Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
116 Cards in this Set
- Front
- Back
AB Design
|
A single subject research design involving a single baseline (A) phase and a single treatment (B) phase
|
|
Analysis of Covariance (ANCOVA)
|
Variation of ANOVA that allows for stat control of extra variables
Ex. In study effects of various types of reading programs in kids, effect of IQ can be stat controlled by ANCOVA |
|
Alpha
|
Prob of rejecting null when its true - Type I
Set by researcher in advance .01 or .05 |
|
Analysis of Variance (ANOVA)
|
Stat test used to asses the diff bw two means
|
|
Alternative Hypothesis
|
Stat hyp that states the oppos of null, predicts that there is a relat bw two variables
|
|
Beta
|
Prob of retain null when its false, Type II
|
|
Blocking
|
Method of controlling extra variable by making it another IV to stat analyze its effect on the DV
|
|
Case Study
|
In depth study of single indiv or unit
|
|
Central Limit Theorum
|
As sample size increases, the shape of sampling distrib of means approaches normality and its means approaches the pop mean
|
|
Chi-square test
|
Infer stat test used when data freq of obs are w/in categories. Test the hyp that obs freq are equal those that would be expected if the null were true
|
|
Coefficient of Determination
|
Square of correl coeff, proportion of variability bw 2 variables share
|
|
Cohort effects
|
effects of being part of a group that was born at a particular time and exposed to unique educational, cultural, and other experiences, confound results of cross sectional
|
|
Confound
|
Extra variable, not of interest, exerts systematic effect on dv
|
|
Correlation coefficient
|
Stat value, bw -1 and +1. Indicates magnitude and direction bw 2 or more variables. Can be pos or neg. Pos same direction. Neg in oppos directions
|
|
Correlational research
|
Variables meas only not manip to determine relationship
|
|
Counterbalancing
|
Technique designed to control order effects in a repeated measure design , admin tx to diff subj in a diff order
|
|
Criterion variable
|
outcome variable in a correl research that one tries to predict
|
|
Critical Value
|
Number above or below which a stat is signif (F or t)
|
|
Cross-sectional study
|
Assesses effects of development by examining indiv of diff ages at one point in time
|
|
Cross-sequential study
|
Assesses the effects of develop by combining cross sec and longit method, examines indiv of diff ages at diff points in time
|
|
Degrees of freedom
|
Number of values that are free to vary in a stat calculation.
|
|
Demand characteristics
|
cues in a research study that give subj clue about how they are expected to behave
|
|
Dependent variable
|
The variable that is hyp to be affected by the IV , meas by the investigator after IV is manip
|
|
Descriptive statistics
|
stat used to describe and summarize at set of data. includes freq distrib, measures of central tendency, and meas of variability
|
|
Discriminant function analysis
|
Stat tech used to find orderly patterns of responses on tests, which distinguish groups of people based on these responses, practical use is to use a series of predictors to predict examinees criterion group membership
|
|
Double blind design
|
Experimental design in which neither the subj or the experimenters are aware of the hyp being tested or which subj are in the control or exp group
|
|
Experimenter expectency effect (Rosenthal Effect)
|
Effect that an experimenter unintentionally exerts on a research study, usually in a direction supporting the research hyp
|
|
Experiment wise error rate
|
The probability of making at least one type I error when mult stat comparisons are made in the course of one experiment
|
|
External validity
|
Degree to which the results of an experiment are generalizable to pop beyond the exp group
|
|
factorial ANOVA
|
2 or more IV and all possible combo of IV are administered. Assess main and interaction effects
|
|
Frequency distribution
|
How often each score in a distrib of scores occurs
|
|
Hawthorne effect
|
Being in an experiment affects subj responses to the IV, due to the attention subj are receiving as research participants
|
|
Heteroscedasticity
|
In a scatterplot unequal variability of Y scores at diff values of X
|
|
History
|
An external event, irrelevant to the IV, systematic effect on DV, threat to internal validity
|
|
Homoscedasticity
|
In a scatterplot, equal variability of Y on all values of X
|
|
Independent variable
|
Variables that is posited as the cause of the effect, manipulated
|
|
Inferential statistics
|
Procedures that allow us to infer the values of pop from samples from that pop
|
|
Internal validity
|
degree to which obs changes in a DV can be attrib to variations of IV
|
|
Interval scale
|
