Statistics

is:

1. presumed cause, antecedent, predictor variable, treatment(s), factor(s), denoted by X.


2. also, a manipulated variable in an experiment or study whose presence or degree determines the change in the dependent variable.

3. independent variables may have one or more levels.

is:

1. outcome measure, presumed effect, conquence, variable predicted to, criterion, observation, denoted by Y.

2. also, the observed variable in an experiment or study whose changes are determined by the presence or degree of one or more independent variables.

1. investigator uses random assignment of subjects to treatment groups.

2. used to establish a cause and effect relationship.

3. investigator manipulates independent variables.

1. investigator must deal with intact groups.

2. used to establish a cause and effect relationship.

3. investigator manipulates independent variables.

are:

1. manipulated by researcher (i.e. treatment type), or

2. potentially manipulable (i.e. subject matter studied.)

are:

1. fixed - not manipulated by researcher (i.e. socioecomonic status).

2. organismic(i.e. age, gender).

3. response (i.e. scores on tests).

is:

1. ex post facto - used to explain or predict.

2. uses two or more groups and one independent variable.

3. one group is a comparison group.

is:

1. ex post facto - used to explain or predict.

2. uses one group and two or more independent variables.

3. does not use a comparison group.

1. after the fact.

2. explains or predicts.

3. does not manipulated the independent variable.

1. describe phenomena as they exist.

2. no independent or dependent variables.

1. is prospective.

2. a study of the development of subjects over an extended period of time.

1. is retrospective.

2. studies subjects of various ages at the same point in time.

3. relies on recollection of subjects.

internal validity is the extent that what we are doing is correct.


external validity is the extent to whom the results can it be generalized.

univariate statistical technique employed to compare one factor, with more than one level, on the basis of one outcome measure.

univariate statistical technique employed to compare two or more factors, each with more than one level, on the basis of one outcome measure.

1. is the extent to which an instrument/a test measures what it is suppose to measure.

2. Validity implies reliability.

is:

1. when an instrument/a test yields consistent results (internal consistency)

2. reliability is sometimes called "poor man's validity".

3. reliability does not imply validity.

1. each element in the population has an equal chance/probability of being included in the sample.

2. sample is representative of the population.

1. each element in the population does not have an equal chance/probability of being included in the sample.

2. sample not representative of the population.

Normal distribution is symmetrical with the mean = medium = mode.

skewed distribution is asymmetrical with the longer tail extending away from the X and Y origin (positively skewed) or with the longer tail extending toward the x and Y origin (negatively skewed).

Statistics are indices (descriptive measures) for sample.


Parameters are indices (descriptive measures) for population.

Univariate statistics have one outcome measure.


Multivariate statistics have more than one outcome measure.

correlation between one X (independent) variable and one Y (dependent) variable.

correlation between one X (independent) variable and one Y (dependent) variable, while partialling out confounding variable(s).

(i.e. r xy.z shows a first order partial correlation between X and Y, partialling out z

or r xy.zw shows a second order partial correlation between X and Y, partialling out z and w.)

Multiple correlation is between more than one X (independent) variables and one Y (dependent) variable.


Canonical correlation is between more than one X (independent) variables and more than one Y (dependent) variables.

Null hypothesis (Ho)is the hypothesis of: no "difference", no "effect", no "relationship".

Research hypothesis (H1)is the researcher's hunch.

Type I error is falsely rejecting a true null hypothesis.


Type II error is not rejecting a false null hypothesis.

Power is correctly rejecting a false null hypothesis (1 - Beta).


Confidence coefficient is correctly not rejecting a true null hypothesis.

One-tailed test is when directionality is specified in the research hypothesis.


Two-tailed test is when directionality is not specified in the research hypothesis.

1. more likely to find statistical significance.

2. has more power.

3. more given to chance.

4. more likely to make Type I error.

5. less likely to make Type II error.

1. less likely to find statistical significance.

2. has less power.

3. less given to chance.

4. less likely to make Type I error.

5. more likely to make Type II error.

1. identifies sample from population.

2. establishes external validity.
(The sample cannot be generalized to the population without random selection.)

1. assigns subjects from sample to independent variables.

2. establishes internal validity.

1. univariate statistical technique employed to compare one factor, with more than one level, on the basis of one outcome measure, while controlling for a confounding variable (covariate).


2. confounding variable should be correlated with outcome variable.

multivariate statistical technique employed to distinguish among groups on the basis of their centroids from more than one outcome measure.

Descriptive statistics describe, organize, summarize data.

1. Inferential statistics infer from the sample to the population.

2. hypothesis testing and estimation are the two major branches of inferential statistics.

parametric statistics are most often used with interval or ratio (quantitative) data.


non-parametric statistics are most often used with nominal (qualitative) or ordinal data.

1. true variance, explained variance.

2. denoted by r(squared).

3. it shows the proportion of variance in y explained by x).

4. measure of practical significance.

1. error variance, unexplained variance.

2. denoted by 1-r(squared).

3. shows the proportion of variance in y not explained by X (i.e. if 25% of the variance in Y is explained by X, 75% is the unexplained variance).

location of point on a distribution, (i.e. mean, mode, medium).

degree of variation in a set of data, (i.e. standard deviation, SIQR, index of dispersion, range).

1. prediction of one outcome variable (Y) from one predictor variables (X).


2. Y is a continuous variable.

1. prediction of one outcome variable (Y) from more than one predictor variables (X).


2. Y is a continuous variable.

1. prediction of one outcome variable (Y) from one or more than one predictor variables (X).


2. Y is a dichotomous (binary) variable.

multivariate statistical technique employed to describe major differences amoung groups based on their group centroids from more than one outcome measure or one categorical outcome measure with more than two levels.

a main effect is an outcome that is a consistent difference between levels of a factor.


an interaction effect exists when differences on one factor depend on which level you are on in another factor.

target population is the ideal/theoretical population the researcher has in mind.


accessible population is the population available to the researcher.

the way we define/measure variables in Experimental Research.


the way we define/measure variables in any other research.

between factor is the among groups/subjects factor.


the within factor is the repeated measures factor.

subdivisions of the independent variable (factor).

measure of practical significance, measure of observed effect (i.e. Cohen's d, r-squared, R-squared, eta squared, phi, Cramer's V.)