Nothing
Postestimation: Assessing a model"
In cSEM: Composite-Based Structural Equation Modeling
Introduction
As indicated by the name, assess() is used to assess a model estimated using the
csem() function.
In cSEM model assessment is considered to be any task that in some way or
another seeks to assess the quality of the estimated model without conducting
a statistical test (tests are covered by the test_* family of functions).
Quality in this case is taken to be a catch-all term for
all common aspects of model assessment. This mainly comprises
fit indices, model selection criteria, reliability estimates, common validity
assessment criteria, effect sizes, and other
related quality measures/indices that do not rely on a formal test procedure.
Hereinafter, we will refer to a generic (fit) index, quality or assessment measure as
a quality criterion.
Currently the following quality criteria are implemented:
- Convergent and discriminant validity assessment:
- The average variance extracted (AVE)
- The Fornell-Larcker criterion
- The heterotrait-monotrait ratio of correlations (HTMT)
- Congeneric reliability ($\rho_C$), also known as e.g.: composite reliability,
construct reliability, (unidimensional) omega, Jöreskog's $\rho$, $\rho_A$,
or $\rho_B$.
- Tau-equivalent reliability ($\rho_T$), also known as e.g.: Cronbach alpha, alpha, $\alpha$,
coefficient alpha, Guttman's $\lambda_3$, KR-20.
- Distance measures
- The standardized root mean square residual (SRMR)
- The geodesic distance (DG)
- The squared Euclidian distance (DL)
- The maximum-likelihood distance (DML)
- Fit indices
- The $\chi^2$-statistic
- The $\chi^2/df$-statistic
- The comparative fit index (CFI)
- The goodness-of-fit index (GFI)
- The standardized root mean square residual (SRMR)
- The root mean square error of approximation (RMSEA)
- The normed fit index (NFI)
- The non-normed fit index (NNFI)
- The comparative fit index (CFI)
- The incremental fit index (IFI)
- The root mean square outer residual covariance ($\text{RMS}_{\theta}$)
- The Goodness-of-Fit (GoF) proposed by @Tenenhaus2004.
- The variance inflation factors (VIF) for the structural equations as
well as for Mode B regression equations (if
.approach_weights = "PLS-PM").
- The coefficient of determination and the adjusted coefficient of determination
($R^2$ and $R^2_{adj}$)
- A measure of the effect size (Cohen's $f^2$).
- Direct, indirect and total effect assessment.
- Several model selection criteria as described in @Sharma2019.
For implementation details see the Methods & Formulae section.
Syntax & Options
assess(
.object = NULL,
.only_common_factors = TRUE,
.quality_criterion = c("all", "aic", "aicc", "aicu", "bic", "fpe", "gm", "hq",
"hqc", "mallows_cp", "ave",
"rho_C", "rho_C_mm", "rho_C_weighted",
"rho_C_weighted_mm", "dg", "dl", "dml", "df",
"effects", "f2", "fl_criterion", "chi_square", "chi_square_df",
"cfi", "gfi", "ifi", "nfi", "nnfi",
"reliability",
"rmsea", "rms_theta", "srmr",
"gof", "htmt", "r2", "r2_adj",
"rho_T", "rho_T_weighted", "vif",
"vifmodeB"),
...
)
.object
: An object of class cSEMResults resulting from a call to csem().
.quality_criterion
: A character string or a vector of character strings naming the quality criterion
to compute. By default all quality criteria are computed ("all").
See assess() for a list of possible candidates.
.only_common_factors
: Logical. Should only concepts modeled as common
factors be included when calculating one of the following quality criteria:
AVE, the Fornell-Larcker criterion, HTMT, and all reliability estimates.
Defaults to TRUE.
...
: Further arguments passed to functions called by assess().
See args_assess_dotdotdot
for a complete list of available arguments.
Like all postestimation functions assess() can be called on any object of
class cSEMResults. The output is a named list of the quality criteria given
to .quality_criterion. By default all possible quality criteria
are calculated (.quality_criterion = "all").
Details {#details}
In line with all of cSEM's postestimation functions, assess() is a generic
function with methods for objects of class cSEMResults_default,
cSEMResults_multi, cSEMResults_2ndorder. In cSEM
every cSEMResults_* object must also have class cSEMResults for internal reasons.
When using one of the major postestimation functions, method dispatch is therefore technically
done on one of the cSEMResults_* class attributes, ignoring the
cSEMResults class attribute. As long as assess() is
used directly, method dispatch is not of any practical concern to the end-users.
Composite models vs. common factor models
Some assessment measures are inherently tied to the common factor model. It
is therefore unclear how to interpret their results in the context of a
composite model. Consequently, their computation is suppressed by default for
constructs modeled as composites. Currently, this applies to the following
quality criteria:
- AVE and validity assessment based thereon (i.e., the Fornell-Larcker criterion)
- HTMT and validity assessment based thereon
- All reliability measures
It is possible to force computation of all quality criteria for constructs
modeled as composites using .only_common_factors = FALSE, however, we explicitly
warn to interpret results, as they may not even have a conceptual meaning.
All quality criteria assume that the estimated loadings, construct correlations
and path coefficients involved in the computation of a specific quality measure
are consistent estimates for their theoretical population counterpart. If
the user deliberately chooses an approach that yields inconsistent estimates (by
setting .disattenuate = FALSE in csem() when the estimated model contains constructs
modeled as common factors) assess() will still estimate
all quantities, however, quantities such as
the AVE or the congeneric reliability $\rho_C$ inherit inconsistency.
Methods & Formulae {#methods}
This section provides technical details and relevant formulae. For the relevant
notation and terminology used in this section, see the
Notation and the
Termionology help files.
Average Variance Extracted (AVE) {#ave}
Definition
The average variance extracted (AVE) was first proposed by @Fornell1981.
Several definitions exist. For ease of comparison to extant literature the most
common definitions are given below:
- The AVE for a generic construct/latent variable $\eta$ is an estimate of
how much of the variation of its indicators is due to the
assumed latent variable. Consequently, the share of unexplained,
i.e. error variation is 1 - AVE.
- The AVE for a generic construct/latent variable $\eta$ is the share of the total
indicator variance (i.e., the sum of the indicator variances of all indicators
connected to the construct), that is captured by the (indicator) true scores.
- The AVE for a generic construct/latent variable $\eta$ is the ratio of the
sum of the (indicator) true score variances (explained variation)
relative to the sum of the total indicator variances (total variation, i.e.,
the sum of the indicator variances of all indicators connected to the construct).
- Since for the regression of $x_k$ on $\eta_k$, the R squared ($R^2_k)$ is equal to the
share of variation of $x_k$ explained by $\eta_k$ relative to the total
variation of $x_k$, the AVE for a generic construct/latent variable $\eta$
is equal to the average over all $R^2_k$.
- The AVE for a generic construct/latent variable $\eta$ is the sum of the
squared correlation between indicator $x_k$ and
the (indicator) true score $\eta_k$ relative to the sum of the indicator variances of
all indicators connected to the construct in question.
It is important to stress that, although different in wording, all definitions
are synonymous!
The AVE is inherently tied to the common factor model. It
is therefore unclear how to interpret the AVE for constructs modeled as composites.
Consequently, the computation is suppressed by default for
constructs modeled as common factors. It is possible to force computation of the
AVE for constructs modeled as composites using .only_common_factors = FALSE,
however, we explicitly warn to interpret results, as they may not even have a
conceptual meaning.
Formulae
Using the results and notation derived and defined in the Notation help file,
the AVE for a generic construct is:
$$ AVE = \frac{\text{Sum indicator true score variances}}{\text{Sum indicator variances}} = \frac{\sum Var(\eta_k)}{\sum Var(x_k)} = \frac{\sum\lambda^2_k}{\sum(\lambda^2_k + Var(\varepsilon_k))}$$
If $x_k$ is standardized (i.e., $Var(x_k) = 1$) the denominator reduces to $K$
and the AVE for a generic construct is:
$$ AVE = \frac{1}{K}\sum \lambda^2_k = \frac{1}{K}\sum \rho_{x_k, \eta}^2$$
As an important consequence, the AVE is closely tied to the communality.
Communality ($COM_k$) is definied as the proportion of variation in an indicator
that is explained by its common factor. Empirically, it is the square of
the standardized loading of the $k$'th indicator ($\lambda^2_k$). Since indicators,
scores/proxies and subsequently loadings are always standardized in cSEM,
the squared loading is simply the squared correlation between the indicator and
its related construct/common factor.
