David A. Kenny
April 9, 2016

 

Multiple Latent Variable Models:
Confirmatory Factor Analysis


Standard Exploratory Factor Analysis Model or EFA
        Every measure loads on each factor
              either uncorrelated (orthogonal) or correlated (oblique)
              generally factors are uncorrelated
        Because with more than one factor, the solution is not unique (i.e., underidentified), it can be rotated.

        To test if k factors are sufficient to explain the covariation between measures estimate the following loading matrix (assuming k = 5) with orthogonal or uncorrelated factors with unit variance:

Measure     1         2        3        4        5
        1          x        0        0        0        0
        2          x        x        0        0        0
        3          x        x        x        0        0
        4          x        x        x        x        0
        5          x        x        x        x        x
        6          x        x        x        x        x
        7          x        x        x        x        x
        8          x        x        x        x        x

If a model with this loading structure is good fitting (see Measures of Fit), then k factors are sufficient.
EFA is useful when the researcher does not know how many factors there are or when it is uncertain what measures load on what factors.
Find out about a book that discusses both EFA and CFA.

Confirmatory Factor Analysis Model or CFA (an alternative to EFA)
        Typically, each variable loads on one and only one factor.
        Factors are correlated (conceptually useful to have correlated factors).
        Generally errors (or uniquenesses) across variables are uncorrelated.
        Variables in CFA are usually called indicators.
Parameters
          Loadings
                    Path from the latent variable to the indicator
                              Standardized path is a factor loading.
                    At least one loadings per factor is fixed to one (marker variable).        
          Error variance for each indicator
          Factor variance (fixed to one in EFA, but not in CFA)
          Factor covariance
                    Unlike EFA, latent variables are correlated.

Degrees of Freedom (df) for CFA Models
Unknowns
          Free loadings (do not count marker variable or loadings set equal)
          Error variances
          Correlated errors
          Factor variances
          Factor correlations
Knowns: k(k + 1)/2
Typically CFA models with several factors and indicators have many df.

CFA

Identification
Given k factors, there must be k2 constraints.
Usually k of these constraints are scaling ones (i.e., marker variables).

The standard EFA model with two or more factors and all the loadings free is not identified.  This is why the solution can be rotated.

Standard CFA model: Simple Structure
        Each measure or indicator loads on one and only one factor which implies no double loadings.

          No correlated errors
          Latent variables correlated
Simple Structure CFA model is identified:
             If there are, at least, two indicators per latent variable and the errors of those two or more indicators are uncorrelated with each other and with at least one other indicator on the other latent variables.

Testing in CFA and Structural Equation Modeling
Principle of nesting:  Model A is said to be nested within Model B, if Model B is a more complicated version of Model A. For example, a one-factor model is nested within a two-factor as a one-factor model can be viewed as a two-factor model in which the correlation between factors is perfect).
Relative fit of a nested model: the chi square difference test, the smaller chi square and its degrees of freedom are subtracted from the larger chi square and degrees of freedom.
In principle, the more complicated model should fit for the test to be valid.

Discriminant Validity
Definition of poor discriminant validity: The correlation between two factors is or is very close to one or minus one.
          Consequences

                    multicollinearity: If the factors are treated as causes of a third factor, the high collinearity leads to very large standard errors.
                    problems of convergence and inadmissabile solutions
          Criteria: A correlation of .85 or larger in absolute value indicates poor discriminant validity
          Test:  Estimate a model that fixes the correlation to one (Do not use a marker variable strategy, but instead fix factor variances to one.) or collapse the two factors and see if the model fit worsens.

Example 1: Unpublished Master’s Thesis of Julie Fenster: “Multidimensional measurement of Religiousness/Spirituality for use in health research assessment developed by the Fetzer Institute”

Three Latent Variables

Daily Spiritual Experiences (DSE)
                    I feel God’s presence.
                    I am touched by the beauty of creation.
Private Religious Practices (PRP)
                    Private prayer.
                    Read the Bible.
Positive Religious and Spiritual Coping (PRSC)
                    Think about life as part of a larger spiritual force.
                    I look to God for strength, support and guidance. 
Correlations
                DSE with PRSC = .869
                PRP with PRSC = .918
                DSE with PRP = .910 

See also “Exploring the Dimensionality of "Religiosity" and "Spirituality" in the Fetzer Multidimensional Measure” by J. A. Neff, Journal for the Scientific Study of Religion, 45, 449‑459.

