767 Lexicon Flashcards
Continuous variable
no breaks between data (time, weight, etc.)
Discrete variable
Breaks between data (children, cars, etc.)
Between subjects
independent samples design. Samples unrelated to one another (e.g. IQ of men and women of this class)
Within subjects
dependent samples / correlated samples. could be repeated measures. could also be paired data (e.g. couples, romantic, client-counselor, members on one team vs. members on the other).
Multivariate
Focus on multiple dependent variables - even if only one independent variable
Univariate
Focus on one dependent variable - singular dependent measure - simple Algebra
Data
- research results from which inferences are drawn
- Usually numerical
- Can also be newspaper and magazine articles, biographical materials, diaries, and so on. Verbal materials
Analysis
- categorizing, ordering, manipulating, and summarizing of data to obtain answers to research questions
- reduce data to intelligible and interpretable form
Interpretation
- based on analysis, makes inferences pertinent to the research, relations studied, and draws conclusions about the relations
- relations within the research study
- almost automatic with analysis
- broader meaning of research data is sought
- link to theory and other findings.
- examine congruence or lack of congruence
Statistics
- theory and method of analyzing quantitative data obtained from samples of observations in order to study and compare sources of variance of phenomena
- make decisions to accept or reject the hypothesized relations between the phenomena
- aid in drawing reliable inferences from empirical observations
Statistic
a measure calculated from a sample
Parameter
a population value
Multiple regression
analyzes the common and separate influences of two or more independent variables on a dependent variable
Canonical correlation
- logical extension of multiple regression
- adds more than one dependent variable to the multiple regression model
- handles relations between sets of independent variables and sets of dependent variables
discriminant analysis
discriminate groups from one another on the basis of sets of measures
factor analysis
- different in kind and purpose from other multivariate methods
- help researcher discover and identify the unities or dimensions, called factors, behind many measures
Path analysis
- graphic method of studying the presumed direct and indirect influences of independent variables on each other and on dependent variables
- method of portraying and testing theories
- requires researchers to make explicit the theoretical framework of research problems
Analysis of covariance structures
- the ultimate approach to the analysis of complex data structures
- the analysis of the varying together of variables that are in structure dictated by theory
- Also called causal modeling and structural equation models
Index
- an observable phenomenon that is substituted for a less-observable phenomenon (e.g. a thermometer gives readings of numbers that represent degrees of temperature)
- number that is a composite of two or more numbers (means, medians, coefficients of correlations)
MAXMINCON
- Maximize experimental relations and effects
- Minimize error variance
- Control extraneous event and variables
Theory
- An abstraction, not reality
- Best available explanation of a phenomenon at some point in time
- Not static. You can never prove a theory. Only disprove it.
- Set of propositions defined by interrelated constructs
- Nature of interrelations
- Are these relationships directional? Does one moderate another
- Explain phenomenon.
- If we can explain it, we can predict it.
- Case conceptualizations are small theories. Micro theories
Science
- Systematic, planned, controlled, empirical, critical investigation into phenomenon.
- Guided by hypothesis and theories about relation between constructs.
- Ideally, a self-correcting, evolutionary process.
Stocastic
Probabilistic
Deductive reasoning
You have known truths and laws and move from theories to data
Inductive reasoning
You do not have known truths and build theory from data
Level of discourse
A set that contains all the objects that enter discussion
Frequency distribution
- Primarily for descriptive purposes 2. Observed distributions can be compared to theoretical distributions (e.g. normal)
Profile analysis
Compare profiles of scores where all scores have been converted into same unit
Theory of errors
Given enough chance errors, they will distribute into a normal curve
Standard error of the mean
Standard deviation of an infinite number of means. Measure of chance or error in its effect on one measure of central tendency
Sub-set
A set that results from selecting sets from an original set (e.g., samples of a population)
Universal set (U)
- Set of all elements under discussion 2. Universe of discourse or level of discourse. 3. Population
Empty set (E)
The set with no numbers in it. Null set.
