Bumped version number, date and added cran-comments. Amended the vign…

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: seminr
 Type: Package
 Title: Domain-Specific Language for Building PLS Structural Equation Models
-Version: 0.5.2
-Date: 2018-10-11
+Version: 0.5.3
+Date: 2018-10-18
 Authors@R: c(person("Soumya", "Ray",
               email = "soumya.ray@gmail.com", role = c("aut", "ths")),
              person("Nicholas", "Danks",

cran-comments.md

@@ -1,21 +1,9 @@
 ## Resubmission
 This is a resubmission. In this version I have:
 
-## Attended to a CRAN error:
-We see such failures occasionally: for the record this time it is
-
- SEMinR Model succesfully bootstrapped── 1. Failure: Seminr evaluates the factor discriminant validity p_values correc
- diff[1, 2] is not strictly less than 0.1. Difference: 0.000314
-
-Seems you failed to heed the warnings in §1.6 of 'Writing R Extensions': please re-read it and correct ASAP (CRAN is shut until Sep 10) and before Sep 21 to safely retain the package on CRAN.
-
-Also try a spell checker --- it is 'successfully'.
-
--- 
-Brian D. Ripley,                  ripley@stats.ox.ac.uk
-Emeritus Professor of Applied Statistics, University of Oxford
-
-Specifically, we updated the bootstrap method to take a seed for reproducability and added a tolerance of 0.0001 for floating point calculation tests. All tests including random processes such as bootstrapping were made reproducible.
+## Attended to a user-reported error:
+Single-item interactions were generating a bug in the code. 
+Issue #84 on https://github.com/sem-in-r/seminr/issues/84
 
 ## Test environments
 * macOS Sierra 10.12.6 (on travis-ci), R 3.5.0

inst/doc/SEMinR.Rmd

@@ -348,7 +348,7 @@ mobi_pls <- estimate_pls(data = mobi,
 
 Dijkstra and Henseler (2015) offer an adjustment to generate consistent weight and path estimates of common factors estimated using PLSPM. SEMinR automatically adjusts for consistent estimates of coefficients for common-factors defined using `reflective()`.
 
-Note: At this point, SEMinR does not adjust for PLSc on models with interactions involving common-factors. Thus in models including common-factors and interactions, the coefficient of common-factors will be subject to bias.
+Note: SEMinR does adjust for PLSc on models with interactions involving common-factors. Models with interactions can be estimated as PLS consistent, but are subject to some bias as per Becker et al. (2018). In small sample sizes, bootstrapping such a PLSc model with interactions can cause errors.
 
 ## Bootstrapping the model for significance
 
@@ -415,6 +415,7 @@ summary(boot_mobi_pls)
 
 ## References
 
+* Becker et al. (2018). Estimating Moderating Effects in PLS-SEM and PLSc-SEM: Interaction Term Generation*Data Treatment
 * Dijkstra, T. K., & Henseler, J. (2015). Consistent Partial Least Squares Path Modeling, MIS Quarterly Vol. 39(X).
 * Dillon, W. R, and M. Goldstein. 1987. Multivariate Analysis: Methods, and Applications. Biometrical Journal 29 (6): 750–756.
 * Fornell, C. and D. F. Larcker (February 1981). Evaluating structural equation models with unobservable variables and measurement error, Journal of Marketing Research, 18, pp. 39-5)

