a walk down that tree with a piece of output in your hand.
Keep a one-page sheet of the conventions you will quote, from effect-size benchmarks to Bonferroni alphas and the Levene row to read. For the qualitative weeks, practise turning a handful of comments into one theme with sub-themes and quotes. Finally, sit the tutorial exercises with the answer sheets closed, and save the practice exam for a timed attempt once every chapter is done.
Your AI Statistics tutor for PYB350
Stuck on a hard PYB350 question? Sia is AskSia’s AI Statistics tutor — ask any PYB350 Advanced Statistical Analysis question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.
