PSY 380 Topic 8 analysis selection report example

Introduction to Probability and Statistics Grand Canyon University Free custom sample in 24 to 48h

This PSY 380 Topic 8 sample paper reasons from a question about caffeine and reaction time to a one-way repeated-measures ANOVA. It was prepared for GCU PSY 380, Introduction to Probability and Statistics, catalog code PSY-380, required in Grand Canyon University's BS in Psychology. The question is made measurable, dose is inventoried as a three-level categorical variable and reaction time as ratio-level, and 30 participants complete every dose. A between-subjects ANOVA and three paired t tests are rejected with reasons, sphericity is checked as Gravetter and Wallnau describe, and the output and design changes that would alter the choice are named.

CoursePSY 380 Introduction to Probability and Statistics
TopicTopic 8
Paper typeAnalysis selection report
LengthAbout 950 words, 2 pages
FormatAPA 7 student paper
SchoolGrand Canyon University (GCU)
ProgramBS in Psychology
UpdatedOctober 2026

Free sample paper for PSY 380 Topic 8

1

From Question to Procedure: Why a Caffeine and Reaction Time Study Calls for a Repeated-Measures ANOVA

[Author Name]

College of Humanities and Social Sciences, Grand Canyon University

PSY-380: Introduction to Probability and Statistics

Topic 8 Analysis Selection Report

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

What this page is doingThe title states the analysis derived. The study is constructed for coursework; the sources are real.
2

From Question to Procedure: Why a Caffeine and Reaction Time Study Calls for a Repeated-Measures ANOVA

The Question, Made Measurable

A composite research team wants to know whether caffeine affects how quickly people respond. As stated, the question has no measurable answer, since caffeine, affect and quickly could mean many things. Restated: does mean simple reaction time, measured in milliseconds on a computer task in which participants press a key as soon as a light appears, differ after consuming 0, 100 and 200 milligrams of caffeine? This report reasons from that question and the data to the analysis, and shows the reasoning.

Inventory of Variables

The study has two variables of interest. The independent variable is caffeine dose, with three levels: 0, 100 and 200 milligrams, delivered in identical-looking drinks so participants do not know which they received. Although dose is numerical, the study uses only three fixed levels and treats them as categories, so it is a categorical variable with three levels. The dependent variable is reaction time, the average of 40 trials per session, recorded in milliseconds. Reaction time is a ratio-level measure: it has a true zero and equal intervals, so means and differences are meaningful.

Design Features

Each of 30 participants completes all three conditions, one per day, on three separate mornings a week apart, in an order counterbalanced across participants so that each order is used equally often. The same people therefore appear in every condition. That fact is decisive. Measurements from the same person are related: someone who is fast at 0 milligrams will probably be fast at 200 as well, because people differ in baseline speed. An analysis must take that relationship into account.

Deriving the Analysis

Three facts determine the analysis. There is one categorical independent variable with three levels; the outcome is a continuous, ratio-level measure; and the same participants are measured at every level. A comparison of means across more than two conditions calls for analysis of variance. Because the conditions are within participants, the appropriate form is the one-way repeated-measures analysis of variance. That procedure separates the variability due to consistent differences between people from the variability due to the conditions, and tests the condition effect against the remaining error. Removing stable individual differences from the error term makes the test more sensitive than a between-groups test on the same data (Maxwell et al., 2018).

The Two Closest Alternatives

The first alternative is a one-way between-subjects analysis of variance, which compares means across three groups. It would be wrong here because it assumes that the three sets of scores come from different, independent people. Applied to these data, it would treat each person's three related scores as unrelated, leaving individual differences in the error term and reducing the power to detect a caffeine effect.

The second alternative is a series of paired-samples t tests, comparing 0 with 100, 0 with 200 and 100 with 200 milligrams. Each test correctly accounts for the pairing, but running three tests at α = .05 pushes the risk that one or more of the three comes out significant by chance to roughly 14 percent if no true differences exist. The repeated-measures analysis tests the overall question in one step at the chosen level, and pairwise comparisons, with a correction for multiple tests, follow only if the overall test is significant (Field, 2018).

Table 1

The Analysis and Its Closest Alternatives

OptionAccounts for repeated measures?Controls error across comparisons?Decision
One-way repeated-measures ANOVAYesYes, single overall testChosen
One-way between-subjects ANOVANoYesRejected: treats related scores as independent
Three paired t testsYesNo, about 14 percent familywise errorRejected: inflated false positives

Why Dose Is Treated as Categorical

Dose could be treated as a continuous predictor instead, using regression to estimate how much reaction time changes per milligram. With only three fixed doses, that approach assumes the relationship is a straight line, which may not hold: caffeine's effect might level off between 100 and 200 milligrams, or reverse at high doses for some people. Treating dose as three categories makes no assumption about the shape and lets the data show it. If the team later tested many doses, a regression approach would become the better choice.

Checking Assumptions

The repeated-measures analysis assumes that reaction times are approximately normally distributed within each condition and that the variances of the differences between each pair of conditions are roughly equal, a condition called sphericity. Normality can be checked with plots; reaction times are often slightly skewed, which averaging over 40 trials per session reduces. Sphericity can be tested, and if it is violated, a correction to the degrees of freedom, such as the Greenhouse-Geisser correction, is applied and reported (Gravetter & Wallnau, 2017). Counterbalancing addresses order effects, such as practice making later sessions faster.

