1/570
Looks like no tags are added yet.
Speed test reliability = best assessed by
Test-retest
Reliability coefficient of .80 means
80% true score
20% measurement error
Cronbach’s alpha
Internal consistency for items with multiple response options.
Standard Error of Measurement (SEM)
Confidence interval around an individual's observed test score → estimate TRUE SCORE
O standard error of measurement ajuda a construir uma faixa provável para o true score.
If SD is 10 and reliability is not given, what range can SEM have?
0-10
Reliability can range com 0-1
Reliability =1 → SEM = 0
Reliability = 0 → SEM = SD = 10
Main problem with percent agreement for inter-reater reliability?
Overestimate reliability because it's affected by chance agreement
Difficulty index (p)
Proportion of people who got the item correctly
High p = EASY (.85 = 85% got it right)
Low p = HARD (.15 = 15% got it right)
Main effects
Effects of each IV, without considering the effects of other IVs in the study
Multitrait-multimethod matrix evaluates [ ] validity
Construct Validity (Convergent + Discriminant validity)
“Oblique” in factor analysis means
Correlated
“Orthogonal” in factor analysis means
Uncorrelated
Key standard score equivalents
Z = 0 → T= 50 → 50th percentile
Z = +1 → T=60 → 84th percentile
Z= +2 → T = 70 → 98th percentile
Mean em T e Z score
T → M = 50, SD = 10
Z → M = 0, SD = 1
A type of counterbalanced design
Latin Square
What threatens a study’s internal validity when an external event occurs during the course of the study and has an effect on the subjects’ scores or status?
History
Multiple regression
2 or more predictor to predict one continuous criterion
Ex:
Predictors: cognitive ability, conscientiousness, years of experience
Criterion: job performance
Simultaneous multiple regression
All predictors are entered at the same time
Multiple baseline research design
introduce treatment at different times across behaviors, settings, or people
treatment does not need to be withdrawn
Example: One patient has 3 phobias → start treatment for each phobia sequentially → multiple baseline across behaviors.
Reversal design (ABAB)
treatment is withdrawn and reintroduced.
* A = baseline
* B = treatment
* A = withdraw treatment
* B = reintroduce treatmen
Linear regression analysis
Uses one predictor to predict one continuous outcome.
Example: SAT score → college GPA.
Discriminant function analys
Uses one or more predictors to categorize people into one of two or more groups / categories
Example: predictors → depressed vs. non-depressed.
A battery of tests to determine which vocational training is more adequate for unemployed young adults
Stepwise multiple regression
Uses multiple predictors and identifies which predictors add the most predictive value.
Useful for finding the fewest predictors needed for accurate prediction.
Factor analysis
reduces many variables into a smaller number of underlying factors.
Exemplo: você tem 12 itens num questionário:
* “I worry a lot”
* “I feel tense”
* “I have trouble relaxing”
* “I feel sad”
* “I have little energy”
* “I have lost interest in things”
Turn into:
Factor 1 = Anxiety
Factor 2 = Depression
Coefficient of determination
A measure of the amount of variability in one variable that is accounted for the variability in another variable.
Obtained squaring the correlation coefficient.
Spearman rho correlation coefficient
V1 = ranked/ordinal
V2 = ranked / ordinal
ranks - 1st, 2nd, 3rd
ordinal - low/medium/high, mild/moderate/severe, promoted/not promoted
Ex:
class rank x supervisor rank
job satisfaction level (low, medium, high) x promotion status (promoted vs. not promoted)
Pearson correlation coefficient
V1 = continuous
V2 = continuous
Linear relationship
Ex:
hours studied x EPPP score
Weight x height
Biserial correlation coefficient
V1 = continuous variable
V2 = artificially dichotomized variable.
Example:
pass/fail created by cutting a continuous score
EPPP score x high/low anxiety
Point-Biserial correlation coefficient
V1 = continuous variable
V2 = true dichotomous variable.
Example:
EPPP score x male/female
yes/no
correct/incorrect.
Contingency coefficient
V1= nominal / categorical
V2 = nominal / categorical
Ex:
gender and diagnosis
Diagnosis and therapy type
What increases the standard error of the mean?
Smaller sample size
Higher population SD
What decreases the standard error of mean?
Larger sample size
Lower population SD
Type I error
False positive = null was true, but was rejected
Rejecting the null hypothesis = ver efeito onde não tinha
Type II error
False negative
Retaining a false null hypothesis
Null was false, but was maintained
Terapia funciona, mas vc conclui que não
Alpha
Probability of Type I error
0.5 = I’m accepting up to 5% of Type I error chance
0.1 = I accept 1% chance of Type I error
Beta
Probability of Type II error
Statistical power
Probability of rejecting a false null hypothesis
Chance de detectar um efeito real
Higher statistical power reduces Type [?] error
Type II
What increases statistical power?
Higher alpha
BUT: more Type I risk
T-test for correlated samples
Measures whether 2 related mean scores are significantly different.
Ex.: pre-test vs post-test in the same 35 psychologists, before/after attending a lecture.
When do you use a single-sample t-test?
To compare one sample mean with a known or hypothesized population mean.
ANCOVA is used to…
ANOVA + statistically CONTROL for a covariate.
Covariate = variable you want to statistically control because it's related to the DV, but is not the main IV you wish to study.
Spearman-Brown formula
Predicts how reliability changes when you change test length.
How can test reliability be increased?
Increase the number of items
Use a more heterogeneous sample
Reliability measures / coefficients
Internal consistency → Cronbach's alpha, KR-20, split-half
Inter-rater → Cohen's Kappa, ICC
Stability over time → Test-retest
Equivalent forms → alternate/parallel forms
KR-20
Internal consistency for dichotomous items (0/1, correct/incorrect).
