Term
| stratified random sampling |
|
Definition
| Efficient for probability sampling; separate sample selected from homogeneous layer of the population. These layers are then combined to form conclusions about each respective subgroup |
|
|
Term
| area probability sampling |
|
Definition
| population divided into geographic areas (i.e. trying to find out urban area population) |
|
|
Term
|
Definition
| Sample selected from undivided population, where each thing in the population has an equal chance of being selected. |
|
|
Term
|
Definition
| How well the measure or design des what it purports to do. |
|
|
Term
|
Definition
| soundness of statements about whether one variable is the cause of a particular outcome |
|
|
Term
|
Definition
degree to which the conceptualization of what is being measured is what it is claimed. |
|
|
Term
|
Definition
| adequate sampling of the relevent material/content that a test purports to measure |
|
|
Term
|
Definition
| convergence of related test or behavior |
|
|
Term
|
Definition
| distinctiveness of unrelated tests or behavior ( part of construct validity) |
|
|
Term
|
Definition
| degree to which a test or questionnaire is correlated with outcome criteria in the present ( concurrent validity) or the future ( predictive validity) |
|
|
Term
|
Definition
| generalizability of an inferred causual relationship over different people, settings, research outcomes |
|
|
Term
| internal consistency relability |
|
Definition
| overall degree of relatedness of all items in a test or all raters in a judgment study |
|
|
Term
|
Definition
| degree of temporal stability (relatedness) of a meauring instrument or test. (one administration to another) |
|
|
Term
|
Definition
| threat to internal validity; an uncontrolled event that occurs between the premeasurement and the post-measurement. (i.e. fire drill interrupting) |
|
|
Term
|
Definition
| threat to internal validity; certain intrinsic changes in the research participants, growing older, wiser, stronger, more experienced between pretest and postest. |
|
|
Term
|
Definition
threat to internal validity; intrinsic changes in the measuring instrument, such as deterioration |
|
|
Term
|
Definition
| threat to internal validity; selection of participants to a particular condition ( unsuspected differences between participants in each condition) |
|
|
Term
|
Definition
aka standard score z = each score - mean/standard deviation |
|
|
Term
|
Definition
| fluctuations that are not random, but slanted in a particular direction ( AKA bias) |
|
|
Term
|
Definition
| chance fluctuations or haphazard errors ( AKA noise) |
|
|
Term
| quasi-experimental design |
|
Definition
| do not randomly assign individual units to treatment conditions; quasi = "resembling" |
|
|
Term
| interrupted time-series design |
|
Definition
| compare the "effects" of an intervention in the situation before and after it occurs |
|
|
Term
| cross-lagged correlation design |
|
Definition
| show relationship between 2 sets of data points, where one point is treated as a lagged value of the outcome variable |
|
|
Term
|
Definition
| study of variation across some dimension over time (time series = data point for each point in time) |
|
|
Term
|
Definition
| correlation coefficient; standard index of linear relationship |
|
|
Term
|
Definition
| Pearsons r where both variables are dichotomous |
|
|
Term
|
Definition
Pearson r computed on scores in ranked form * no different than Pearson except for ranked numbers rather than a rating scale* |
|
|
Term
| point biserial correlation |
|
Definition
| scores for one variable is on a continuum and the other variables are dichotomous ( i.e. one continuous and one dichotomous) |
|
|
Term
|
Definition
| both variables are dichotomous |
|
|
Term
|
Definition
response set in which the bias results from the judge's overextending a favorable impression of someone, based on some central trait, to the person's other characteristics. |
|
|
Term
|
Definition
| rating error in which ratings are consistently more positive than they should be |
|
|
Term
|
Definition
| respondents are overly agreeable * go along with almost any statement* |
|
|
Term
|
Definition
| respondents hesitate to give extreme ratings and instead tend to cluster their responses around the center choice. |
|
|
Term
| multiple regression analysis |
|
Definition
| taking multiple predictors and trying to make a correlation between them to form a line ( regression) |
|
|
Term
|
Definition
| location of the bulk of a distribution; measured by means, medians, modes |
|
|
Term
|
Definition
| statistic used to test the degree of agreement between frequency ( count) data actually obtained and those expected |
|
|
Term
|
Definition
| procedure in which some subjects receive treatment A before treatment B and others receive B before A (in Latin Square Designs) |
|
|
Term
|
Definition
the number of observations minus the number of restrictions limiting the observations freedom to vary (in a chart, it is already N-1 so you must add one to obtain how many levels there are for a given factor - i.e. Age says 3 so it is 4 df ( 4-1 = 3) with independent t test - (n-1) + ( n-1) I so 2 groups with 10 people in each ( 10-1)+(10-1) 18 |
|
|
Term
|
Definition
| general name for a variable, the independent variable |
|
|
Term
|
Definition
| research design with more than one factor and two or more levels of each factor |
|
|
Term
|
Definition
| ratio of mean squares that are distributed as F when the null hypothesis is true, where F is a test of significance to judge the tenability of the null hypothesis of no relationship between 2 or more variables |
|
|
Term
|
Definition
| equality of the population variance of the groups to be compared |
|
|
Term
|
Definition
|
|
Term
|
Definition
| a design with 2 or more factors, with at least one between and one within subject |
|
|
Term
|
Definition
| meaning of a variable in terms of the operations used to measure it or the experimental methods involved in its determination |
|
|
Term
|
Definition
| statistic for estimating the replicability of an obtained effect |
|
|
Term
|
Definition
| estimatino of the effective powe of a statistical test, or of the sample size needed to detect an obtained effect given a specificed level of power |
|
|
Term
|
Definition
| stastical design in which the sampling units generate two or more measurements |
|
|
Term
|
Definition
| aka root mean square; an index of the variability of a set of data around the mean value in a distribution |
|
|
Term
|
Definition
| error of rejecting null hypothesis when it is true |
|
|
Term
|
Definition
| error of failing to reject the null hypothesis when it is false |
|
|
Term
|
Definition
| mean of the squared deviations of scores from their means in a population |
|
|