What the Output Would Answer

The analysis would produce an F statistic for the caffeine effect, with its degrees of freedom, here 2 and 58, and a p value, answering whether mean reaction time differs across the three doses more than chance would explain. It would produce an effect size, partial eta squared, answering how much of the within-person variation in reaction time caffeine accounts for. The condition means and standard deviations would show the direction and size of differences, and corrected pairwise comparisons would show which doses differ from which.

What Would Change the Choice

One change in the design would make a different procedure correct: if 90 participants were each randomly assigned to only one dose, 30 per dose, the conditions would be independent and a one-way between-subjects analysis of variance would be appropriate. If reaction times were badly skewed and could not be transformed, a nonparametric alternative for repeated measures, the Friedman test, could replace the parametric analysis.

References

Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). SAGE.

Gravetter, F. J., & Wallnau, L. B. (2017). Statistics for the behavioral sciences (10th ed.). Cengage Learning.

Maxwell, S. E., Delaney, H. D., & Kelley, K. (2018). Designing experiments and analyzing data: A model comparison perspective (3rd ed.). Routledge.

PSY 380 Topic 8 instructions, in plain terms

Topic 8 in PSY-380 asks students to select an analysis by reasoning from the question and the data. Strong reports restate the question in measurable form, inventory each variable with how it was recorded and its level of measurement, and set out design features such as the number of conditions and whether participants appear more than once. Students should derive a single analysis from those facts with the reasoning written out, reject the two closest alternatives with reasons tied to the data, check assumptions, describe what the output would answer and name the change that would make a different procedure correct. Naming a test without the derivation earns little.

How this PSY 380 Topic 8 example is built

The finished report reasons from the question to the procedure and shows the reasoning. It restates the psychological question in a form that has a measurable answer, then inventories the variables: what each one is, how it was recorded, and at what level of measurement. Design features are set out next, including the number of conditions and whether the same participants appear more than once. From those facts a single analysis is derived, and the derivation is written out rather than announced. The two closest alternatives are considered and rejected with reasons attached to the data. Assumptions are checked against what is available. The report says what output would be produced and what each part of it would answer, then names the change that would make a different procedure correct.

Where the marks sit in the PSY 380 Topic 8 rubric

Faculty reward an analysis derived, not announced. An Excellent report makes the caffeine question measurable and inventories dose and reaction time with their measurement levels. Credit goes to deriving the repeated-measures ANOVA from three conditions and the same participants, after Maxwell, Delaney and Kelley. The rubric values rejecting a between-subjects ANOVA and three paired t tests with a 14 percent familywise error, in a table, and explaining why dose stays categorical. Strong reports check sphericity and counterbalancing, as Field and Gravetter and Wallnau describe. Top marks go to stating which result would reverse the choice.

PSY 380 Topic 8 help before you submit

The costly failure is choosing a procedure out of familiarity when the level and structure of the data call for something else, and it is detectable from the design description alone. Running a comparison of means on data recorded as ranks or categories produces output that cannot be interpreted, however clean it looks. Treating repeated measurements on the same people as independent samples is the error this topic most often catches. Reports that name an analysis without deriving it have supplied an answer with no reasoning to award. Alternatives dismissed in a phrase leave the defense untested. A report that never returns to the psychological question has selected a procedure for nothing in particular.

Get a PSY-380 Topic 8 example written to your instructions

Send the PSY-380 Topic 8 assignment instructions, your classroom rubric and the scenario, variables or data set your section provided. We write a custom example against those criteria, with the question restated, the variables and design laid out, the procedure derived and defended against alternatives, and the assumptions addressed, in 24 to 48 hours. The first one is free.

More PSY 380 papers and BS in Psychology samples

PSY 380 Topic 8 questions, answered

What decides the analysis at this level?

Four things, taken in order: what the question asks about, how many groups or conditions are involved, whether the same participants contribute to more than one of them, and the measurement level of the outcome. Working through those in writing usually leaves one procedure standing, and it also produces the justification the rubric wants rather than a bare name.

Does measurement level really rule things out?

It does, and ignoring it produces numbers with no meaning. An average of categories that happen to be coded as one through five is arithmetic performed on labels, and no amount of correct calculation makes the result interpretable. Naming the level of each variable early is the cheapest way to avoid an error the instructor can see immediately in the output.

What if an assumption is violated?

Say so, say how you know, and state the response. Options at this level usually include a procedure that makes fewer assumptions, a transformation where the course has covered one, or reporting the result with the limitation stated openly. What loses marks is running the original analysis, noticing nothing, and presenting the output as though the requirement had been satisfied.

Where can I find a free PSY 380 Topic 8 sample paper?

The selection report is printed here in full: the question made measurable, variables, design, the derivation, two alternatives, why dose is categorical, assumptions, the output and what would change the choice. Share your PSY 380 Topic 8 prompt for a free report.

Why not run three paired t tests in PSY 380 Topic 8?

Each test at .05 adds to the risk of a false positive, which reaches about 14 percent across three. The repeated-measures ANOVA tests the overall question once, with corrected follow-ups.