The research glossary defines terms used in conducting social science and policy research, for example those describing methods, measurements, statistical procedures, and other aspects of research; the child care glossary defines terms used to describe aspects of child care and early education practice and policy.
Interval Variable
A variable wherein the distance between units is the same but the zero point is arbitrary.
Intervening Variable
An intervening variable or mediating variable is something that effects or helps to explain the relationship between an independent and a dependent variable. An intervening variable is the link between an independent and a dependent variable. It is predicted by the independent variable, and it predicts the dependent variable. In early childhood research, intervening variables are often used to identify the processes that underlie the relationships between independent variables and dependent variables. For example, the time parents spent reading to children may explain the relationship between parental education and children's reading scores on a standardized test.
Intervention
The situation or variable introduced to affect the dependent variable (outcome); manipulations of the subject or the subject's environment performed for research purposes. In early education, interventions often involve the introduction and use of one or more instructional approaches designed to improve children's learning. Other interventions might include professional development training designed to improve the quality of teaching in the classroom and ultimately children's learning.
Interviewer Error
A type of non-sampling error caused by mistakes made by the interviewer. These may include influencing the respondent in some way, asking questions in the wrong order, or using slightly different phrasing (or tone of voice) than other interviewers. It can include intentional errors such as cheating and fraudulent data entry.
Jackknife Technique
A (usually) computer-intensive resampling method used to estimate population parameters (for example, means and percentage), and/or to gauge uncertainty in these estimates (e.g.,standard error). The name is derived from the approach that involves removing each observation (i.e., cut with a knife) one at a time (or two at a time for the second-order Jackknife, and so on), calculating the mean for each new sample (original sample minus the omitted case) and then averaging the means of the new samples.
Kurtosis
A statistic that measures how outlier-prone a distribution is. The kurtosis of a normal distribution is 0. If the kurtosis is different from 0, then the distribution produces outliers that are either more extreme (positive kurtosis) or less extreme (negative kurtosis) than are produced by the normal distribution.
Latent Growth Model
Latent growth modeling (LGM) is a class of statistical methods that are used to study change (growth) in behavior or attitudes over time. Traditional approaches to the study of change such as regression analysis and ANOVA focus on mean change (average change of a group or subgroups of study participants) and treat differences in change between participants as error. LGM examines individual (within-person) change over time as well as differences in the individual change (between-person). It is used to model change over time and to investigate factors that affect the level and rates of change. It can be used to examine differences in which groups of people change.
Latent Variables
In statistics, latent variables are variables not directly observed and measured but inferred from other observed and measured variables. Mathematical models (e.g., factor analysis, structural equation modeling, item response theory models) are used to examine the relationships between a set of observed variables (indicators) in order to identify the latent variable. For example, the latent variable 'teacher attitudes toward math' may be modeled from a series of survey items asking about their feelings toward math and how they feel when doing math.
Least Squares
A commonly used method for calculating a regression equation. This method minimizes the difference between the observed data points and the data points that are estimated by the regression equation.
Level of Significance
See significance level.