PSYCH 306 | Research Methods in Psychology | Semester 1, 2023
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
PSYCH 306 | Research Methods in Psychology | Semester 1, 2023
Instructions for Midsemester Data Analysis Assignment
30% of course grade | Due 7th May 2023, 23:59pm
Submit by uploading on Canvas
This data analysis assignment is like the R exercises you have worked through in markdown files during labs – the dataset is just larger, there are more analyses to run, and you will need to make your own decisions about which ones to run and how to interpret them.
But you won’t need to do anything you haven’t already done in one of the labs!
Overview of steps
Step 1: Download your individual data file
In the folder “Files” > “Assignment_Datasets” on Canvas, you will find hundreds of .csv files, each containing a dataset, identified by student UPI (e.g. “jdoe123.csv”). Locate and download the one corresponding to your UPI. Triple check that you have the right UPI, because you will be marked correct only if the statistical values, plots, and conclusions in your completed assignment are correct for your assigned dataset.
Step 2: Download the Assignment Template markdown file
Download the “assignment_template_PSYCH306.Rmd” file from Canvas. You will complete the assignment entirely within this markdown file.
Step 3: Read the “Dataset Description” and “Overview of Questions and Marks” sections below
Carefully read the description of the study and variables in the dataset, and the list of questions and how many marks are allocated to each. They are both on the following pages of this document.
Step 4: Work through the questions
The “assignment_template_PSYCH306.Rmd” Markdown file contains detailed instructions, and blocks of code and text for you to complete each question. Write all the R code you use to answer the questions within code blocks in the assignment Markdown document. Save your work regularly!
Step 5: Knit and submit
When you are finished, “Knit” the completed Markdown file as a Word or PDF file and upload it on Canvas under “Assignments” > “Midsemester Data Analysis Assignment” by May 7th 23:59pm.
Dataset Description
These are data from a fictional longitudinal study looking at how literacy in childhood is affected by relationships and family environment. The study has been revisiting the same children each year for several years, so the researchers are excited to find out what it might be able to tell us about any effects the Covid-19 lockdowns may have had on literacy development.
They have hired you as a research assistant, given you their dataset, and asked you to investigate!
Below is a description of the measures involved in the study. You should consult these descriptions when you are deciding which analyses to run, when you are deciding how to label graphs, and when you are reporting and interpreting the outcomes of analyses.
Participants and study design: Participants were 270 children in New Zealand, born between 2014-2019. For the purpose of looking at the effect of Covid-19 lockdowns, we will focus on measures of their literacy taken in December 2021 (after two years of intermittent lockdowns and educational disruptions) and in December 2022 (after they had returned to relatively normal schooling for one year).
“Literacy” includes a child’s proficiency in speaking, understanding, reading, and writing their native language. In this study it was measured by a standardised test, and the score is expressed as a percentile relative to other children of the same age. A score of 50 therefore indicates that a child is at the 50th percentile of children their age, i.e. exactly average. A normally-developing seven-year old will therefore have the same percentile score as a normally-developing two-year old, even though their language abilities will of course be much higher.
Overview of Questions and Marks (total 30 marks)
Part 1: Inspecting the data
Question 1.1: Complete text block describing variables (1 mark)
Part 2: Is there an association between parental income and childhood literacy?
Question 2.1: Show a scatterplot of literacy_Dec2021 as a function of parental_income (1 mark)
Question 2.2: Complete text block stating null and alternative hypotheses (1 mark)
Question 2.3: Complete text block reporting test of the association between parental_income and literacy_Dec202 (1 mark)
Question 2.4: Complete text block predicting literacy percentile from parental income (1 mark)
Part 3: Is there an association between age of child and strain on parents’ relationship during Covid-19?
Question 3.1: Show the contingency table between age_group and parental_relationship (1 mark)
Question 3.2: Complete text block reporting test of the association between age_group and parental_relationship (1 mark)
Part 4: Has literacy improved in the year since Covid-19 lockdowns ended?
Question 4.1: Complete text block reporting tests of whether 2021 and 2022 literacy scores are different from expected (1 mark)
Question 4.2: Show plot to visually assess whether 2021 literacy scores are normally distributed (0.5 mark)
Question 4.3: Complete text block reporting test of whether 2021 literacy scores are normally distributed (0.5 mark)
Question 4.4: Show plot to visually assess whether 2022 literacy scores are normally distributed (0.5 mark)
Question 4.5: Complete text block reporting test of whether 2022 literacy scores are normally distributed (0.5 mark)
Question 4.6: Complete text block reporting test of whether 2022 literacy scores are higher than 2021 literacy scores (2 marks)
Part 5: Is there evidence that the impacts of Covid-19 on literacy differed by age?
Question 5.1: Show table of number of children in each age group (0.5 marks)
Question 5.2: Show boxplot of literacy_Dec2021 grouped by age (1 mark)
Question 5.3: Complete text block reporting test of whether literacy scores in 2021 differ by age (2 marks)
Question 5.4: Complete text block reporting tests of whether data meet the assumptions of this analysis (2 marks)
Question 5.5: Complete text block reporting follow-up tests, if appropriate (2 marks)
Part 6: Could age and attachment style have affected how children's literacy recovered after Covid-19?
