Term
|
Definition
| An independent variable is the variable in a study you have control over; you manipulate it to achieve some sort of experimental effect with the DV. |
|
|
Term
| Interorganizational validity |
|
Definition
| A term specific to industrial/organization psychology, interorganizational validity is an external validity related term that refers to whether or not a practice that works in one part/department/area of a company is capable of working in another area of the same company. |
|
|
Term
|
Definition
| A predictor score is a numerical measure on a particular assessment designed to give information about an applicant that be used for selection decisions. Often, multiple predictor scores can be combined together to obtain a multifaceted assessment of an applicant for a particular job. |
|
|
Term
|
Definition
| Situational specificity refers to a state when observed validity coefficients will vary considerably from study to study even when the jobs and tests are quite similar in nature. Based on this find, researchers have decided that a validation study must be conducted when a test is used in a different job, situation, or company or with a different set of applicants applying for a job. |
|
|
Term
|
Definition
| Transportability refers to a state when an individual with a certain set of professional skills can be moved to a completely different context. For instance, a person with an education degree may now have skills that would be relevant in an HR job. |
|
|
Term
|
Definition
| Validity is the extent a test measures what it actually claims to measure. It refers to the overall accuracy of a test. An intelligence test would have a high validity if it consistently was accurate in assessing intelligence. |
|
|
Term
|
Definition
| A validity generalization is an application of meta-analysis to the correlations between the results of a particular employment test and a predetermined criterion. Particularly, validity generalization is designed to get a good idea of how valid a test is over time, attempting to test and mange the situation specificity hypothesis. A mean validity coefficient is obtained across all the studies and applied to some preset criterion. |
|
|
Term
| Concurrent validation design |
|
Definition
| A concurrent validation design is a study that is oriented towards establishing a selection system. This validation type will have the selection system be tested on current employees and compare the predictor measurements to the current criterion of work performance. |
|
|
Term
|
Definition
| Construct validity is a way of assessing the validity of an instrument. When an instrument possesses high construct validity, it is measuring the underlying constructs of the behavior accurately. |
|
|
Term
|
Definition
| A content validity ratio is a measure based on a panelist discussion on determining whether or not a test item is essential to measure a skill related to a job. Items are eliminated if they fail to meet the CVR significance. |
|
|
Term
|
Definition
| Content based validity is essentially an estimate of much a particular measure fully captures the domain of work behavior expected of a job and using it as a predictor of future behavior. If a test has good content validity, it will be able to accurately assess all of the information needed to make a decision about employment selection. |
|
|
Term
|
Definition
| A type of construct validity, convergent validity refers to when two particular measures are measuring the same construct, even when the measures might be gathered independently of each other. ie: NEO PI-R and the FFI measuring personality related to the FFM. |
|
|
Term
| Corrected validity coefficient |
|
Definition
| Corrected validity coefficients are typically used when there is a restriction in range. They need to be corrected in order to interpret validity coefficients properly. There are three different types of information that need to be taken into consideration to decide which correction to perform: 1) whether restriction occurs on the predictor, criterion, or a third variable; 2) whether the unrestricted variances for the relevant variables are known; 3) whether the third variable, if even included involved the measure at all, is measured or not measured. |
|
|
Term
|
Definition
| Criterion related validity exists when a particular test has demonstrated a certain amount of effectiveness in predicting the criterion or a particular indicator of a construct. Concurrent validity and predictive validity are two types of criterion-related validity. |
|
|
Term
|
Definition
| A cross validation refers to the extent regression weights derived from a sample can be used to predict outcomes to an equivalent standard in a population as a whole or in samples drawn from the same population. If cross-validity is low, the use of different assessment tools and predictions related to one sample are inappropriate to use with other samples from the same population |
|
|
Term
|
Definition
| A type of construct validity. Discriminant validity refers to the fact that two different types of measures that measures two different types of things should theoretically not be related on what constructs they measure. For example, the MMPI-2 and the Beck Depression inventory measure two completely different things and display a high degree of discriminant validity. |
|
|
Term
|
Definition
| Dust bowl empiricism refers to when no theory is involved behind a study. Anything based off of dust bowl empiricism is solely referring to instances arising from entirely inductive processes. Bottom line, there is no theory involved in the research here whatsoever. |
|
|
Term
|
Definition
| Closely related to cross validation, a holdout sample is set aside part of the data to see how the selected model performs on it. This provides the model with a stronger test of its predictive validity compared to testing it on the same data set from which the particular model of testing was derived. |
|
|
Term
|
Definition
| A meta-analysis is the statistical analysis of a large collection of past analysis results for the sole purpose of putting findings together and summarizing them. This will help derive a common theory about a particular line of research being performed. |
|
|
Term
| Multitrait-multimethod matrix |
|
Definition
| The MTMM is an approach to assessing the construct validity on a set of measures in a study, specifically the convergent validity and the discriminant validity. The matrix presents itself as a correlation matrix where multiple traits within multiple methods are being compared to see if there is construct validity being achieved. Correlations will be performed within and across methods and trait to achieve this matrix. |
|
|
Term
| Predictive validation design |
|
Definition
| A predictive validation design is used to validate a particular selection process. Data are collected on a group of hires that are being admitted into the company or applying for it. This predictor data will be compared to some criterion data after a certain time period in order to see if the selection procedure is valid. |
|
|
