everything i’m doing to reach an (advanced) undergraduate level of psychology knowledge
goal: score a 780 on the GRE psychology subject test.
that would put me at the 95th percentile, which seems like a reasonable target considering Clinical Psychology PhD programs accept around 5% of applicants.
16.12.2024
Intro to Stats Modules V - VI
I completed Module V last week, so starting with a review of that material.
12.12.2024
Intro to Stats Modules II - IV
II: Producing Data and Sampling
random sampling - goal is to avoid bias
some types of sampling bias: selection bias, nonreponse bias, voluntary response bias
estimate = parameter + bias + chance error
chance error can be reduced with a larger sample size
from observational studies, we can only say things about correlation, not causation
to determine causation, we need controlled experiments
III: Probability
Compliment rule: P(not A) = 1 - P(A)
Addition rule: if A and B are independent, then P(A or B) = P(A) + P(B)
Multiplication rule: if A and B are independent, then P(A and B) = P(A) * P(B)
Bayes rule:
P(A|B) = ( P(B|A) * P(A) ) / P(B)
P(B) can be expanded to P(B|A) * P(A) + P(B|not A) * P(not A)
example: we know… approximately percentage of people in the US are infected with COVID-19? P(A) the probability that someone who has covid will test positive (P(B|A)) the percentage of people who test positive (P(B))
we can determine P(A|B), ie: the probability that someone has covid given they test positive so we can figure out false positive rates
test question:
There are three boxes on the table: The first box contains 2 quarters, the second box contains 2 nickels, and the last box contains 1 quarter and 1 nickel. You choose a box at random, then you pick a coin at random from the chosen box.
If the coin you picked is a quarter, what’s the chance that the other coin in the box is also a quarter?
correct answer is given as: (1/1) * (1/3) / (1/1) * (1/3) + (1/2) * (1/3), which is just 2/3
i’m not sure how to get to this from Bayes rule… but i think that’s what they’re doing.
another way to think about it is to enumerate all the scenarios.
you chose box 1, and selected q1 you chose box 1, and selected q2 you chose box 3, and selected q
intuitively, should be 2/3
maybe i can work backwards from their given answer.
P(A|B) = ( P(B|A) * P(A) ) / P(B|A) * P(A) + P(B|not A) * P(not A)
if we frame it as… probability we’re in box 1, given we selected a quarter
A = we’re in box 1 B = we selected a quarter first
in that case, P we selected a quarter, given we’re in box 1 is 1/1 probability we’re in box 1 overall is 1/3… this is checking out so far for the denominator, first part of the sum is the numerator second part: probability we selected a quarter, given not in box 1… 1/2 (i’m a little confused about why this isn’t 1/3) probability we’re not in box 1… 1/3
this still isn’t inuitive to me - it seems like the first 1/3 should have been 1/2, because we drop the double nickel box for figuring out the other probabilities… :thinking emoji:
IV: Normal Approximation and Binomial Distribution
Bell shaped histogram = normal distribution
Empirical rule: if you have a large sample size, you can use the normal approximation to the binomial distribution ~ about 2/3 of data points will be within 1 standard deviation of the mean ~ about 95% of data points will be within 2 standard deviations of the mean ~ about 99.7% of data points will be within 3 standard deviations of the mean
We can standardize data to z-scores (unit-less)
z = (x - mean) / standard deviation
Test question:
A fair coin is tossed 400 times. Approximately what are the chances to get more than 210 tails? (Use the empirical rule and the normal approximation to the binomial distribution.)
(a) 32% (b) 16% (c) 5%
from chatgpt: For a binomial distribution, the standard deviation is calculated using the formula:
s = sqrt(n * p * (1 - p))
Where: n = Total number of trials (coin flips) p = Probability of success on a single trial (probability of getting tails) (1 - p) = Probability of failure on a single trial (probability of not getting tails, i.e., getting heads)
10 = sqrt(400 * 0.5 * (1 - 0.5))
so, 210 is one standard deviation above the mean of 200 in this case.
empirical rule tells us about 66% of the distribution is within 1 standard deviation of the mean.
so, the tail above the standard deviation should be (100 - 66) / 2 = 16%
11.12.2024
Before getting into coursework - I read through the paper Therapist Reflective Functioning, Therapist Attachment Style and Therapist Effectiveness. TLDR: therapists with high reflective functioning perform better, as measured by client outcomes.