The AVE is also directly related to the indicator reliability,
defined as the squared correlation between an indicator $k$ and its related
proxy true score (see section Reliability below),
which is again simply $\lambda^2_k$. Therefore in cSEM we always have:
$$ AVE = \frac{1}{K}\sum COM_k = \frac{1}{K}\sum \text{Indicator reliability}_k = \frac{1}{K}\sum\lambda^2_k = \frac{1}{K}\sum R^2_k $$
Implementation
The function is implemented as: calculateAVE().
See also
The AVE is the basis for the Fornell-Larcker criterion.
Degrees of freedom{#df}
Definition
Degrees of freedom are calculated as the difference between the number
of non-redundant free elements of the empirical indicator correlation matrix $\bm{S}$
and the model parameters.
Although, composite-based estimators retrieve parameters of the
postulated models by forming composites, which involves the estimation of weights
the computation of the degrees of freedom eventually depends on the postulated
model and the parameters implied by the model.
Most notably, a common factor model estimated by a composite-based
approach such as PLS has the same degrees of freedom compared to e.g., classical
maximum likelihood estimation of the same model.
Formulae
$$
\begin{align}
\text{df} &= \text{# non-redundant off-diagonal elements of the empirical indicator correlation matrix $\bm{S}$} \
&- \text{# model parameters}
\end{align}
$$
If the model contains only linear terms the model parameters are:
- # free correlations between exogenous constructs
- # specified correlations between endogenous constructs
- # structural parameters
In addition, for each construct $\eta_j$:
- # of loadings if $\eta_j$ is modeled as a common factor
- # of specified measurement error correlations between items of constructs
$\eta_j$ if $\eta_j$ is modeled as a common factor
- # of weights of $\eta_j$ minus 1 if $\eta_j$ is modeled as a composite. One
weight per block is fixed and hence not counted as a model parameter since
the variance of the composite is scaled to be unity.
- # of non-redundant off-diagonal elements of $\bm{\Sigma}_j$ if $\eta_j$ is modeled as
a composite.
If the model contains second-order terms the model parameters are similar:
- # free correlations between exogenous constructs
- # specified correlations between endogenous constructs
- # structural parameters. Note: relations between constructs measuring/forming
the second-order construct are not path!
In addition, for each construct $\eta_j$ (including the second-order constructs):
- # of loadings if $\eta_j$ is modeled as a common factor
- # of specified measurement error correlations between items of constructs
$\eta_j$ if $\eta_j$ is modeled as a common factor
- # of weights of $\eta_j$ minus 1 if $\eta_j$ is modeled as a composite. One
weight per block is fixed and hence not counted as a model parameter since
the variance of the composite is scaled to be unity.
- # of non-redundant off-diagonal elements of $\bm{\Sigma}_j$ if $\eta_j$ is modeled as
a composite.
Notes
- If all constructs are allowed to freely covary, i.e., there is no
structural model and no structural parameters, all constructs are considered
exogenous.
- If the strucutral model contains nonlinear terms (e.g., $\eta^2_1$ or $\eta_1\eta_2$),
degrees of freedom computation is currently unclear (at least to us).
A warning is printed to inform the user that the calculation is may not be
correct.
Implementation
The function is implemented as calculateDf().
See also:
Degrees of freedom are required for several fit measures.
Fit Indices {#fit_indices}
Definition
Fit indices for confirmatory factor analysis (CFA) were first introduced by @Bentler1980.
Since then a large number of indices has been defined. Contrary to exact tests
of model fit, the purpose of fit indices is to measure the fit of a structural
equation model on a continuous scale. For normed fit indices this scale is
between 0 and 1. Fit indices can be divided into two classes:
- 'badness of fit' (resp. 'lack of fit') indices; a smaller value indicates a better fit.
- 'goodness of fit' indices; a higher value represents a better fit.
Several studies have analyzed the empirical and theoretical properties of fit
indices in the context of CFA where concepts are expressed by latent variables.
only little is known about the properties and the performance of fit indices
in composite models and for models estimated using a composite-based approach.
cSEM offers a number of fit indices that are known from factor-based SEM.
However, applied users should be aware that only little is known about their applicability,
intuition, and interpretability in the context of models containing constructs
modeled as composites or for models estimated using a composite-based approach.
Independent of the approach and model used, a particularily controversial issue
are cutoff values for fit indices [e.g., @Marsh2004]. In factor-based SEM cutoff
values are rather popular. The basis for these are numerous simulation studies, most notably @Hu1999.
In contrast for composite models - for better or worse - no cutoff values have been suggested.[^1]
Using assess() to calculate fit indices, the user should always keep in mind
that the value of a fit index is just some indication of good or bad fit.
Other aspects related to model fit must be considered as well.
It is unreasonable to make a binary decision about rejection or non-rejection
of a model by soley comparing the value of a fit index with a (more or less)
arbitrary cutoff value.
The definitions of fit indices calculated by assess() are given in the following:
- The $\chi^2$-statistic is the value of the fitting function times the
sample size minus 1.
- The $\chi^2/df$-ratio is the $\chi^2$-statistic divided by its
degrees of freedom.
- The goodness-of-fit index (GFI) measures the relative increase in fit of
the specified model compared to no model at all.
- The standardized root mean square residual (SRMR) is the square root of
the mean of squared residual correlations.
- The root mean square error of approximation (RMSEA) is the square root
of the discrepancy due to approximation per degree of freedom.
- The normed fit index (NFI) measures the increase in fit when specifying
the model under consideration relative to the fit of a certain baseline model
called the "null model".
- The non-normed fit index (NNFI) accounts for the degrees of freedom
of the involved models. It is the ratio of the distance between the fit of
the baseline model and the fit of the specified model (each per degree of freedom)
and the distance beetween the fit of the baseline model and the expected fit of the specified model (each per degree of freedom).
- The comparative fit index (CFI) estimates the relative decrease in
non-centrality when specifying the model under consideration instead of the baseline model.
- The incremental fit index (IFI) is the ratio of the distance between the
fit of the baseline model and the fit of the specified model and the distance
between the fit of the baseline model and the expected fit of the specified model.
Its definition differs only marginally from the definition of the NNFI.
- The root mean square outer residual covariance ($\text{RMS}_{\theta}$) is
defined as the square root of the mean squared covariances of the residuals of
the outer model. The calculation of the indicator's residual covariance matrix
involves the calculation of the construct's covariance matrix. See @Lohmoeller1989.
It should be stressed again that (with the possible exception of the $\text{RMS}_{\theta}$)
none of the above mentioned fit indices were originally designed for composite models.
The indices RMSEA and CFI are non-centrality
based and require specific assumptions on model and data typically made in CFA.
The same applies for IFI and NNFI since their calculation relies on the properties
(primarily the expectation) of the test statistic when data follows a normal
distribution. In general, those assumptions are not made in composite models
and composite-based estimators, respectively. For this reason, the intuition
behind these indices does not hold for composite-based SEM. Nevertheless,
calculation of these indices is also possible in this case. Whether the values
of these indices are still meaningful in a sense that they can be used for assessment of
model fit is an open question. Furthermore, values of fit indices for
composite-based estimators and factor-based estimators may not be compared.
Users should always keep this aspect and the general limitations of fit indices
in mind.
Formulae
The exact formulae of the fit indices as implemented in cSEM are given in
the following. The term $F = F(\bm{S}, \bm{\Sigma}(\hat{\bm{\theta}})) = F(\bm{S}, \hat{\bm{\Sigma}})$
stands for the value of the maximum likelihood fitting function evaluated at $\bm{S}$
(the empirical covariance matrix of the indicators) and $\hat{\bm{\Sigma}}$
(the estimated model-implied covariance matrix of the indicators). The value of
the maximum likelihood fitting function is computed by calculateDML().
The $\chi^2$-statistic
The $\chi^2$-statistic is defined as:
$$ \chi^2 = (N-1)\cdot F$$
where $N$ is the sample size.
Main reference: @Joereskog1969
The $\chi^2/\text{df}$-ratio
The $\chi^2/\text{df}$-statistic is defined as:
$$ \chi^2 = (N-1)\cdot F/\text{df}_M$$
where $N$ the sample size and $\text{df}_M$ the degrees of freedom of the estimated model.