           Example 2: Salovey, P., & Rodin, J. (1984). Some antecedents and consequences of social-comparison jealousy. Journal of Personality and Social Psychology, 47, 780-792.

          One latent variable model χ²(5) = 24.305
                    Two latent variable model χ²(4) = 8.669

                    chi square difference test: χ²(1) = 15.636, p < .001
                    conclusion: two latent variables are needed

Salovey & Rodin Example with Standardized Estimates

Discriminant Validity

           Example 3: Braze, D., Katz, L., Magnuson, J. S., Mencl, W. E., Tabor, W., Van Dyke, J. A., Gong, T., Johns, C. L., & Shankweiler, D. P. (2016). Vocabulary does not complicate the Simple View of Reading. Reading and Writing, 29, 435-451: In this paper, they show that language comprehension (LC) and reading comprehension (RC) have poor discriminant validity.

Braze et al. Example (Standardized Estimates)

Discriminant Validity

Respecification (see Respecification page for more detail)
Criteria
    Empirical (again see Respecification page for more details)
          Correlation matrix
          Modification indices (also called Lagranian multipliers)
                     The estimated change in chi square if the parameter were freely estimated.
          Standardized residuals
                      If model is correctly specified, large values (greater than 1.96 in absolute value) indicate correlations poor fitted.
                      In my experience, these values tend to be conservative (i.e., too small).
    Theoretical: All respecifications require some rationale and that rationale should be extended to other cases.

Types of Respecifications    
  
Resulting in a MORE COMPLEX MODEL (i.e., more parameters)
          Another factor
          Correlated errors
                    definition:  Variance not explained by theoretical constructs may covary across two measures.  Such covariance is referred to as a correlated error.
          Double loadings
 
Resulting in a SIMPLER MODEL (i.e., few parameters)
Note that making the model simpler, while often a very reasonable thing to do, does not improve the fit of the model.
          Fewer factors
          Equal loadings (should be done using the covariance matrix or raw data)

Specialized Issues
          How many indicators per factor?
                    2 is the minimum
                    3 is safer, especially if factor correlations are weak
                    4 provides safety
                    5 or more is more than enough (If too many indicators then combine indicators into sets)
          What to do about “too many” indicators? Parcels or “testlets”
                    Definition: Adding (or averaging) sets of indicators up to create a smaller number of indicators
                    Strategies
                              random
                              conceptually similar sets
                              sets that may contain items with correlated errors
                    Disadvantages of parceling
                              loss of information
                              possibility of specification error that is missed and becomes undetectable                   
                    Advantages of parceling
                              smaller models (better participant to parameter ratio)
                              more “normal” distributions of variables
                              usually better fit
                    Compromise strategy:  Run individual CFA on each latent variable and then parcel.

Single Indicators
          measures with no measurement error
                   Treat as variable in most programs but LISREL requires
                         fix loading to one
                         free variance if exogenous or disturbance if endogenous
                         fix error variance to zero
                         do not correlate its "error variance" with anything 
          measures with measurement error
                    fix loading to one
                    free variance if exogenous or disturbance variance if endogenous
                    error variance
                              fix to a known value (see Williams and Hazer) or
                              find an instrumental variable

Models with Means
Strategy 1
                fix factor mean (if exogenous) or intercept (if endogenous) to zero
                free all indicator intercepts 
Strategy 2
                free factor mean (if exogenous) or intercept (if endogenous)
                fix the marker variable’s intercept to zero
                free all other indicators’ intercepts
The model fit and other parameter estimates (e.g., loadings) are the same for both strategies.  Most people find the second strategy simpler and easier to work with. 


Go to the next SEM page.
Description: Causal Modeling logo
Go to the main SEM page.