As a later exercise - I plan to do a deep dive on that paper - read through the cited work, and make sure I have firm grasp of all the statistical concepts the researchers used.
For now though, back to my undergraduate studies.
Intro Stats Module I: Introduction and Descriptive Statistics
visualizations
bar graphs, dot plots, and pie charts are for categorical data histograms are for quantitative data box plot (box and whiskers) summarizes the distribution of a set of values (min/max, quartiles, median) scatter plot (2d) summarizes the distribution of a set of values (x and y)
numerical summaries mean and median interquartile range standard deviation
how standard deviation is calculated:
i - find the mean of the data ii - subtract the mean from each data point iii - square each of the differences (this makes them all positive) iv - find the sum of the squared differences - this is the variance v - take the square root of the variance
I took and FAILED! the test. I got a lot of questions wrong about how changing data by % or fixed amounts effects mean, median, interquartile range, and standard deviation.
To work on my intuition - I’m going to make a data set in google sheets and experiment with changes to the data.
median changes the same as % or absolute uniform change quartiles changes the same as % or absolute uniform change mean changes the same as % or absolute uniform change SD does not change with absolute change, increased proportionally with percentage change IQR does not change with absolute change, increased proportionally with percentage change
100% when I retook the test.
10.12.2024
Next step: review undergrad level statistics. It’s been over 10 years since I’ve studied it in a classroom.
I enjoyed the Intro Psychology course, so I’m going to stick with Coursera.
‘Introduction to Statistics’ from Stanford University looks good.
I’ll aim to finish it before the end of next week.
After that - I want to do another pass on the intro psych material. As a rough plan - I’ll spend a day on each of the six categories listed in the GRE practice test summary pg 3. Instead of going through another course, I think I’ll google and make my own study sheets.
After that - a break to get deeper into reading papers.
First week of January - I’ll take the GRE practice test again (I didn’t look at the answers - I just scored it blindly) to see how much I improve.
9.12.2024
Finished the Coursera course! (‘Introduction to Psychology’ from Yale University)
Coursera Module VI: Variation
Part 1 was about differences - personality and intelligence mostly
most surprising nugget from the lectures and readings was that “shared” environmental factors (ie parenting) don’t matter much
the big influences are genetics, and non-shared environmental factors (peers, life experiences)
lots of twin studies
Flynn Effect - IQ drifting upwards over the past 100 years (when not normalized to 100)
Personality mnemonic: OCEAN - openness, conscientiousness, extroversion, agreeableness, neuroticism
Part 2 was Clinical Psychology
personality disorders in the DSM: Borderline, OCD, Narcissistic, Histrionic, Antisocial, Avoidant, Dependent, Schizoid, Schizotypal, Paranoid
dissociative disorders are controversial, some doubt they exist “epidemic” of them in the 90s after cultural focus on them in the late 80s we all dissociate sometimes - daydreaming during a familiar car ride, for example maybe related to the dreaming / sleep deprivation
schizophrenia: paranoid, disorganized, catatonic
mood disorders: Depression, Bipolar Bipolar I is more extreme than Bipolar II SSRI’s for depression Lithium and careful management of sleep hygiene for bipolar
anxiety disorders: Generalized Anxiety Disorder, Panic Disorder, Social Anxiety Disorder, Obsessive-Compulsive Disorder, Post-Traumatic Stress Disorder
Module VII was happiness / positive psychology
richer countries / people are happier, to a point
basic needs matter
comparative wealth (relative to peers) matters
people have a “set point” that is basically inherent in their temperament
lots of reversion to the mean
8.12.2024
Coursera Module V - part 1: Emotions
Lectures started with the evolutionary underpinning of emotion. We have emotions because we evolved that way.
The best part of this material for me were the videos, particularly the hour long interview Jaak Panksepp
To me, his descriptions make emotions sound like internal senses.
He discovered the “Play” circuit in the brain. His way of talking about emotions made a lot of sense to me.