inst/doc/SEMinR.html

@@ -288,7 +288,7 @@ <h3>Describe all model interactions with <code>interactions()</code></h3>
 #&gt;     intxns_list &lt;- lapply(all_intxns, create_interaction)
 #&gt;     return(intxns_list)
 #&gt;   }
-#&gt; &lt;environment: 0x1149fa470&gt;</code></pre>
+#&gt; &lt;environment: 0x10e7240d0&gt;</code></pre>
 <p>Not that these functions themselves return functions that are not resolved until passed as a parameter to the <code>estimate_pls()</code> function for model estimation.</p>
 </div>
 <div id="describe-an-interaction-with-interaction_ortho-or-interaction_scaled" class="section level3">
@@ -399,7 +399,7 @@ <h2>PLS SEM Model Estimation</h2>
 <div id="consistent-pls-plsc-for-common-factors" class="section level3">
 <h3>Consistent PLS (PLSc) for common-factors</h3>
 <p>Dijkstra and Henseler (2015) offer an adjustment to generate consistent weight and path estimates of common factors estimated using PLSPM. SEMinR automatically adjusts for consistent estimates of coefficients for common-factors defined using <code>reflective()</code>.</p>
-<p>Note: At this point, SEMinR does not adjust for PLSc on models with interactions involving common-factors. Thus in models including common-factors and interactions, the coefficient of common-factors will be subject to bias.</p>
+<p>Note: SEMinR does adjust for PLSc on models with interactions involving common-factors. Models with interactions can be estimated as PLS consistent, but are subject to some bias as per Becker et al. (2018). In small sample sizes, bootstrapping such a PLSc model with interactions can cause errors.</p>
 </div>
 </div>
 <div id="bootstrapping-the-model-for-significance" class="section level2">
@@ -420,7 +420,7 @@ <h2>Bootstrapping the model for significance</h2>
                                  nboot = 1000,
                                  cores = 2)
 #&gt; Bootstrapping model using seminr...
-#&gt; SEMinR Model succesfully bootstrapped</code></pre>
+#&gt; SEMinR Model successfully bootstrapped</code></pre>
 <p>Notably, bootstrapping can also be meaningfully applied to models containing interaction terms and readjusts the interaction term (Henseler and Chin 2010) for every sub-sample. This leads to slightly increased processing times, but provides accurate estimations.</p>
 </div>
 <div id="reporting-the-pls-sem-model" class="section level2">
@@ -480,19 +480,19 @@ <h3>Reporting the bootstrapped <code>boot_seminr_model</code></h3>
 #&gt; 
 #&gt; Structural Path t-values:
 #&gt;                   Satisfaction
-#&gt; Image                    8.031
-#&gt; Expectation              2.847
-#&gt; Value                    4.496
-#&gt; Image*Expectation        0.936
-#&gt; Image*Value              0.050
+#&gt; Image                    7.954
+#&gt; Expectation              2.754
+#&gt; Value                    4.590
+#&gt; Image*Expectation        0.997
+#&gt; Image*Value              0.032
 #&gt; 
 #&gt; Structural Path p-values:
 #&gt;                   Satisfaction
 #&gt; Image                    0.000
-#&gt; Expectation              0.005
+#&gt; Expectation              0.006
 #&gt; Value                    0.000
-#&gt; Image*Expectation        0.349
-#&gt; Image*Value              0.960</code></pre>
+#&gt; Image*Expectation        0.319
+#&gt; Image*Value              0.974</code></pre>
 <ol start="2" style="list-style-type: decimal">
 <li><code>boot_model_summary &lt;- summary(boot_seminr_model)</code> returns an object of class <code>summary.boot_seminr_model</code> which contains the following accessible objects:
 <ul>
@@ -506,6 +506,7 @@ <h3>Reporting the bootstrapped <code>boot_seminr_model</code></h3>
 <div id="references" class="section level2">
 <h2>References</h2>
 <ul>
+<li>Becker et al. (2018). Estimating Moderating Effects in PLS-SEM and PLSc-SEM: Interaction Term Generation*Data Treatment</li>
 <li>Dijkstra, T. K., &amp; Henseler, J. (2015). Consistent Partial Least Squares Path Modeling, MIS Quarterly Vol. 39(X).</li>
 <li>Dillon, W. R, and M. Goldstein. 1987. Multivariate Analysis: Methods, and Applications. Biometrical Journal 29 (6): 750–756.</li>
 <li>Fornell, C. and D. F. Larcker (February 1981). Evaluating structural equation models with unobservable variables and measurement error, Journal of Marketing Research, 18, pp. 39-5)</li>

vignettes/SEMinR.Rmd

@@ -348,7 +348,7 @@ mobi_pls <- estimate_pls(data = mobi,
 
 Dijkstra and Henseler (2015) offer an adjustment to generate consistent weight and path estimates of common factors estimated using PLSPM. SEMinR automatically adjusts for consistent estimates of coefficients for common-factors defined using `reflective()`.
 
-Note: At this point, SEMinR does not adjust for PLSc on models with interactions involving common-factors. Thus in models including common-factors and interactions, the coefficient of common-factors will be subject to bias.
+Note: SEMinR does adjust for PLSc on models with interactions involving common-factors. Models with interactions can be estimated as PLS consistent, but are subject to some bias as per Becker et al. (2018). In small sample sizes, bootstrapping such a PLSc model with interactions can cause errors.
 
 ## Bootstrapping the model for significance
 
@@ -415,6 +415,7 @@ summary(boot_mobi_pls)
 
 ## References
 
+* Becker et al. (2018). Estimating Moderating Effects in PLS-SEM and PLSc-SEM: Interaction Term Generation*Data Treatment
 * Dijkstra, T. K., & Henseler, J. (2015). Consistent Partial Least Squares Path Modeling, MIS Quarterly Vol. 39(X).
 * Dillon, W. R, and M. Goldstein. 1987. Multivariate Analysis: Methods, and Applications. Biometrical Journal 29 (6): 750–756.
 * Fornell, C. and D. F. Larcker (February 1981). Evaluating structural equation models with unobservable variables and measurement error, Journal of Marketing Research, 18, pp. 39-5)