Main reference: @Joereskog1969
The goodness-of-fit index (GFI)
The GFI is generally defined in analogy to the coefficient of determination ($R^2$) known
from regression analysis as 1 minus the share of the weighted unexplained variance (SSE; the difference between
$\bm{S}$ and $\hat{\bm{\Sigma}}$) relative
to the weighted total variance (SST; the variance of $\bm{S}$):
$$ GFI = 1 - \frac{\text{trace}\left{\left(\bm{W}^{-\frac{1}{2}}\lbrack\bm{S} - \hat{\bm{\Sigma}}\rbrack\bm{W}^{-\frac{1}{2}}\right)^2\right}}{\text{trace}\left{\left(\bm{W}^{-\frac{1}{2}}\bm{S}\bm{W}^{-\frac{1}{2}}\right)^2\right}} $$
The matrix $\bm{W}$ is a weight matrix. Depending on the estimation technique used
to obtain $\hat{\bm\theta}$ different types of GFI may be computed by choosing
a particular weight.
- If $\bm{W} = \hat{\bm{\Sigma}}$, the GFI is based on the SSE and the SST from a maximum
likelihood estimation.
- If $\bm{W} = \hat{\bm{S}}$, the GFI is based on SSE and the SST from a generalized least
squares (GLS) estimation.
- If $\bm{W} = \hat{\bm{I}}$, the GFI is based on SSE and the SST from a unweighted least
squares (ULS) estimation.
Note that for any quadratic matrix \bm{X}, we have: $\text{trace}(\bm{X}^2) = \sum_{i,j} x^2_i$.
Main references: @Joereskog1982, @Mulaik1989 and @Tanaka1985
The standardized root mean square residual (SRMR)
The SRMR is defined as
$$ \text{SRMR} = \sqrt{2 \sum_{j=1}^{K} \sum_{i=1}^{j} \frac{ \lbrack (s_{ij} - \hat{\sigma}{ij})/(s{ii} s_{jj})^{1/2} \rbrack^{2}}{K (K+1)}} $$
where $K$ stands for the number of indicators, $s_{ij}$ for the empirical covariance
between indicators $i$ and $j$, and $\hat{\sigma}{ij}$ for the estimated model-implied
counterpart. The SRMR describes with which distance the observed correlations
are reproduced on average by the model. Therefore, smaller values are associated
with a better fit. If data is standardized, $s{ii} = s_{jj} = 1$ holds,
and the formula reduces to:
$$ \text{SRMR} = \sqrt{2 \sum_{j=1}^{K} \sum_{i=1}^{j} \frac{(s_{ij} - \hat{\sigma}_{ij})^2}{K(K+1)}} $$
Main reference: @Bentler2006
The root mean square error of approximation (RMSEA)
The RMSEA is defined as
$$ \hat{\epsilon} = \sqrt{\frac{\hat{F}0}{\text{df}{M}}} \quad \text{where} \quad \hat{F}{0} = \max \Bigl( 0, F - \frac{\text{df}{M}}{N-1} \Bigr) $$
In this formula, $\text{df}{M}$ stands for the degrees of freedom of the specified model
(see the Degrees of Freedom section for details on how the degrees of freedom are calculated).
The term $\hat{F}{0}$ is an estimator for the discrepancy due to approximation.
Thus, the RMSEA measures the discrepancy due to approximation per degree of freedom.
Main reference: @Browne1992
The normed and non-normed fit index (NFI and NNFI)
The fit indices NFI and NNFI were among the first fit indices to be introduced [@Bentler1980].
They are defined as:
$$ \text{NFI} = \frac{F_{B} - F_{M}}{F_{B}} \quad \text{and} \quad \text{NNFI} = \frac{F_{B}/\text{df}{B} - F{M}/\text{df}{M}}{F{B}/\text{df}{B} - 1/(N-1)} $$
The term $F{B}$ refers to the value of the fitting function in the null model,
$F_{M}$ to the value of the fitting function in the model under consideration.
Thus, the NFI measures the increase in fit relative to the fit of the null model
when specifying the model. The intuition of NNFI is that (in factor-based methods)
the expectation of $F_{M}/\text{df}_{M}$ is equal to $1/N-1$. This does not automatically
hold for composite-based estimators.
The NNFI measures the relative departure of the numerator's
term from it's expectation (in the denominator). That is why, the NNFI is not
normed and can take values larger than $1$.
Main reference: @Bentler1980
The comparative fit index (CFI)
The CFI is defined as:
$$ \text{CFI} = 1 - \frac{\max(0, (N-1) F_{M}-\text{df}{M})}{\max(0, (N-1) F{M}-\text{df}{M}, (N-1)F{B}-\text{df}_{B})} $$
Like the RMSEA, the CFI is a non-centrality based index. It measures the
increase in fit (that is to say the reduction in non-centrality) when
specifying the model under consideration relative to the fit of the null model.
The CFI is a normed index with a value of $1$ indicating the best fit.
Since it makes use of the assumptions in factor-based methods, its intuition
does not apply to composite-based estimators.
Main reference: @Bentler1990.
The incremental fit index (IFI)
The IFI is defined as:
$$ \text{IFI} = \frac{F_{B} - F_{M}}{F_{B} - df_{M}/(N-1)} $$
The rationale underlying the IFI is that the term $F_{B} - F_{M}$ (in the numerator)
is compared with its expectation $F_{B} - \text{df}_{M}/(N-1)$ (in the denominator).
Main reference: @Bollen1989
The root mean square outer residual covariance
Implementation
The functions are implemented as: calculateChiSquare(), calculateChiSquareDf(),
calculateCFI(), calculateNFI(),
calculateNNFI(), calculateIFI(), calculateGFI(), calculateRMSEA(),
calculateRMSTheta(), calculateSRMR().
See also
Several fit indices require a fitting function, i.e., a distance
measure like the geodesic distance, the squared Euclidean distance or the maximum
likelihood distance. These are implemented as: calculateDG(), calculateDL(),
and calculateDML().
Reliability {#reliability}
Definition
Reliability is the consistency of measurement, i.e., the degree to which a hypothetical
repetition of the same measure would yield the same results. As such, reliability
is the closeness of a measure to an error free measure. It is not to be confused
with validity as a perfectly reliable measure may be invalid.
Practically, reliability must be empirically assessed based on a theoretical
framework. The dominant theoretical framework against which
to compare empirical reliability results to is the well-known true score framework
which provides the foundation for the measurement model described
in the Notation help file.
Based on the true score framework and using the terminology and notation of
the Notation and Termniology help files, reliability
of a generic measurement is defined as:
- The amount of proxy true score variance, $Var(\bar\eta)$, relative to the
the proxy or test score variance, $Var(\hat\eta)$.
- This is identical to the squared correlation between the common
factor and its proxy/composite or test score: $\rho_{\eta, \hat\eta}^2 = Cor(\eta, \hat\eta)^2$.
This "kind" of reliability is commonly referred to as
internal consistency reliability.
Based on the true score theory three major types of measurement models are
distinguished. Each type implies different assumptions which give rise
to the formulae written below. The well-established names for the different
types of measurement model provide natural naming candidates for their corresponding
(internal consistency) reliabilities measure:
- Parallel -- Assumption: $\eta_{kj} = \eta_j \longrightarrow \lambda_{kj} = \lambda_j$ and $Var(\varepsilon_{kj}) = Var(\varepsilon_j)$.
- Tau-equivalent -- Assumption: $\eta_{kj} = \eta_j \longrightarrow \lambda_{kj} = \lambda_j$ and $Var(\varepsilon_{kj}) \neq Var(\varepsilon_{lj})$.
- Congeneric -- Assumption: $\eta_{kj} = \lambda_{kj}\eta_j$ and $Var(\varepsilon_{kj}) \neq Var(\varepsilon_{lj})$.
In principal the test score $\hat\eta$ is a weighted linear combinations of
the indicators, i.e., a proxy or stand-in for the true score/common factor.
Historically, however, the test score is generally assumed to be a simple sum
score, i.e., a weighted sum of indicators with all weights assumed to be equal to
one. Hence, well-known reliability measures such as Jöreskog's $\rho$ or Cronbach's $\alpha$ are defined
with respect to a test score that indeed represents a simple sum score.
Yet, all reliability measures originally developed assuming a sum score may
equally well be computed with respect to a composite, i.e., a weighted score
with weights not necessarily equal to one.
Apart form the distinction between congeneric (i.e., Jöreskog's $\rho$) and tau-equivalent reliability
(i.e., Cronbach's $\alpha$) we therefore distinguish between reliability estimates based on
a test score (composite) that uses the weights of the weight approach used
to obtain .object and a test score (proxy) based on unit weights. The former
is indicated by adding "weighted" to the original name.