Split-half reliability
Correlates scores from two halves of the same test.
Internal consistency.
Inter-rater reliability
Do different raters agree?
Categorical → Kappa
Continuous → ICC
Cohen's Kappa
Inter-rater agreement for categorical/nominal data.
Corrects for chance
ICC
Inter-rater reliability for continuous/numerical ratings.
Test-retest reliability
Same test + same people + two times → stability over time.
Alternate/parallel forms reliability
Two equivalent forms of a test → consistency across forms.
Item discrimination index (D) function and range
How well a test discriminate between low scorers and high scorers
-1.0 to +1.0
Concurrent validity
How well a new predictor estimates current scores on a criterion.
Ex: Determining how well a new screening test estimates current scores in a validated measure of depression
Correlation coefficient of .60 indicates ____% of variability explained.
36%
Square the correlation coefficient: .60² = .36 = 36%
Item characteristic curve
Probability of a correct answer for examinees at different ability levels.
Most effective way of controlling extraneous variables
Random assignment of subjects to the different treatment groups
Test specificity = ?
Ability to identify true negatives
What do you use to compare an obtained sample mean to a known population mean?
T-test for single sample
Manova
Multiple DVs
2+ dependent variables
Usually 1+ categorical IVs
Compares groups on multiple outcomes simultaneously
Example: Therapy type → depression + anxiety scores
Two-way ANOVA
2 independent variables
1 continuous dependent variable
Tests 2 main effects + interaction
Example: Therapy type × Gender → depression score
Test sensitivity = ?
Test’s ability to accurately identify people who have the disease or other attribute the test was designed to identify
Which formula is used to estimate the effect of increasing a predictor’s reliability on its criterion-related validity coefficient?
Correction for attenuation formula
Mixed research design
Within-subjects AND between-subjects
What increases statistical power
INCREASING alpha
INCREASING effect size
Parametric test
Item Characteristic Curve (ICC) - point at which the curve intercepts the y-axis indicates…
Chance of chosing the correct answer by guessing
Item Characteristic Curve (ICC) - Slope of the curve indicates
Discrimination (capacidade de diferenciar bem pessoas com alta ou baixa habilidade)
Item Characteristic Curve (ICC) - Position of the curve indicates
Difficulty of the item
Group design: time-series design is a group version of which single-subject design?
AB
The DV is measured multiple times before and after the IV/treatment is introduced.
SEE - Standard Error of Estimate
Used to construct an interval around the predicted criterion score
When measuring the relationship between two variables, a restriction in range of scores on the variables will most likely produce a correlation coefficient that:
Underestimates the actual relationship between the variables
Central Limit Theorem: as sample size increases, sampling distribution of means approach the shape of…
Normal distribution, regardless of the shape of the population distribution
What triggers lapses following abstinence period?
HIgh-risk situations
DSM diagnosis: 12 yo boy with 2 motor tics for 9 months
Provisional tic disorder (less than one year, less than 18 yo)
Disinhibited social engagement disorder: diagnostic criteria
History of extreme insufficient care
Developmental age of 9 months
Moffit: life-course antisocial behaviour is due to
Neurological deficits
Opioid withdrawal symptoms
Flu-like + agitation/anxiety, aches, diarrhea, vomiting, sweating, dilated pupils.
Stimulant withdrawal symptoms
Fatigue, low energy, vivid and unpleasant dreams, insomnia or hypersomnia, increased appetite, psychomotor agitation or retardation
Alcohol withdrawal symptoms
Seizures, delirium tremens. Tremor, autonomic hyperactivity, anxiety, insomnia, hallucinations,
Most dangerous
OCD: low levels of [neurotransmitter]
Serotonin
Alzheimer's vs Lewy Bodies - differences
Early Alzheimer’s: prominent memory impairment
Early Lewy Body: fluctuating cognition, visual hallucinations, and/or parkinsonism; memory may be less prominent initially.
Intermittent explosive disorder age
At least 6 years old
Pervasive social inhibition, feelings of inadequacy, hypersensitivity to negative evaluation. Diagnosis?
Avoidant personality disorder
Treatments for binge-eating disorder?
CBT + interpersonal psychotherapy
Gene that is the link between standard trisomy 21 and early-onset Alzheimer's?
APP
Lewhinson's theory of depression
Low rate of contingent reinforcement
Drugs used to treat premature ejaculation?
Drugs that increase serotonin
SSRIs (paroxetine, dapoxetine) and antidepressants with similar effects
Most effective intervention for childhood-onset fluency disorder
Habit reversal training
First-line treatment for bulimia
CBT
Fairchild study with male adolescentes with conduct disorder
Hyporeactive in cortisol and cardiovascular responses
OCD - brain activity?
Elevated activitity in the caudate nucleus
Genuine memory loss
Beginning and end are gradual and blurred
Patient can recall fragments of events during that period
PTSD brain abnormalities
Decreased activity: ventromedial prefrontal cortex
Increased activity: amygdala
Specific phobia - brain activity
Increased activity: amygdala, insula, thalamus, dorsal anterior cingulate cortex, medial prefrontal cortex, orbitofrontal cortex
Decreased activity: ventromedial prefrontal cortex and ventral anterior cingulate cortex
Intellectual developmental disorder: deficits in intellectual functioning with scores of _____ or more standard deviations below the mean in standardized intelligence tests
2
Concordance rate for schizophrenia for non-twin siblings
9%
Erotomanic delusional disorder
Central theme: another person [often a higher status person] is in love with the individual
Bipolar disorder heritability rates
90%
Bipolar II
At least one hypomanic episode
At least one major depressive episode
Bipolar I
Requires a full manic episode
Does not require depressive episode
The link between chronic stress and depression is mediated by the
Hyperreactivity of the HPA axis