Cross-partition
New partitioning that arises from successively partitioning the same set U by forming all subsets of the form A ^ B
Regression line
Expresses the relation between X and Y including its direction and magnitude
Relation
A set of ordered pairs
Ordered pair
Two objects, or a set of two elements, in which there is a fixed order for the objects to appear
Function
A special kind of relation in which each element of the domain is paired with only one member of the range
Cartesian product
When each individual member of A is paired with each individual member of B and all possible pairs between the two sets are obtained
Nominal
Name only - categories of an underlying construct
Ordinal
Continuum of an underlying construct with categories built upon the continuum – ranks
Interval
Categories are equidistant (equal interval) along the continuum of the underlying construct
Ratio
There is an absolute, meaningful zero (complete absence of construct)
Mode
Category that has highest frequency. Can be bi-modal or multimodal
Median
Point at which half of scores are above and half are below (50 percentile)
Percentile
Percentage of scores that fall below target score (involves ranking)
Mean
Average Sum X/N
Range
Difference between high and low score (high - low)
Variance
Dispersion of data from the mean – sums of squares / n (or n-1)
Standard deviation
Square root of variance. Take square root to put into original metric.
Deviation scores
Scores represented as deviations from their mean
Standard scores
Deviation scores divided by standard deviation
Kurtosis
Peakedness or flatness of the distribution. 4th moment about the mean
General Linear Model
Examines associations between variables. Can have multiple IVs and multiple DVs. Everything we do will fit into the GLM.
Sums of squares
Sum of squared dispersions of the mean
Cross-product of deviation scores
Allows us to compare variables. Another type of squared deviation score.
Sums of cross-products
Tell us how variables relate – gives direction of association
Covariance
SCP/N
Pearson product moment
Correlation coefficient (r)
Effect size
r squared (r^2)
Sums of Squares Total (SST) is equivalent to
Sums of squares Y (SSY)
Error =
1- r^2
r^2 =
SSM/SST
SSY =
SSM + SSE
Randomization
Random assignment
Formula of the General Linear Model
y = a + bx + e
Our model in the GLM
a + bx
In GLM, a =
Intercept - what is predicted value of y when x = 0?
In GLM, b =
Slope. Rise/run. How much change in x based on change in y?
In GLM, e =
Error. What we can’t explain.
In GLM, a + bx =
Our model. What we can explain
y hat y^
Predictive value of y
Formula for predictive value of y
y-hat = a + bx
y =
y-hat + e
e =
y - y-hat
Criterion of least squares (error)
Minimize SSe - Sums of Squares Error - only one possible line will give us this – our optimal line for our data
Probability
Chance of something being yes vs. no
b =
SCP/SSx
Conditional probability
Given x (a certain condition), what’s the probability of y?
Target population
Who you want to sample
De facto population
Who sample really represents
Z significance level
1.96z = 95%
Poisson curve
Non-normal curve
Probability density function
By formula, we know the exact shape of a distribution and know exact shape of curve
Point prediction
b = some specific value that is not 0 (or between two values)
Standard error
Estimate of chance fluctuation - measure against which outcomes of experiments are checked. Is the difference a real difference or merely a consequence of chance?
Probability sample
Use some form of random sampling in one or more of its stages
Nonprobability sample
Does not use random sampling.
Quota sampling
Knowledge of the strata of the population– sex, race, religion, etc.– is used to select sample members that are representative, “typical” and suitable for certain research purposes
Strata
Partitioning of the universe or population into two or more nonoverlapping (mutually exclusive groups)
Purposive sampling
Characterized by use of judgment and a deliberate effort to obtain representative samples by including presumably typical areas or groups in the sample
Accidental sampling
Taking available samples at hand
Stratified sampling
Population is first divided into strata and then random samples are drawn from each strata. Capitalizes on between-strata differences.
Cluster sampling
Partitioning the population into clusters then sampling the clusters randomly (streets, schools, etc.) Can have multi steps of cluster sampling
Randomness
We cannot predict outcomes.