Methods Flashcards
Cohort effects
when subjects of the same age/demographics impact the study due to age-related factors
Ex. The UKB! Everyone gets older, more chronic pain
Face validity
Whether a test measures what it is supposed to measure
Do experts say it measures the thing it is supposed to?
internal validity
if the test is congruent with itself (same results for person across contexts)
external validity
If the test results are generalizable to different contexts or settings or other measures
concurrent validity
extent to which the results of a test correspond to those of a previously established measurement for the same construct
do pain scales have concurrent validity with eachother?
construct validity
Does the test measure the underlying constructs/theories it seeks to measure?
Are pain scales for chronic pain actually measuring pain for pain patients?
grouping data: ordinal
classifies data into ordered classes (such as if there is a hierarchy in the data)
grouping data: interval scale
shows numerical values of distance between any two adjacent attributes
F-scale
personality test for fascism (authoritarism) in different contexts
grouping data: nominal
groups data into categories and other non-quantitative values
what type of study is best to look at differences among multiple demographics?
cross sectional study
Q-sort test:
participant sorts cards by which are characteristic for them and which are not characteristic
sorting qualities
Minnesota multi phasic personality inventory
Personality measure (originally for psychological disorders)
-550 true/false not sure questions
-cons: have not been shown to discriminate between different disorders
pros: high validity
Does NOT measure interpersonal skills
California Personality Inventory
personality test (less clinical groups than the MMPI)
Measures interpersonal skills
Myers Briggs
-Derived from carl jung’s personality theory
-Given 4 letter personality type
Julian Rotter’s internal-external locus of control scale
pretty self explanatory
curvilinear correlation:
two variables that are positively correlated up to a certain point, when they become negatively correlated
ratio variables:
Has an order, equal intervals, and an absolute zero:
examples:
enzyme activity, dose amount, reaction rate, flow rate, concentration, pulse, weight, length, temperature in Kelvin (0.0 Kelvin really does mean “no heat”), survival time
power in a statistical test
ability to find a difference in distributions when there actually is one
ordinal variable:
something that can be divided up into ranked categories but cannot be definied numerically always (like education experience, high school, college, etc)
inter-rater reliability:
Answers to a test will yield the same results irregardless of who is grading it
wilcoxon signed ranks tests
non parametric tests that compare related samples, matched samples or repeated measures on a single sample
projective personality tests
tests where people “project” parts of their psyche on objects/choices
intentionally ambiguous
affected by rater’s bias
type 1 eror
False positive: person is diagnosed with an illness although they don’t have one
type 2 error
False negative: an illness is falsely not diagnosed and you miss it
chi-squared test is used:
to describe how much difference there is between observed and expected data
One way vs. two way ANOVA
one-way: 3 or more groups differ significantly (one independent var)
two-way: means of 3 or more groups differ significantly for more than one independent var
T-test
If two groups differ significantly
Difference between bar graph and a histogram:
Histogram: bars touch, can be good to show distributions
spurious variable
spurious variable