Emotions are pre-verbal - we layer cognition on top of them, but at they’re core we share the emotions with animals (especially mammals)
From memory, he identifies 7 emotions. 3 primitive: Anger, Fear, Seeking. 4 more advanced: Lust, Care, Grief, and Play.
Jaak didn’t mention “Seeking/Wanting” vs “Consuming/Liking”, but Kent Berridge talked about this distinction. Drug addiction is often more about “wanting” than “liking”.
Coursera Module V - part 2: Social
Ebay bias study:
Same baseball cards, presented by two different hands - black or white.
Photos with white hands got higher bids.
I took the test and got 10 / 10
7.12.2024
Coursera Module IIII - part 1: Perception and Attention
lots about the visual system, with optical illusions to illustrate
fovea vs periphery
rods - less detail, more sensitive to protons, slow to recover, no color, more in periphery cones - more detail, less sensitive to protons, faster to recover, color vision, more in fovea
emphasis on constructed nature of perception
early in the pipeline, visual system finds contrast and edges
attention is limited
different ideas about at what level filtering occurs
dichotic listening studies - two stories, one in each ear
Coursera Module IIII - part 2: Memory
encoding, storage, and retrieval
5 +/- 2 rule of working memory
expertise is more efficient packing of information into those ~5 chunks (individual chess pieces vs patterns)
visual and spatial memory as an aid (pegging, memory palace)
Took the test and got 9/10. The question I missed:
Essay questions tend to be more difficult than multiple choice because with essay questions: (a) there are more cues to stimulate memory, (b) recall is required rather than recognition, (c) there is more maintenance rehearsal
I chose (a), but should have selected (b)
6.12.2024
Coursera Module III - part 1: Developmental Psychology
Piaget’s 4 stages of development (sensorimotor, pre-operational, concrete operational, formal operational)
More modern experiments:
puppets - object permanence and primitive math in infants puppets - theory of mind - deceiving the mean puppet puppets - moral sense - who’s the good guy?
Reading: Developmental Research Design and Methods longitudinal, cross-section, sequential (mix of latter two) within individuals vs between individuals
Some good TedTalks:
Patricia Kuhl - babies hear all language sounds, build distributions for the languages they hear (its why language acquisition is easier for children)
Laura Schulz - babies draw statistical inference from sparse data (is the toy broken?)
Solomon Ash “The fundamental task of psychology is to remove the vale of self evidence from things”
Coursera Module III - part 2: Language
phonemes, morphemes, syntax
Took the test and got 9/10. The question I missed:
According to Professor Bloom’s lecture, which of the following established research finding(s) provide(s) evidence that language may not be necessary for abstract thought?
(a) individuals with certain language aphasias understand false beliefs (b) nonverbal infants understand false beliefs on nonverbal measures (c) nonverbal infants understand basic mathematical concepts, such as 1 + 1 = 2 (d) all of the above
I chose (a), because I didn’t think (b) was true, and (c) seemed not complicated enough to count as “abstract thought”.
Still not sure I agree with this one - I thought there was an example in the lectures of young children not understanding false beliefs.
Correct answers (allegedly) was (d).
5.12.2024
Coursera Module II - part 3: Skinner
Took the test and got 9/10. The question I missed:
Child is annoyed by father complaining about her not cleaning her room. She cleans her room to stop his complaining. Father used… (a) positive punishment (b) positive reinforcement (c) negative reinforcement
I chose (a) positive punishment.
From google AI: PR: add something desirable -> increase behavior NR: remove something undesirable -> increase behavior PP: add something undesirable -> decrease behavior NP: remove something desirable -> decrease behavior
reinforcements increase behavior punishments decrease behavior
positive = adding stimulus negative = removing stimulus
child increases behavior of cleaning room (it’s a reinforcement), when the result is less complaints from father (its negative)
i should have selected (c) negative reinforcement.
4.12.2024
Coursera Module II - part 2: Freud
Freud was wrong (or at least unscientific) about most of what he said. But his idea ’the unconscious is important’ was monumental.
3.12.2024
took the practice test - i got 100 / 144, which works out to a score of 630 - not even in the top 50th percentile! lots of work to do here.
started on the Coursera course ‘Introduction to Psychology’ from Yale University. completed Module I and started II (on the brain)