Numbers arranged in order and intervals are equal, no absolute zero point, mult and division are not possible
|
|
Latin square design
|
Type of counterbalanced design in which treatments are presented only once in each position (1st, 2nd, 3rd) of the sequence in which they are admin
|
|
least squares criterion
|
the principle underlying the construction of the regression line
|
|
Longitudinal studies
|
conducted over extended period of time, using the same group of subj, to assess changes over time
|
|
Main effect
|
the affect of one IV acting alone, w/o consideration of the effects of other IV
|
|
Mann-Whitney U Test
|
Nonparametric equivalent of the t-test for independent samples, used when a study involves 2 indep groups and the data are rank ordered
|
|
matching
|
creating pairs or groups of subj on the basis of their similarity on an extra variable and then randomly assigning members of each group to the tx group, designed to increase internal validity by ensuring equivalence bw groups in regards to status on extra variable
|
|
Maturation
|
naturally occurring process w/in research subj that threatens a study's internal validity by exterting a systematic effect on the DV
|
|
Mean
|
Arithmetic average
|
|
Mean square
|
In ANOVA, index of variability that is used to derive F ration. Equal to sum of squares divided by the degrees of freedom. A one-way ANOVA includes MSB and MSW
|
|
Median
|
Middle score in a distrib of scores that have been ordered from low to high q
|
|
Meta-analysis
|
Method of analyzing a group of indep studies with a common conceptual basis, using each study as if it were a separate score, resulting in an effect size indicating the average effect of a tx across many studies
|
|
Mode
|
Most frequently occurring score in a group of scores
|
|
Multicollinearity
|
Describes a situation in which predictors in a multiple regression equation overlap (highly correl). Means that predictor are providing redundant info about the criterion
|
|
Multiple baseline design
|
Single subj - in which an iv is sequentially admin across 2 or more subj, bx, or settings (across baselines)
|
|
multiple regression
|
method of combining test scores (predictor) involving the use of an equation that yields a predicted criterion score for an examinee based on his perf on predictors
|
|
Multivariate analysis of variance (MANOVA)
|
More than one IV is analyzed at same time
|
|
negatively skewed distribution
|
more scores fall at high end
|
|
nominal scale
|
classifies data into unordered categories
|
|
nonparametric tests
|
inferential stat tests used w/ nominally and ordered data, don't make parametric assumptions (normal distrib, homogeneity of variance) and therefore less powerful
|
|
Normal Distribution
|
Symmetric distrib in which half fall above and below, when plotted appears bell shaped
|
|
Null hypothesis
|
Stat hyp that there is no diff bw or among tx/control groups. Experimenter attempts to disprove.
|
|
One-way ANOVA
|
A stat hyp test used when a research study has one IV and involves comparing means obtained from more than 2 groups, assesses whether means are equal
|
|
One-tailed test
|
Alt hyp is directional (one pop mean is greater than the other)
|
|
Order effects (carry over, mult tx interference)
|
repeated meas, effects of order of tx admin
|
|
Ordinal scale
|
Scale, ranks, arranges data in order categories
|
|
Parametric tests
|
Inferential stat tests used when data meas on a interval or ration scale, assumptions normal distrib, homogeneity of variance
|
|
Path analysis
|
Stat tech used to test hyp regarding causal relat among mult variables via correl methods (w/o manip)
|
|
Pearson r
|
Correl coeffe used to meas the relat bw 2 intervally or ratio scaled variables
|
|
Percentile ranks
|
Score expressed as a percentage of cases in group lying below the given score value
|
|
Positively skewed distrib
|
Scores fall a low end
|
|
Post-hoc test
|
After ANOVA - Scheffe and Tukey, identify precise pattern of diff among groups
|
|
Power
|
Prob of rejecting a false null, ability of stat test to correctly id tx effect
|
|
Power
|
Prob of rejecting a false null, ability of stat test to correctly id tx effect
|
|
Pretest sensitization
|
Alters subj reactions to IV, threatens external validity, can't be generalized to situations when pretest not admin
|
|
Pretest sensitization
|
Alters subj reactions to IV, threatens external validity, can't be generalized to situations when pretest not admin
|
|
Quasi-experiment
|
Subj are not randomly assigned to groups
|
|
Quasi-experiment
|
Subj are not randomly assigned to groups
|
|
Random assignment
|
Ensure all subj have equal change, create initial equivalency, ensure internal validity
|
|
Random assignment
|
Ensure all subj have equal change, create initial equivalency, ensure internal validity
|
|
Range
|