Formulae
The most general formula for reliability is the (weighted) congeneric reliability:
$$ \rho_{C; \text{weighted}} = \frac{Var(\bar\eta)}{Var(\hat\eta_k)} = \frac{(\bm{w}'\bm{\lambda})^2}{\bm{w}'\bm{\Sigma}\bm{w}}$$
Assuming $\bm w = \bm\iota$, i.e., unit weights, the "classical" formula for congeneric reliability
(i.e., Jöreskog's $\rho$), follows:
$$ \rho_C = \frac{Var(\bar\eta)}{Var(\hat\eta_k)} = \frac{\left(\sum\lambda_k\right)^2}{\left(\sum\lambda_k\right)^2 + Var(\bar\varepsilon)}$$
Using the assumptions imposed by the tau-equivalent measurement model we obtain
the (weighted) tau-equivalent reliability, i.e., (weighted) Cronbach's alpha):
$$ \rho_{T; \text{weighted}} = \frac{\lambda^2(\sum w_k)^2}{\lambda^2(\sum w_k)^2 + \sum w_k^2Var(\varepsilon_k)}
= \frac{\bar\sigma_x(\sum w_k)^2}{\bar\sigma_x[(\sum w_k)^2 - \sum w_k^2] + \sum w_k^2Var(x_k)}$$
where we used the fact that if $\lambda_k = \lambda$ (tau-equivalence),
$\lambda^2$ equals the average covariance between indicators:
$$\bar\sigma_x = \frac{1}{K(K-1)}\sum^K_{k=1}\sum^K_{l=1} \sigma_{kl}$$
Again, assuming $w_k = 1$, i.e., unit weights, the "classical" formula for
tau-equivalent reliability (Cronbach's $\alpha$) follows:
$$ \rho_T = \frac{\lambda^2K^2}{\lambda^2K^2 + \sum Var(\bar\varepsilon_k)}
= \frac{\bar\sigma_xK^2}{\bar\sigma_x[K^2 - K] + K Var(x_k)}$$
Using the assumptions imposed by the parallel measurement model we obtain
the parallel reliability:
$$ \rho_P = \frac{\lambda^2(\sum w_k)^2}{\lambda^2(\sum w_k)^2 + Var(\varepsilon)\sum w_k^2} =
\frac{\bar\sigma_x(\sum w_k)^2}{\bar\sigma_x[(\sum w_k)^2 - \sum w_k^2] + Var(x)\sum w_k^2} $$
In cSEM indicators are always standardized and weights are chosen such
that $Var(\hat\eta_k) = 1$. This is done by scaling the weight vector $\bm{w}$
by $(\bm{w}'\bm{\Sigma}\bm{w})^{-\frac{1}{2}}$. This simplifies the
formulae:
$$
\begin{align}
\rho_{C; \text{weighted}} &= (\sum w_k\lambda_k)^2 = (\bm{w}'\bm{\lambda})^2 \
\rho_{T; \text{weighted}} = \rho_{P; \text{weighted}} &= \bar\rho_x(\sum w_k)^2 \
\end{align}
$$
where $\bar\rho_x = \bar\sigma_x$ is the average correlation between indicators. Consequently,
parallel and tau-equivalent reliability are always identical in cSEM.
So far formulae have been motivated theoretically. Since $\bm{\Sigma}$ is unknown
it can be replaced by $\bm{S}$ (the empirical indicator correlation matrix) or
$\hat{\bm{\Sigma}}$ (the model-implied indicator correlation matrix), however,
$\bm{S}$ and $\hat{\bm{\Sigma}}$ are generally not equal. The practical implication
is that if $\rho_{C}$ is computed as $(\bm{w}'\bm{\lambda})^2$ using unit weights
the weights can in fact be scaled by both $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$
or $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$! Similarly, $\rho_{C; \text{weighted}}$
can be computed using weights scaled using either $\bm{S}$ or $\hat{\bm{\Sigma}}$.
Consequently there are in fact four types of congeneric reliability depending the type
of weight and the type of scaling for the weights. Hence, the calculation is
of "the" congeneric reliability is always:
$$(\bm{w}'\bm{\lambda})^2$$
where $\bm{w}$ can be:
- a vector of unit weights scaled by $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$. This
is typically what people refer to as the congeneric reliability (Jöreskog's $\rho$).
We label this type of reliability estimate $\rho_C$.
- a vector of unit weights scaled by $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$.
This has no known name. Its usefulness is an open question. We label this type
of reliability estimate $\rho_{C;mm}$.
- a vector of weights obtained using a composite-based estimator (e.g. PLS-PM)
scaled by $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$. This is
Dijkstra Henseler's $\rho_A$. We label this type of reliability estimate $\rho_{C;\text{weighted}}$.
- a vector of weights obtained using a composite-based estimator (e.g. PLS-PM)
scaled by $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$.
This has no known name. Its usefulness is an open question.
We label this type of reliability estimate $\rho_{C;\text{weighted};mm}$
A note on the terminology
A vast bulk of literature dating back to seminal work by Spearman (e.g., Spearman (1904))
has been written on the subject of reliability. Inevitably, definitions, formulae,
notation and terminology conventions are unsystematic and confusing. This is
particularly true for newcomers to structural equation modeling
or applied users whose primary concern is to apply the appropriate method
to the appropriate case without poring over books and research papers
to understand each intricate detail.
In cSEM we seek to make working with reliabilities as consistent as
possible by relying on a paper by @Cho2016 who proposed uniform
formula-generating methods and a systematic naming conventions for all
common reliability measures. Naturally, some of the conventional terminology
is deeply entrenched within the nomenclatura of a particular filed (e.g., coefficient
alpha alias Cronbach's alpha in pychonometrics) such that a new, albeit consistent,
naming scheme seems superfluous at best. However, we belief the merit of a
"standardized" naming pattern will eventually be helpful to all users as it
helps clarify potential misconceptions thus preventing potential misuse, such as
the (ab)use of Cronbach alpha as a reliability measure for congernic measurement models.
Apart from these considerations, this package takes a pragmatic stance
in a sense that we use consistent naming because it naturally provides
a consistent naming scheme for the functions and the systematic formula generating
methods because they make code maintenance easier. Eventually, what matters is
the formula and more so its correct application. To facilitate the translation between
different naming systems and conventions we provide a "translation table" below:
Systematic names Mathematical Synonymous terms
Parallel reliability $\rho_P$ Spearman-Brown formula, Spearman-Brown prophecy,
Standardized alpha, Split-half reliability
Tau-equivalent reliability $\rho_T$ Cronbach's alpha, $\alpha$, Coefficient alpha
Guttmans $\lambda_3$, KR-20
Tau-equivalent reliability $\rho_{T;\text{weighted}}$ --
weighted
Congeneric reliability $\rho_C$ Composite reliability, Jöreskog's $\rho$,
Construct reliability, $\omega$,
reliability coefficient, Dillon-Goldsteins's $\rho$
Congeneric reliability $\rho_{C;\text{weighted}}$ Dijkstra-Henseler's $\rho_A$
weighted
Table: Systematic names and common synonymous names for the
reliability estimates found in the literature
Closed-form confidence interval
@Trinchera2018 proposed a closed-form confidence interval (CI) for
the tau-equivalent reliability (Cronbach's alpha). To compute the CI, set
.closed_form_ci = TRUE when calling assess() or invoke
calculateRhoT(..., .closed_form_ci = TRUE) directly. The level of the CI
can be changed by supplying a single value or a vector of values to .alpha.
Implementation
The functions are implemented as calculateRhoC() and calculateRhoT().
The Goodness of Fit (GoF)
Definition
Calculate the Goodness of Fit (GoF) proposed by @Tenenhaus2004.
Note that, contrary to what the name suggests, the GoF is not a
measure of (overall) model fit in a $\chi^2$-fit test sense.
See e.g. @Henseler2012a for a discussion.
Formulae
The GoF is defined as:
$$\text{GoF} = \sqrt{\varnothing \text{COM}k \times \varnothing R^2{structural}} =
\sqrt{\frac{1}{k}\sum^K_{k=1} \lambda^2_k + \frac{1}{M} \sum^M_{m = 1} R^2_{m;structural}} $$
where $COM_k$ is the communality of indicator $k$, i.e. the variance in the indicator
that is explained by its connected latent variable and $R^2_{m; structural}$ the
R squared of the $m$'th equation of the structural model.
Implementation
The function is implemented as: calculateGoF().
The Heterotrait-Monotrait-Ratio of Correlations (HTMT)
Definition
The heterotrait-monotrait ratio of correlations (HTMT) was first proposed by
@Henseler2015 to assess convergent and discriminant validity.
Formulae
See: @Henseler2015 on page 121 (equation (6))
Implementation
The function is implemented as: calculateHTMT().
