Classical Test Theory: A method of scoring psychometric assessment that assumes all items included in the assessment contribute equally to the observed outcome.
The Spearman Brown Prophecy Formula: A mathematical formula used by psychometricians to predict the reliability of an assessment when the length of the test has been changed.
Shared Variance: Variance pertains to the amount of an outcome that can be explained by measuring a particular construct. Shared Variance is therefore the amount of overlapping information explained by more than one variable.
Items: Items are part of an assessment or test, in the context of a personality measure they are the statements upon which individuals rate themselves.
Reliability: is the degree to which we can trust a measurement to be accurate.
Internal Consistency: is a part of reliability in which we determine how well items in a construct measure that particular construct based on how well the items relate to each other. If different questions should all be measuring the same thing; we’d expect those question to be related to each other.
Alternate Form: is a type of reliability in which participants complete different measures of the same measure of an assessment at different points. If the assessment is reliable, we’d expect it to have similar outcomes across various forms. An example of this might be completing an assessment on a laptop compared to a handheld device; we’d expect a person’s results to be similar on both.
Test Re-test: is another method of evaluating reliability in which the same test is administered at different time points and the outcome appears similar. For instance, considering a personality measure, if a great amount of time hasn’t passed and the person’s circumstances are similar, we’d expect two measures of that individual’s personality to be similar over time.
Personality Scale: a model of items which fit together to measure a certain construct. For instance, several items could be assessed together for the scale of ‘Thought’.
Scale Hierarchy: is a method of structuring scales in a measure from very specific measures at the bottom of the hierarchy to broader, more generalizable constructs at the top of the model.
Traits: measurable constructs of an individual’s personality.
Big Five: the most widely accepted model of personality explains that personality consists of five broad constructs. These are thought to be Openness to experience; Conscientiousness; Extraversion; Agreeableness and Emotional Stability.
Likert Formats: a bipolar scale upon which people are prompted to select the degree to which they agree or disagree with a statement.
Validity: is the quality of your findings truly representing the thing that you’re trying to measure. For example, if a measure validly measures Openness to experience, it must genuinely measure this trait rather than relating to a different characteristic.
Cross-validation: a method of evaluating validity across different contexts to ensure that an assessment, for example, still does what it’s meant to be doing in different circumstance.
Convergent validity: is a method of demonstrating that a measure evaluates what it’s intended to evaluate by looking at how well it relates to other measures that are theoretically related. For instance, a valid measure of self-confidence ought to also converge with a valid measure of Self-worth as we would expect these traits to be related to one another.
Discriminant validity: is the converse of convergent validity. Traits which are not theoretically related ought not be found to related with different measures. That is, it seems unlikely that a person’s degree of Conscientiousness should be related to their level of Extraversion, so we’d expect a valid measure of these traits to indicate this relationship as such.
Adverse Impact: refers to employment practices which are intended to be neutral towards different demographics but are found to have an underlying discriminatory effect on a certain group of people