Qualitative and Quantitative
Research Methods in Psychology (Seminars)
Your Seminar Teachers (we)
Your Seminar Teachers (we)
- Pozdniakov Ivan (Research Assistant at CDM HSE, PhD student)
- Seminar teacher
Your Seminar Teachers (we)
- Solodkov Roman (Research Assistant at CDM HSE, Master student)
- Teaching Assistant
Plans for seminars
Plans for seminars
Osin's Lectures
- I module: Philosophy, methodology, descriptive statistics
- II module: Experimental planning with focus on non-experimental methods and quazi-experiments
- III module: Statistical tests with focus on multidimensional methods, qualitative methods
Seminars
- I module: Introduction to R, descriptive statistics and visualization in R
- II module: Experimental planning with focus on true experiments (no R!)
- III module: Introduction to inferential statistics, statistical tests in R and (optional) SPSS with focus on basic tests (t-test, ANOVA, linear regression)
So, yes, we will learn programming in R!
Don't worry!
Why R?
- Specific tasks:
- Data processing (usually some kind of tables or matrices)
- Visualization (to present your data AND to understand it better)
- Statistics
- Cognitive Neuroscience: Signal processing
- Linguistics: Text processing
- Speed (usually) doesn’t matter: very high level languages with (what?)
R
- Was created by statisticians for **statistics* (1993, but S was created in 1976)
- High-level language, created for data analysis
- Very popular in science (especially in social sciences)
- … and cognitive sciences too
- Active development
- Complex environment for creating dynamic documents, presentations, web-applications and even web-sites!
In addition…
- RMarkdomn:
- Dynamic html documents and presentations (like this one)
- Writing PDF with good LaTeX quality
- Relatively easy to convert documents to presentations and vice versa
- Dynamic html documents and presentations (like this one)
- Shiny web-applications
- Even web-sites!
Other options
- MATLAB (MATrix LABoratory) - almost nessecary for signal processing in cognitive neuroscience (+EEGLAB, SPM, Brainstorm etc.)- you have another course for that =)
- Python - general purpose language, good for language processing and machine learning
- Lower level languages (C, C++, Java) - excellent but not for data analysis
- Julia – new language that tries to be better than MATLAB, Python and R (not yet)
- SPSS (initial release – 1968!!) – very common program that used by many psychologists for statistics. It is popular because doesn’t require writing a code. Nevertheless, you can =)
- Stata, SAS – not for cognitive science
- Excel - well, why not? Limited data analysis
Comparison
| Aspects | MATLAB | R | Python |
|---|---|---|---|
| Data Analysis | + | +++ | ++ (with pandas) |
| Speed of processing | + | + | ++ |
| Text processing and parsing | - | + | +++ |
| Visualization | + | +++ (with ggplot2) | ++ (with additional libs) |
| Signal Processing | +++ | ++ (with additional libs) | ++ (with additional libs) |
| General purposes | - | - | +++ |
| Nice IDEs | +++ | ++ | + |
| Matrix and vector operations | +++ | ++ | ++ (with numpy) |
| Easy to learn | Matter | Of | Opinion |
| Different OS compatability | ++ | ++ | +++ (in Linux - by default) |
Why R?
Why R?
Why R?
Why R?
What's better?
- Cognitive psychology: R or SPSS
- Linguistics: Python or R
- Cognitive Neuroscience: MATLAB and (R or SPSS)
- Computational neuroscience: MATLAB or Python
- Computational psychology: Python or MATLAB
- Differential psychology: SPSS or R
- Advanced differential psychology (with SEM): SPSS + Amos, Mplus, EQS or… R
So, we will learn R
and some SPSS too
Our plans for this module
I Module (topic 1): R as a calculator, RStudio, variables, functions, data types
I Module (topic 1): R as a calculator, RStudio, variables, functions, data types
I Module (topic 2): Data import, data.frames, text data
I Module (topic 3): Data reshaping, descriptive stats
data.tablevsdplyrdata.tablebasics- descriptive stats
I Module (topic 4) Visualization in ggplot2 and plotly
Rules of the game
Rules of the game
- Coursework = 0.5 * H + 0.3 * T + 0.2 * S.
- Final Score = 0.6 * Coursework + 0.4 * FinalExamScore.
- The scores S, H, and T are not rounded. The total score is rounded to the nearest integer.
- H = BIG homeworks at the end of the II and III module
- T = end-of-the-module (I, II and III) tests (40 multiple choice questions)
- S = seminar activity and small homeworks
- “Automatic” pass policy:
- Option 1) Those students whose average score on the end-of-module tests (T) equals 7.5 or above, have the option of having this score counted as final exam score.
- Option 2) Those students whose Coursework score (H, T, S combined) equals 7.5 or above, have the option of having this score counted as course final score.
- No-fail exam policy: If a student who is eligible to get an “Automatic pass” (Option 1 or Option 2) chooses to take the final exam, his/her exam score will only be counted in case it makes the exam / course total score higher, compared to the “automatic pass” score.
- Tests: One low-scored test can be rewritten to improve the grade and one missed test can be written
Homeworks
- 2 BIG HOMEWORKS (H):
- Experimental planning : 3 research plans on one hypothesis – deadline is around the end of module II
- Your own data analysis of some real data (e.g., from someone’s experiment or some open dataset from Internet)
- Small homeworks after seminars – will be evaluated as a part of your seminar activity (S) as well as attendance (S).
Next classes for this module
- Will be separated by two groups before and after Osin's lecture. Please choose 1 group that fits your schedule better
- Install R and RStudio!
- Bring your own laptop with R and RStudio (but it is OK if you haven't)
- It will be (hopefully) at computer class (room 308)
Thank you for your attention!