Literature
[^1]: There are some cutoffs such as e.g., the SRMR should be less than
0.08 or 0.1, however, these values are essentially arbitrary as they have
never been formally motivated. Reference is usually done to @Hu1999 which
based the cut-off on a simulation using factor-based SEM.
Try the cSEM package in your browser
Any scripts or data that you put into this service are public.
cSEM documentation built on Nov. 25, 2022, 1:05 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Introduction
As indicated by the name, assess() is used to assess a model estimated using the
csem() function.
In cSEM model assessment is considered to be any task that in some way or
another seeks to assess the quality of the estimated model without conducting
a statistical test (tests are covered by the test_* family of functions).
Quality in this case is taken to be a catch-all term for
all common aspects of model assessment. This mainly comprises
fit indices, model selection criteria, reliability estimates, common validity
assessment criteria, effect sizes, and other
related quality measures/indices that do not rely on a formal test procedure.
Hereinafter, we will refer to a generic (fit) index, quality or assessment measure as
a quality criterion.
Currently the following quality criteria are implemented:
- Convergent and discriminant validity assessment:
- The average variance extracted (AVE)
- The Fornell-Larcker criterion
- The heterotrait-monotrait ratio of correlations (HTMT)
- Congeneric reliability ($\rho_C$), also known as e.g.: composite reliability, construct reliability, (unidimensional) omega, Jöreskog's $\rho$, $\rho_A$, or $\rho_B$.
- Tau-equivalent reliability ($\rho_T$), also known as e.g.: Cronbach alpha, alpha, $\alpha$, coefficient alpha, Guttman's $\lambda_3$, KR-20.
- Distance measures
- The standardized root mean square residual (SRMR)
- The geodesic distance (DG)
- The squared Euclidian distance (DL)
- The maximum-likelihood distance (DML)
- Fit indices
- The $\chi^2$-statistic
- The $\chi^2/df$-statistic
- The comparative fit index (CFI)
- The goodness-of-fit index (GFI)
- The standardized root mean square residual (SRMR)
- The root mean square error of approximation (RMSEA)
- The normed fit index (NFI)
- The non-normed fit index (NNFI)
- The comparative fit index (CFI)
- The incremental fit index (IFI)
- The root mean square outer residual covariance ($\text{RMS}_{\theta}$)
- The Goodness-of-Fit (GoF) proposed by @Tenenhaus2004.
- The variance inflation factors (VIF) for the structural equations as
well as for Mode B regression equations (if
.approach_weights = "PLS-PM"). - The coefficient of determination and the adjusted coefficient of determination ($R^2$ and $R^2_{adj}$)
- A measure of the effect size (Cohen's $f^2$).
- Direct, indirect and total effect assessment.
- Several model selection criteria as described in @Sharma2019.
For implementation details see the Methods & Formulae section.
Syntax & Options
assess( .object = NULL, .only_common_factors = TRUE, .quality_criterion = c("all", "aic", "aicc", "aicu", "bic", "fpe", "gm", "hq", "hqc", "mallows_cp", "ave", "rho_C", "rho_C_mm", "rho_C_weighted", "rho_C_weighted_mm", "dg", "dl", "dml", "df", "effects", "f2", "fl_criterion", "chi_square", "chi_square_df", "cfi", "gfi", "ifi", "nfi", "nnfi", "reliability", "rmsea", "rms_theta", "srmr", "gof", "htmt", "r2", "r2_adj", "rho_T", "rho_T_weighted", "vif", "vifmodeB"), ... )
.object
: An object of class cSEMResults resulting from a call to csem().
.quality_criterion
: A character string or a vector of character strings naming the quality criterion
to compute. By default all quality criteria are computed ("all").
See assess() for a list of possible candidates.
.only_common_factors
: Logical. Should only concepts modeled as common
factors be included when calculating one of the following quality criteria:
AVE, the Fornell-Larcker criterion, HTMT, and all reliability estimates.
Defaults to TRUE.
...
: Further arguments passed to functions called by assess().
See args_assess_dotdotdot
for a complete list of available arguments.
Like all postestimation functions assess() can be called on any object of
class cSEMResults. The output is a named list of the quality criteria given
to .quality_criterion. By default all possible quality criteria
are calculated (.quality_criterion = "all").
Details {#details}
In line with all of cSEM's postestimation functions, assess() is a generic
function with methods for objects of class cSEMResults_default,
cSEMResults_multi, cSEMResults_2ndorder. In cSEM
every cSEMResults_* object must also have class cSEMResults for internal reasons.
When using one of the major postestimation functions, method dispatch is therefore technically
done on one of the cSEMResults_* class attributes, ignoring the
cSEMResults class attribute. As long as assess() is
used directly, method dispatch is not of any practical concern to the end-users.
Composite models vs. common factor models
Some assessment measures are inherently tied to the common factor model. It is therefore unclear how to interpret their results in the context of a composite model. Consequently, their computation is suppressed by default for constructs modeled as composites. Currently, this applies to the following quality criteria:
- AVE and validity assessment based thereon (i.e., the Fornell-Larcker criterion)
- HTMT and validity assessment based thereon
- All reliability measures
It is possible to force computation of all quality criteria for constructs
modeled as composites using .only_common_factors = FALSE, however, we explicitly
warn to interpret results, as they may not even have a conceptual meaning.
All quality criteria assume that the estimated loadings, construct correlations
and path coefficients involved in the computation of a specific quality measure
are consistent estimates for their theoretical population counterpart. If
the user deliberately chooses an approach that yields inconsistent estimates (by
setting .disattenuate = FALSE in csem() when the estimated model contains constructs
modeled as common factors) assess() will still estimate
all quantities, however, quantities such as
the AVE or the congeneric reliability $\rho_C$ inherit inconsistency.
Methods & Formulae {#methods}
This section provides technical details and relevant formulae. For the relevant notation and terminology used in this section, see the Notation and the Termionology help files.
Average Variance Extracted (AVE) {#ave}
Definition
The average variance extracted (AVE) was first proposed by @Fornell1981. Several definitions exist. For ease of comparison to extant literature the most common definitions are given below:
- The AVE for a generic construct/latent variable $\eta$ is an estimate of how much of the variation of its indicators is due to the assumed latent variable. Consequently, the share of unexplained, i.e. error variation is 1 - AVE.
- The AVE for a generic construct/latent variable $\eta$ is the share of the total indicator variance (i.e., the sum of the indicator variances of all indicators connected to the construct), that is captured by the (indicator) true scores.
- The AVE for a generic construct/latent variable $\eta$ is the ratio of the sum of the (indicator) true score variances (explained variation) relative to the sum of the total indicator variances (total variation, i.e., the sum of the indicator variances of all indicators connected to the construct).
- Since for the regression of $x_k$ on $\eta_k$, the R squared ($R^2_k)$ is equal to the share of variation of $x_k$ explained by $\eta_k$ relative to the total variation of $x_k$, the AVE for a generic construct/latent variable $\eta$ is equal to the average over all $R^2_k$.
- The AVE for a generic construct/latent variable $\eta$ is the sum of the squared correlation between indicator $x_k$ and the (indicator) true score $\eta_k$ relative to the sum of the indicator variances of all indicators connected to the construct in question.
It is important to stress that, although different in wording, all definitions are synonymous!
The AVE is inherently tied to the common factor model. It
is therefore unclear how to interpret the AVE for constructs modeled as composites.
Consequently, the computation is suppressed by default for
constructs modeled as common factors. It is possible to force computation of the
AVE for constructs modeled as composites using .only_common_factors = FALSE,
however, we explicitly warn to interpret results, as they may not even have a
conceptual meaning.
Formulae
Using the results and notation derived and defined in the Notation help file, the AVE for a generic construct is: $$ AVE = \frac{\text{Sum indicator true score variances}}{\text{Sum indicator variances}} = \frac{\sum Var(\eta_k)}{\sum Var(x_k)} = \frac{\sum\lambda^2_k}{\sum(\lambda^2_k + Var(\varepsilon_k))}$$ If $x_k$ is standardized (i.e., $Var(x_k) = 1$) the denominator reduces to $K$ and the AVE for a generic construct is: $$ AVE = \frac{1}{K}\sum \lambda^2_k = \frac{1}{K}\sum \rho_{x_k, \eta}^2$$ As an important consequence, the AVE is closely tied to the communality. Communality ($COM_k$) is definied as the proportion of variation in an indicator that is explained by its common factor. Empirically, it is the square of the standardized loading of the $k$'th indicator ($\lambda^2_k$). Since indicators, scores/proxies and subsequently loadings are always standardized in cSEM, the squared loading is simply the squared correlation between the indicator and its related construct/common factor. The AVE is also directly related to the indicator reliability, defined as the squared correlation between an indicator $k$ and its related proxy true score (see section Reliability below), which is again simply $\lambda^2_k$. Therefore in cSEM we always have:
$$ AVE = \frac{1}{K}\sum COM_k = \frac{1}{K}\sum \text{Indicator reliability}_k = \frac{1}{K}\sum\lambda^2_k = \frac{1}{K}\sum R^2_k $$
Implementation
The function is implemented as: calculateAVE().
