The first step in fitting any linear models should be exploratory graphing with the raw data. In many cases, this step will provide preliminary answers to the hypotheses and will provide a general idea of the effect sizes for the predictors. Sometimes, descriptive graphs will not provide the whole picture, in particular when the response […]
Path analysis is a form of a structural equation model in which all the variables in the model are observed and all the paths of the model are estimated simultaneously. Simply put, path analysis is a series of linear regressions. Hence, the same assumptions that we have for linear regression hold for path analysis. Path […]
Path analysis is a form of a structural equation model in which all the variables in the model are observed and all the paths of the model are estimated simultaneously. Simply put, path analysis is a series of linear regressions. Hence, the same assumptions that we have for linear regression hold for path analysis. Path […]
Choice models can be used to describe how any individual makes a choice between a discrete set of choice alternatives. These methods are widely used in many fields including social sciences, economics, marketing, transportation, health, and animal behavior. This workshop will consist of a detailed description of the multinomial logit model, and the related conditional […]
Penalized regression is a modern statistical method that has gained much attention over the last decade as researchers in many fields are able to measure far more variables than ever before. Linear regression suffers in two important ways as the number of predictors becomes large. First, overfitting may occur, meaning that the fitted model does […]
The JMP DOE platform provides a user friendly way to construct study designs that accommodate a variety of types of factors and constraints common to experimental research. This workshop will illustrate how to use this platform to design, evaluate and compare: Full factorial designs Randomized block designs Incomplete block designs Split plot designs Response surface […]
One of the first steps in planning a new study is determining an appropriate sample size. A sample size should be large enough to have a high probability of detecting an effect of a treatment but small enough that it is within the confines of the study?s budget and minimizes the potential risks to human […]
Multilevel models (also referred to as hierarchical models or mixed models) are a class of statistical models that can be used when observations are not independent. Non-independence can occur when data is clustered due to the study design (e.g., data collected on households and their individual members or blocked agricultural studies) or when data is […]
In a longitudinal study data has been collected on the same observation over time. Since the repeated measures taken on the same observations are not independent from each other this is a type of multilevel data and special care needs to be taken during the analysis to account for this non-independence. These types of data […]
Panel data, also known as longitudinal data, consists of multiple observations of the same subject over time. Many disciplines will analyze such data using a random intercept or random coefficients model. In this workshop, another way of modeling such data will be discussed: using fixed effects. Fixed effects methods can be used in longitudinal studies […]