内容节选：

The Structural Equation Model

A structural equation model consists of two parts: (1) the measurement model and (2) the structural model.  The measurement model is the part of the conceptual model that specifies the relationships between the observed variables and the constructs (or latent variables).  The structural model, on the other hand, specifies the causal relations among the constructs or latent variables (i.e., the hypothesis you want to test).  In a conceptual/structural equation model, constructs or latent variables can be classified into two groups: exogenous variables and endogenous variables.  Exogenous variables are latent variables that only affect other latent variables (i.e., endogenous variables); they are not affected by any latent variable at all.  Endogenous variables are latent variables that are affected by at least one exogenous variable.  They may or may not affect other endogenous variables.

In the example I used in the class, I hypothesized that intention to buy (η2) is positively affected by attitude toward the brand (η1), which is in turned positively affected by perceived quality (ξ1) and negatively affected by perceived price (ξ2).  In this example, perceived quality and perceived price were exogenous variables and were labeled as ξ1 and ξ2, respectively.  Attitude toward the brand and intention to buy were endogenous variables and were labeled as η1 and η2, respectively.  In the example, perceived quality, perceived price, attitude toward the brand, and intention to buy were measured by 4 items (x1 to x4), 3 items (x5 to x7), 3 items (y1 to y3), and 3 items (y4 to y6), respectively.

内容节选：


Steps in Research Methodology

There are seven steps in (causal) research: (1) problem formulation, (2) hypothesis formulation, (3) research design, (4) sample and sampling, (5) data collection, (6) data analysis, and (7) manuscript writing.

。。。。。。。。。。

内容节选：

Steps in Using Multivariate Data Analysis Techniques

There are 6 steps in using multivariate data analysis techniques:
•        Step 1: Define the research problem, objectives, and multivariate technique to be used.  It is important for your research to be theory-driven, not data-driven.
•        Step 2: Develop the analysis plan.  Issues such as sample size, scales of the variables, and estimation method need to be addressed.
•        Step 3: Evaluate the assumptions underlying the multivariate technique.  Check your data to see if any assumption for the particular multivariate technique is violated.  Any serious violation of one or more assumptions may render the data useless.
•        Step 4: Estimate the multivariate model and assess the overall model fit.  A proper estimation procedure is used and the overall model fit is assessed in this step.  Model re-specification is possible in this step.
•        Step 5: Interpret the variate.  The nature of multivariate relationship is investigated and interpreted at this step.
•        Step 6: Apply the diagnostics to the results.  This step normally concerns the duplication of the multivariate technique with a holdout sample or a new sample.

In practice, many researchers skip Step 2, Step 3, and Step 6.  In fact, every step is important and useful and should be followed during the pretest as well as the main study.

自己顶一下

内容不错，顶一下