See also
The AVE is the basis for the Fornell-Larcker criterion.
Degrees of freedom{#df}
Definition
Degrees of freedom are calculated as the difference between the number of non-redundant free elements of the empirical indicator correlation matrix $\bm{S}$ and the model parameters.
Although, composite-based estimators retrieve parameters of the postulated models by forming composites, which involves the estimation of weights the computation of the degrees of freedom eventually depends on the postulated model and the parameters implied by the model. Most notably, a common factor model estimated by a composite-based approach such as PLS has the same degrees of freedom compared to e.g., classical maximum likelihood estimation of the same model.
Formulae
$$ \begin{align} \text{df} &= \text{# non-redundant off-diagonal elements of the empirical indicator correlation matrix $\bm{S}$} \ &- \text{# model parameters} \end{align} $$
If the model contains only linear terms the model parameters are:
- # free correlations between exogenous constructs
- # specified correlations between endogenous constructs
- # structural parameters
In addition, for each construct $\eta_j$:
- # of loadings if $\eta_j$ is modeled as a common factor
- # of specified measurement error correlations between items of constructs $\eta_j$ if $\eta_j$ is modeled as a common factor
- # of weights of $\eta_j$ minus 1 if $\eta_j$ is modeled as a composite. One weight per block is fixed and hence not counted as a model parameter since the variance of the composite is scaled to be unity.
- # of non-redundant off-diagonal elements of $\bm{\Sigma}_j$ if $\eta_j$ is modeled as a composite.
If the model contains second-order terms the model parameters are similar:
- # free correlations between exogenous constructs
- # specified correlations between endogenous constructs
- # structural parameters. Note: relations between constructs measuring/forming the second-order construct are not path!
In addition, for each construct $\eta_j$ (including the second-order constructs):
- # of loadings if $\eta_j$ is modeled as a common factor
- # of specified measurement error correlations between items of constructs $\eta_j$ if $\eta_j$ is modeled as a common factor
- # of weights of $\eta_j$ minus 1 if $\eta_j$ is modeled as a composite. One weight per block is fixed and hence not counted as a model parameter since the variance of the composite is scaled to be unity.
- # of non-redundant off-diagonal elements of $\bm{\Sigma}_j$ if $\eta_j$ is modeled as a composite.
Notes
- If all constructs are allowed to freely covary, i.e., there is no structural model and no structural parameters, all constructs are considered exogenous.
- If the strucutral model contains nonlinear terms (e.g., $\eta^2_1$ or $\eta_1\eta_2$), degrees of freedom computation is currently unclear (at least to us). A warning is printed to inform the user that the calculation is may not be correct.
Implementation
The function is implemented as calculateDf().
See also:
Degrees of freedom are required for several fit measures.
Fit Indices {#fit_indices}
Definition
Fit indices for confirmatory factor analysis (CFA) were first introduced by @Bentler1980. Since then a large number of indices has been defined. Contrary to exact tests of model fit, the purpose of fit indices is to measure the fit of a structural equation model on a continuous scale. For normed fit indices this scale is between 0 and 1. Fit indices can be divided into two classes:
- 'badness of fit' (resp. 'lack of fit') indices; a smaller value indicates a better fit.
- 'goodness of fit' indices; a higher value represents a better fit.
Several studies have analyzed the empirical and theoretical properties of fit indices in the context of CFA where concepts are expressed by latent variables. only little is known about the properties and the performance of fit indices in composite models and for models estimated using a composite-based approach. cSEM offers a number of fit indices that are known from factor-based SEM. However, applied users should be aware that only little is known about their applicability, intuition, and interpretability in the context of models containing constructs modeled as composites or for models estimated using a composite-based approach.
Independent of the approach and model used, a particularily controversial issue
are cutoff values for fit indices [e.g., @Marsh2004]. In factor-based SEM cutoff
values are rather popular. The basis for these are numerous simulation studies, most notably @Hu1999.
In contrast for composite models - for better or worse - no cutoff values have been suggested.[^1]
Using assess() to calculate fit indices, the user should always keep in mind
that the value of a fit index is just some indication of good or bad fit.
Other aspects related to model fit must be considered as well.
It is unreasonable to make a binary decision about rejection or non-rejection
of a model by soley comparing the value of a fit index with a (more or less)
arbitrary cutoff value.
The definitions of fit indices calculated by assess() are given in the following:
- The $\chi^2$-statistic is the value of the fitting function times the sample size minus 1.
- The $\chi^2/df$-ratio is the $\chi^2$-statistic divided by its degrees of freedom.
- The goodness-of-fit index (GFI) measures the relative increase in fit of the specified model compared to no model at all.
- The standardized root mean square residual (SRMR) is the square root of the mean of squared residual correlations.
- The root mean square error of approximation (RMSEA) is the square root of the discrepancy due to approximation per degree of freedom.
- The normed fit index (NFI) measures the increase in fit when specifying the model under consideration relative to the fit of a certain baseline model called the "null model".
- The non-normed fit index (NNFI) accounts for the degrees of freedom of the involved models. It is the ratio of the distance between the fit of the baseline model and the fit of the specified model (each per degree of freedom) and the distance beetween the fit of the baseline model and the expected fit of the specified model (each per degree of freedom).
- The comparative fit index (CFI) estimates the relative decrease in non-centrality when specifying the model under consideration instead of the baseline model.
- The incremental fit index (IFI) is the ratio of the distance between the fit of the baseline model and the fit of the specified model and the distance between the fit of the baseline model and the expected fit of the specified model. Its definition differs only marginally from the definition of the NNFI.
- The root mean square outer residual covariance ($\text{RMS}_{\theta}$) is defined as the square root of the mean squared covariances of the residuals of the outer model. The calculation of the indicator's residual covariance matrix involves the calculation of the construct's covariance matrix. See @Lohmoeller1989.
It should be stressed again that (with the possible exception of the $\text{RMS}_{\theta}$) none of the above mentioned fit indices were originally designed for composite models. The indices RMSEA and CFI are non-centrality based and require specific assumptions on model and data typically made in CFA. The same applies for IFI and NNFI since their calculation relies on the properties (primarily the expectation) of the test statistic when data follows a normal distribution. In general, those assumptions are not made in composite models and composite-based estimators, respectively. For this reason, the intuition behind these indices does not hold for composite-based SEM. Nevertheless, calculation of these indices is also possible in this case. Whether the values of these indices are still meaningful in a sense that they can be used for assessment of model fit is an open question. Furthermore, values of fit indices for composite-based estimators and factor-based estimators may not be compared. Users should always keep this aspect and the general limitations of fit indices in mind.
Formulae
The exact formulae of the fit indices as implemented in cSEM are given in
the following. The term $F = F(\bm{S}, \bm{\Sigma}(\hat{\bm{\theta}})) = F(\bm{S}, \hat{\bm{\Sigma}})$
stands for the value of the maximum likelihood fitting function evaluated at $\bm{S}$
(the empirical covariance matrix of the indicators) and $\hat{\bm{\Sigma}}$
(the estimated model-implied covariance matrix of the indicators). The value of
the maximum likelihood fitting function is computed by calculateDML().
The $\chi^2$-statistic
The $\chi^2$-statistic is defined as: $$ \chi^2 = (N-1)\cdot F$$ where $N$ is the sample size.
Main reference: @Joereskog1969
The $\chi^2/\text{df}$-ratio
The $\chi^2/\text{df}$-statistic is defined as: $$ \chi^2 = (N-1)\cdot F/\text{df}_M$$ where $N$ the sample size and $\text{df}_M$ the degrees of freedom of the estimated model.
Main reference: @Joereskog1969
The goodness-of-fit index (GFI)
The GFI is generally defined in analogy to the coefficient of determination ($R^2$) known from regression analysis as 1 minus the share of the weighted unexplained variance (SSE; the difference between $\bm{S}$ and $\hat{\bm{\Sigma}}$) relative to the weighted total variance (SST; the variance of $\bm{S}$): $$ GFI = 1 - \frac{\text{trace}\left{\left(\bm{W}^{-\frac{1}{2}}\lbrack\bm{S} - \hat{\bm{\Sigma}}\rbrack\bm{W}^{-\frac{1}{2}}\right)^2\right}}{\text{trace}\left{\left(\bm{W}^{-\frac{1}{2}}\bm{S}\bm{W}^{-\frac{1}{2}}\right)^2\right}} $$ The matrix $\bm{W}$ is a weight matrix. Depending on the estimation technique used to obtain $\hat{\bm\theta}$ different types of GFI may be computed by choosing a particular weight.