Question 6.1: Show table of number of children in each combination of age_group and attachment_style (0.5 marks)
Question 6.2: Show boxplot of literacy_Dec2022 grouped by age and attachment style (1 mark)
Question 6.3: Complete text block reporting test of whether literacy scores in 2021 differ by age, attachment style, and/or the interaction of the two (3 marks)
Question 6.4: Complete text block reporting tests of whether data meet the assumptions of this analysis (2 marks)
Question 6.5: Complete text block reporting follow-up tests, where appropriate (3 marks)
Part 7: What are your conclusions about the data?
Question 7.1: Complete text block summarising what the last two analyses tell us (1 mark)
Additional information
Extensions
Extensions will only be granted in exceptional circumstances. If you need an extension to the due date (7th May, 23:59pm), fill in the form on Canvas (“Modules” > “General course resources” > “How to request an extension”) and send it, together with any documentation, to Thanos Kyritsis via Canvas mail, at least two full business days before the deadline (extension requests received after 9am Thursday 4th May may not be processed in time).
How can I get help with the assignment?
Ask questions on Piazza. Please don’t share your answers or code for assignment questions – but you are very welcome to discuss points of confusion, ask for clarification, or post and discuss code from labs.
Go to tutors’ office hours (either your tutor or another one) – see course home page on Canvas for times. Note that tutor office hours will not be held during the midsemester break (10th April–21st April), or on public holidays (7th April and 25th April).
Go to Kate Storrs’ office hours – 9– 11am Thursdays, including during midsemester break (10th April–21st April).
Form study groups. Although you have different datasets, talking with others will be a very helpful in understanding what the appropriate response to each question is.
The lab in week 7 will include time to discuss the assignment. We recommend you start working on it well before then though.
The second half of the Week 6 Lecture 2 lecture contained more information and advice on completing the assignment.
Can I work with others on the assignment?
You are encouraged to discuss the assignment and help one another understand what each question is asking for. Because you have been assigned a dataset that is specific to you, you will often not arrive at exactly the same values or results as other students analysing their datasets. The final work must be your own – i.e. you should not use shared code or text, write code or text for other students, or use human or AI services to write code or text for you.
What if I have problems creating the Word file?
The final submitted file must be either a Word or a PDF document that you have created by “Knitting” the .Rmd markdown file. Try Knitting the document often and early on, as you work on it, so that you can get help via Piazza / tutor email / in lab if you encounter problems. Submitting as a PDF is fine if you prefer (it looks nicer!), though this tends to throw errors in R unless you install further packages. Do not submit the raw .Rmd markdown file.
What if I run the right analysis but make a mistake reporting it, or vice versa?
For the questions that ask you to complete a text block reporting the result of a statistical test, you will be marked on both the values you report in text and the analysis code you ran to get those values. You will receive full marks only if you report the results correctly and they accurately reflect the output of your analysis code. If you have the correct analysis code but make a mistake in reporting the result in text, you will receive partial marks. If you report the correct result in text but you have not shown your analysis code, or have performed the wrong analysis, you will get no marks for that question. It is therefore very important to make sure all the R code underlying your analyses is in the submitted assignment file.
Do the plots and reports of statistics have to be in APA format?
Roughly, but we won’t be super strict about it. When reporting statistical results, you should follow the APA numbers & stats guidelines on which abbreviations to use (e.g. M for mean, SD for standard deviation), how many decimal places to use, etc. When in doubt about how to report a result, follow the examples that have been given in lectures . For plots, the internal components of the plot should follow APA figure guidelines (see image below). In particular, they should use a simple uncluttered design, with readable font sizes, and simple descriptive labels and titles on the x-axis and y-axis (and legend, if more than one factor is shown on the plot). You don’t need to worry about any of the surrounding figure attributes though (e.g. figure number, title, or caption).
Image from: https://apastyle.apa.org/style-grammar-guidelines/tables-figures/figures
What will the plots be marked on? Do they have to look nice? Do I have to use ggplot?
Plots will be marked on whether they display the correct data clearly (e.g. simple design, readable fonts, clear labels, upper and lower limits to the x and y axes that allow the full range of data to be seen clearly). The package “ggplot2”, which you have been using in labs, allows for greater flexibility in customising plots, and generally leads to nicer- looking results. Provided your plots follow the guidelines explained here and in the previous question though, you may use whichever plotting commands you prefer.
Do I have to use only the code I’ve been taught in labs?
No! You are welcome to use additional packages or functions that we haven’t covered in labs. For example, maybe you know other ways of creating tables, or manipulating dataframes, or improving the aesthetics of plots – feel free to use these. However, make sure that when it comes to statistical analyses, you run the tests we have talked about in lectures and labs. There are usually many different ways to statistically answer a given research question, but for the purpose of this assignment please stick to the tests we have presented in lectures and covered in labs.
Can I add more code blocks and text to the markdown file?
Try to write all your code within the empty code blocks provided, and to only edit text within the green quote blocks (i.e. beginning “>”). If you would like you add more code blocks you may, as long as the existing ones aren’t deleted.
Can I / should I add comments to my code?
Feel free to comment your code (i.e. explain what certain lines are doing, by adding text beginning with "#"). This is not essential but is good programming practice, and may help us understand your response if it differs from what we are expecting!
2023-05-12