diff bw highest and lowest score in a distrib, meas of variation
|
|
Range
|
diff bw highest and lowest score in a distrib, meas of variation
|
|
Ratio scale
|
arranged in order and intervals are equal, has an absolute zero, mult and division are possible
|
|
Ratio scale
|
arranged in order and intervals are equal, has an absolute zero, mult and division are possible
|
|
Regression equation
|
Prediction of an unknown value of one variable from the known value of another
|
|
Regression equation
|
Prediction of an unknown value of one variable from the known value of another
|
|
Regression Line
|
Straight line (line of best fit) that best depicts a relat bw 2 variables
|
|
Regression Line
|
Straight line (line of best fit) that best depicts a relat bw 2 variables
|
|
Reversal design
|
Single subj design that involves the w/d of the tx so that it can be determined if bx reverts to baseline, ABA and ABAB
|
|
Reversal design
|
Single subj design that involves the w/d of the tx so that it can be determined if bx reverts to baseline, ABA and ABAB
|
|
Sampling distrib of means
|
Theoretical distrib of means of various sample one can draw of equal size from one pop. Each sample will have a mean close to pop mean . Yield a distrib of samples means that will approach normal shape, clustered around pop mean, and have a mean equal to pop mean
|
|
Sampling distrib of means
|
Theoretical distrib of means of various sample one can draw of equal size from one pop. Each sample will have a mean close to pop mean . Yield a distrib of samples means that will approach normal shape, clustered around pop mean, and have a mean equal to pop mean
|
|
Power
|
Prob of rejecting a false null, ability of stat test to correctly id tx effect
|
|
Pretest sensitization
|
Alters subj reactions to IV, threatens external validity, can't be generalized to situations when pretest not admin
|
|
Quasi-experiment
|
Subj are not randomly assigned to groups
|
|
Random assignment
|
Ensure all subj have equal change, create initial equivalency, ensure internal validity
|
|
Range
|
diff bw highest and lowest score in a distrib, meas of variation
|
|
Ratio scale
|
arranged in order and intervals are equal, has an absolute zero, mult and division are possible
|
|
Regression equation
|
Prediction of an unknown value of one variable from the known value of another
|
|
Regression Line
|
Straight line (line of best fit) that best depicts a relat bw 2 variables
|
|
Reversal design
|
Single subj design that involves the w/d of the tx so that it can be determined if bx reverts to baseline, ABA and ABAB
|
|
Sampling distrib of means
|
Theoretical distrib of means of various sample one can draw of equal size from one pop. Each sample will have a mean close to pop mean . Yield a distrib of samples means that will approach normal shape, clustered around pop mean, and have a mean equal to pop mean
|
|
Sampling error
|
Extent to which sample values (stat) deviate from pop values (parameters)
|
|
Standard deviation
|
Square root of variance of distrib of scores, measure of variability of distrib
|
|
Standard error of the mean
|
Index of how much a given sample can be expected to deviate from the pop mean. Derived by dividing the standard deviation by the square root of the sample size
|
|
Statistical regression
|
Stat artifact predicting that extreme scores will by less extreme on re-testing. May threaten internal validity when extreme scorers are being used as research subj
|
|
T-score
|
Standard score with a mean of 50 and an sd of 10
|
|
t-test
|
parametric test assess diff bw 2 means
|
|
Time-series design
|
subj meas on dv on mult occasions at regular intervals before and after iv is admin
|
|
transformed scores
|
scores (z-scores, percentile ranks) expressed in terms of a scale diff than raw score scale, permit comparison of indiv score to other scores in the distrib
|
|
Trend analysis
|
Stat tech looking for a pattern of change bw 2 quantitative variables, linear, quadratic, cubic, quartic
|
|
True experiment
|
Involves experimental control of variables and random assign of subj, stronger in terms of internal validity
|
|
Two-tailed test
|
Stat test used when alt hyp is nondirectional (2 means are diff no specification of direction)
|
|
Type I error
|
Erroneously rejects null - no diff among pop, conclude that a diff exists, when there is isn't one, finding something when nothing is there
|
|
Type II error
|
Erroneously accepts null, no diff in pop when there is one, there is a tx effect but failed to find it
|
|
Variability
|
Dispersion, spread of scores in distrib
|
|
Variance
|
Average of squared differences from mean of each score, meas of variability
|
|
z-scores
|
Meas of how many sd a raw score is from the mean. z-score of +2.0 indic 2 sd above mean. Formula is (X-M)/sd
X = raw score |