- If $\bm{W} = \hat{\bm{\Sigma}}$, the GFI is based on the SSE and the SST from a maximum likelihood estimation.
- If $\bm{W} = \hat{\bm{S}}$, the GFI is based on SSE and the SST from a generalized least squares (GLS) estimation.
- If $\bm{W} = \hat{\bm{I}}$, the GFI is based on SSE and the SST from a unweighted least squares (ULS) estimation.
Note that for any quadratic matrix \bm{X}, we have: $\text{trace}(\bm{X}^2) = \sum_{i,j} x^2_i$.
Main references: @Joereskog1982, @Mulaik1989 and @Tanaka1985
The standardized root mean square residual (SRMR)
The SRMR is defined as $$ \text{SRMR} = \sqrt{2 \sum_{j=1}^{K} \sum_{i=1}^{j} \frac{ \lbrack (s_{ij} - \hat{\sigma}{ij})/(s{ii} s_{jj})^{1/2} \rbrack^{2}}{K (K+1)}} $$ where $K$ stands for the number of indicators, $s_{ij}$ for the empirical covariance between indicators $i$ and $j$, and $\hat{\sigma}{ij}$ for the estimated model-implied counterpart. The SRMR describes with which distance the observed correlations are reproduced on average by the model. Therefore, smaller values are associated with a better fit. If data is standardized, $s{ii} = s_{jj} = 1$ holds, and the formula reduces to: $$ \text{SRMR} = \sqrt{2 \sum_{j=1}^{K} \sum_{i=1}^{j} \frac{(s_{ij} - \hat{\sigma}_{ij})^2}{K(K+1)}} $$
Main reference: @Bentler2006
The root mean square error of approximation (RMSEA)
The RMSEA is defined as $$ \hat{\epsilon} = \sqrt{\frac{\hat{F}0}{\text{df}{M}}} \quad \text{where} \quad \hat{F}{0} = \max \Bigl( 0, F - \frac{\text{df}{M}}{N-1} \Bigr) $$ In this formula, $\text{df}{M}$ stands for the degrees of freedom of the specified model (see the Degrees of Freedom section for details on how the degrees of freedom are calculated). The term $\hat{F}{0}$ is an estimator for the discrepancy due to approximation. Thus, the RMSEA measures the discrepancy due to approximation per degree of freedom.
Main reference: @Browne1992
The normed and non-normed fit index (NFI and NNFI)
The fit indices NFI and NNFI were among the first fit indices to be introduced [@Bentler1980]. They are defined as: $$ \text{NFI} = \frac{F_{B} - F_{M}}{F_{B}} \quad \text{and} \quad \text{NNFI} = \frac{F_{B}/\text{df}{B} - F{M}/\text{df}{M}}{F{B}/\text{df}{B} - 1/(N-1)} $$ The term $F{B}$ refers to the value of the fitting function in the null model, $F_{M}$ to the value of the fitting function in the model under consideration. Thus, the NFI measures the increase in fit relative to the fit of the null model when specifying the model. The intuition of NNFI is that (in factor-based methods) the expectation of $F_{M}/\text{df}_{M}$ is equal to $1/N-1$. This does not automatically hold for composite-based estimators.
The NNFI measures the relative departure of the numerator's term from it's expectation (in the denominator). That is why, the NNFI is not normed and can take values larger than $1$.
Main reference: @Bentler1980
The comparative fit index (CFI)
The CFI is defined as: $$ \text{CFI} = 1 - \frac{\max(0, (N-1) F_{M}-\text{df}{M})}{\max(0, (N-1) F{M}-\text{df}{M}, (N-1)F{B}-\text{df}_{B})} $$ Like the RMSEA, the CFI is a non-centrality based index. It measures the increase in fit (that is to say the reduction in non-centrality) when specifying the model under consideration relative to the fit of the null model. The CFI is a normed index with a value of $1$ indicating the best fit. Since it makes use of the assumptions in factor-based methods, its intuition does not apply to composite-based estimators.
Main reference: @Bentler1990.
The incremental fit index (IFI)
The IFI is defined as: $$ \text{IFI} = \frac{F_{B} - F_{M}}{F_{B} - df_{M}/(N-1)} $$ The rationale underlying the IFI is that the term $F_{B} - F_{M}$ (in the numerator) is compared with its expectation $F_{B} - \text{df}_{M}/(N-1)$ (in the denominator).
Main reference: @Bollen1989
The root mean square outer residual covariance
Implementation
The functions are implemented as: calculateChiSquare(), calculateChiSquareDf(),
calculateCFI(), calculateNFI(),
calculateNNFI(), calculateIFI(), calculateGFI(), calculateRMSEA(),
calculateRMSTheta(), calculateSRMR().
See also
Several fit indices require a fitting function, i.e., a distance
measure like the geodesic distance, the squared Euclidean distance or the maximum
likelihood distance. These are implemented as: calculateDG(), calculateDL(),
and calculateDML().
Reliability {#reliability}
Definition
Reliability is the consistency of measurement, i.e., the degree to which a hypothetical repetition of the same measure would yield the same results. As such, reliability is the closeness of a measure to an error free measure. It is not to be confused with validity as a perfectly reliable measure may be invalid.
Practically, reliability must be empirically assessed based on a theoretical framework. The dominant theoretical framework against which to compare empirical reliability results to is the well-known true score framework which provides the foundation for the measurement model described in the Notation help file. Based on the true score framework and using the terminology and notation of the Notation and Termniology help files, reliability of a generic measurement is defined as:
- The amount of proxy true score variance, $Var(\bar\eta)$, relative to the the proxy or test score variance, $Var(\hat\eta)$.
- This is identical to the squared correlation between the common factor and its proxy/composite or test score: $\rho_{\eta, \hat\eta}^2 = Cor(\eta, \hat\eta)^2$.
This "kind" of reliability is commonly referred to as internal consistency reliability.
Based on the true score theory three major types of measurement models are distinguished. Each type implies different assumptions which give rise to the formulae written below. The well-established names for the different types of measurement model provide natural naming candidates for their corresponding (internal consistency) reliabilities measure:
- Parallel -- Assumption: $\eta_{kj} = \eta_j \longrightarrow \lambda_{kj} = \lambda_j$ and $Var(\varepsilon_{kj}) = Var(\varepsilon_j)$.
- Tau-equivalent -- Assumption: $\eta_{kj} = \eta_j \longrightarrow \lambda_{kj} = \lambda_j$ and $Var(\varepsilon_{kj}) \neq Var(\varepsilon_{lj})$.
- Congeneric -- Assumption: $\eta_{kj} = \lambda_{kj}\eta_j$ and $Var(\varepsilon_{kj}) \neq Var(\varepsilon_{lj})$.
In principal the test score $\hat\eta$ is a weighted linear combinations of the indicators, i.e., a proxy or stand-in for the true score/common factor. Historically, however, the test score is generally assumed to be a simple sum score, i.e., a weighted sum of indicators with all weights assumed to be equal to one. Hence, well-known reliability measures such as Jöreskog's $\rho$ or Cronbach's $\alpha$ are defined with respect to a test score that indeed represents a simple sum score. Yet, all reliability measures originally developed assuming a sum score may equally well be computed with respect to a composite, i.e., a weighted score with weights not necessarily equal to one.
Apart form the distinction between congeneric (i.e., Jöreskog's $\rho$) and tau-equivalent reliability
(i.e., Cronbach's $\alpha$) we therefore distinguish between reliability estimates based on
a test score (composite) that uses the weights of the weight approach used
to obtain .object and a test score (proxy) based on unit weights. The former
is indicated by adding "weighted" to the original name.
Formulae
The most general formula for reliability is the (weighted) congeneric reliability:
$$ \rho_{C; \text{weighted}} = \frac{Var(\bar\eta)}{Var(\hat\eta_k)} = \frac{(\bm{w}'\bm{\lambda})^2}{\bm{w}'\bm{\Sigma}\bm{w}}$$ Assuming $\bm w = \bm\iota$, i.e., unit weights, the "classical" formula for congeneric reliability (i.e., Jöreskog's $\rho$), follows: $$ \rho_C = \frac{Var(\bar\eta)}{Var(\hat\eta_k)} = \frac{\left(\sum\lambda_k\right)^2}{\left(\sum\lambda_k\right)^2 + Var(\bar\varepsilon)}$$ Using the assumptions imposed by the tau-equivalent measurement model we obtain the (weighted) tau-equivalent reliability, i.e., (weighted) Cronbach's alpha):
$$ \rho_{T; \text{weighted}} = \frac{\lambda^2(\sum w_k)^2}{\lambda^2(\sum w_k)^2 + \sum w_k^2Var(\varepsilon_k)} = \frac{\bar\sigma_x(\sum w_k)^2}{\bar\sigma_x[(\sum w_k)^2 - \sum w_k^2] + \sum w_k^2Var(x_k)}$$ where we used the fact that if $\lambda_k = \lambda$ (tau-equivalence), $\lambda^2$ equals the average covariance between indicators: $$\bar\sigma_x = \frac{1}{K(K-1)}\sum^K_{k=1}\sum^K_{l=1} \sigma_{kl}$$ Again, assuming $w_k = 1$, i.e., unit weights, the "classical" formula for tau-equivalent reliability (Cronbach's $\alpha$) follows: $$ \rho_T = \frac{\lambda^2K^2}{\lambda^2K^2 + \sum Var(\bar\varepsilon_k)} = \frac{\bar\sigma_xK^2}{\bar\sigma_x[K^2 - K] + K Var(x_k)}$$ Using the assumptions imposed by the parallel measurement model we obtain the parallel reliability:
$$ \rho_P = \frac{\lambda^2(\sum w_k)^2}{\lambda^2(\sum w_k)^2 + Var(\varepsilon)\sum w_k^2} = \frac{\bar\sigma_x(\sum w_k)^2}{\bar\sigma_x[(\sum w_k)^2 - \sum w_k^2] + Var(x)\sum w_k^2} $$
In cSEM indicators are always standardized and weights are chosen such that $Var(\hat\eta_k) = 1$. This is done by scaling the weight vector $\bm{w}$ by $(\bm{w}'\bm{\Sigma}\bm{w})^{-\frac{1}{2}}$. This simplifies the formulae: $$ \begin{align} \rho_{C; \text{weighted}} &= (\sum w_k\lambda_k)^2 = (\bm{w}'\bm{\lambda})^2 \ \rho_{T; \text{weighted}} = \rho_{P; \text{weighted}} &= \bar\rho_x(\sum w_k)^2 \ \end{align} $$ where $\bar\rho_x = \bar\sigma_x$ is the average correlation between indicators. Consequently, parallel and tau-equivalent reliability are always identical in cSEM.
So far formulae have been motivated theoretically. Since $\bm{\Sigma}$ is unknown it can be replaced by $\bm{S}$ (the empirical indicator correlation matrix) or $\hat{\bm{\Sigma}}$ (the model-implied indicator correlation matrix), however, $\bm{S}$ and $\hat{\bm{\Sigma}}$ are generally not equal. The practical implication is that if $\rho_{C}$ is computed as $(\bm{w}'\bm{\lambda})^2$ using unit weights the weights can in fact be scaled by both $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$ or $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$! Similarly, $\rho_{C; \text{weighted}}$ can be computed using weights scaled using either $\bm{S}$ or $\hat{\bm{\Sigma}}$. Consequently there are in fact four types of congeneric reliability depending the type of weight and the type of scaling for the weights. Hence, the calculation is of "the" congeneric reliability is always: $$(\bm{w}'\bm{\lambda})^2$$ where $\bm{w}$ can be:
- a vector of unit weights scaled by $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$. This is typically what people refer to as the congeneric reliability (Jöreskog's $\rho$). We label this type of reliability estimate $\rho_C$.
- a vector of unit weights scaled by $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$. This has no known name. Its usefulness is an open question. We label this type of reliability estimate $\rho_{C;mm}$.
- a vector of weights obtained using a composite-based estimator (e.g. PLS-PM) scaled by $(\bm{w}'\bm{S}\bm{w})^{-\frac{1}{2}}$. This is Dijkstra Henseler's $\rho_A$. We label this type of reliability estimate $\rho_{C;\text{weighted}}$.
- a vector of weights obtained using a composite-based estimator (e.g. PLS-PM) scaled by $(\bm{w}'\hat{\bm{\Sigma}}\bm{w})^{-\frac{1}{2}}$. This has no known name. Its usefulness is an open question. We label this type of reliability estimate $\rho_{C;\text{weighted};mm}$
A note on the terminology
A vast bulk of literature dating back to seminal work by Spearman (e.g., Spearman (1904)) has been written on the subject of reliability. Inevitably, definitions, formulae, notation and terminology conventions are unsystematic and confusing. This is particularly true for newcomers to structural equation modeling or applied users whose primary concern is to apply the appropriate method to the appropriate case without poring over books and research papers to understand each intricate detail.
In cSEM we seek to make working with reliabilities as consistent as possible by relying on a paper by @Cho2016 who proposed uniform formula-generating methods and a systematic naming conventions for all common reliability measures. Naturally, some of the conventional terminology is deeply entrenched within the nomenclatura of a particular filed (e.g., coefficient alpha alias Cronbach's alpha in pychonometrics) such that a new, albeit consistent, naming scheme seems superfluous at best. However, we belief the merit of a "standardized" naming pattern will eventually be helpful to all users as it helps clarify potential misconceptions thus preventing potential misuse, such as the (ab)use of Cronbach alpha as a reliability measure for congernic measurement models.
Apart from these considerations, this package takes a pragmatic stance in a sense that we use consistent naming because it naturally provides a consistent naming scheme for the functions and the systematic formula generating methods because they make code maintenance easier. Eventually, what matters is the formula and more so its correct application. To facilitate the translation between different naming systems and conventions we provide a "translation table" below:
Systematic names Mathematical Synonymous terms
Parallel reliability $\rho_P$ Spearman-Brown formula, Spearman-Brown prophecy, Standardized alpha, Split-half reliability
Tau-equivalent reliability $\rho_T$ Cronbach's alpha, $\alpha$, Coefficient alpha Guttmans $\lambda_3$, KR-20
Tau-equivalent reliability $\rho_{T;\text{weighted}}$ -- weighted
Congeneric reliability $\rho_C$ Composite reliability, Jöreskog's $\rho$, Construct reliability, $\omega$, reliability coefficient, Dillon-Goldsteins's $\rho$
Congeneric reliability $\rho_{C;\text{weighted}}$ Dijkstra-Henseler's $\rho_A$ weighted
Table: Systematic names and common synonymous names for the reliability estimates found in the literature
Closed-form confidence interval
@Trinchera2018 proposed a closed-form confidence interval (CI) for
the tau-equivalent reliability (Cronbach's alpha). To compute the CI, set
.closed_form_ci = TRUE when calling assess() or invoke
calculateRhoT(..., .closed_form_ci = TRUE) directly. The level of the CI
can be changed by supplying a single value or a vector of values to .alpha.
Implementation
The functions are implemented as calculateRhoC() and calculateRhoT().
The Goodness of Fit (GoF)
Definition
Calculate the Goodness of Fit (GoF) proposed by @Tenenhaus2004. Note that, contrary to what the name suggests, the GoF is not a measure of (overall) model fit in a $\chi^2$-fit test sense. See e.g. @Henseler2012a for a discussion.
Formulae
The GoF is defined as:
$$\text{GoF} = \sqrt{\varnothing \text{COM}k \times \varnothing R^2{structural}} = \sqrt{\frac{1}{k}\sum^K_{k=1} \lambda^2_k + \frac{1}{M} \sum^M_{m = 1} R^2_{m;structural}} $$ where $COM_k$ is the communality of indicator $k$, i.e. the variance in the indicator that is explained by its connected latent variable and $R^2_{m; structural}$ the R squared of the $m$'th equation of the structural model.
Implementation
The function is implemented as: calculateGoF().
The Heterotrait-Monotrait-Ratio of Correlations (HTMT)
Definition
The heterotrait-monotrait ratio of correlations (HTMT) was first proposed by
@Henseler2015 to assess convergent and discriminant validity.
Formulae
See: @Henseler2015 on page 121 (equation (6))
Implementation
The function is implemented as: calculateHTMT().
Literature
[^1]: There are some cutoffs such as e.g., the SRMR should be less than 0.08 or 0.1, however, these values are essentially arbitrary as they have never been formally motivated. Reference is usually done to @Hu1999 which based the cut-off on a simulation using factor-based SEM.
Try the cSEM package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.