Usage Notes

Some key questions and answers to facilitate the correct use of the CLC Estimator

• What can the CLC Estimator do?  
  The app only estimates unidimensional latent constructs based on congeneric approaches. CLC Estimator is not intended for statistical procedures commonly available in existing statistical packages, such as exploratory factor analysis, principal component analysis or confirmatory factor analysis.

• When should the CLC Estimator be used?
  CLC Estimator performs latent construct estimation via congeneric approaches when the available statistical packages do not include this function.

• At what point in the data analysis process the CLC Estimator should be used?
  CLC Estimator should be integrated into the data analysis process after review (including reliability analysis) of the items to be retained in the latent construct estimation.

• When should the CLC Estimator not be used?
  CLC Estimator assumes a common factor model, i.e. a reflective latent construct. Therefore, it is not appropriate when the latent construct is a composite (Rhemtulla et al., 2020).

• How do you deal with missing data, outliers, and other data issues in the CLC Estimator?
  CLC Estimator does not handle missing data, outliers or incomplete data. Users are encouraged to address these issues before using the CLC Estimator. Specifically for missing data, note that the CLC Estimator uses the standard missing data handling procedures embedded in the 'psych' and 'mirt' R packages. For the EFA model, the correlation matrix is computed based on pairwise complete observations. For the graded IRT model, full information maximum likelihood is used. If the data contain missing values, the CLC Estimator warns the user immediately.

• How to create a usable .csv file for the CLC Estimator?
  CLC Estimator supports standard .csv files generated by popular statistical packages. The first row of the .csv file should contain the labels of the variables (i.e. the header), while the other rows should contain the data. The .csv file loading interface allows users to select the delimiter between values (comma, semicolon, tab). The .csv file should be formatted in UNICODE, with decimal symbols expressed in points according to international scientific conventions. A sample dataset in .csv format (comma delimiter) is available at: https://dx.doi.org/10.6084/m9.figshare.21786335

• What estimation method should be used to estimate the latent constructs?
  When social scientists use previously validated scales, the same estimation method should be used. If neither the estimation method is specified by the original developers of the scales nor such scales have been previously validated, maximum likelihood estimation (MLE) is recommended (for a detailed discussion of this topic, see Thompson, 2004).

• What type of output does the CLC Estimator produce?
  CLC Estimator generates a standard .csv file with variable names in the first row and values in the remaining rows. The generated .csv file output can be easily imported into popular statistical packages.

References

Rhemtulla, M., van Bork, R., & Borsboom, D. (2020). Worse than measurement error: Consequences of inappropriate latent variable measurement models. Psychological Methods, 25(1), 30–45. https://doi.org/10.1037/met0000220

Thompson, B. (2004). Exploratory and confirmatory factor analysis: Understanding concepts and applications. American Psychological Association. https://doi.org/10.1037/10694-000

Information on the CSV input file

The CLC Estimator accepts as input only .csv files, which stands for "comma-separated values".  This format is commonly used by many statistical software packages to store and exchange data. A .csv file is made up of rows and columns, much like a spreadsheet. The first row, called the header, should have labels for each column to help identify the data. The rows below the header contain the actual data, separated by a specific character called a delimiter.

For example, if we have a small dataset of people's names and ages, it might look like this in a .csv file:

Name,Age

Alice,30

Bob,25

Carol,22

Here, the first row is the header, which has the labels "Name" and "Age." The following rows contain the data, with each value separated by a comma (which is the delimiter in this case).

When using the CLC Estimator, you can choose the delimiter that separates the values in your .csv file. It can be a comma, semicolon, or a tab. This flexibility allows the tool to work with different .csv file formats. It's also important to note that the .csv file should be saved in UNICODE format, which is a standard for encoding text and characters. Additionally, decimal numbers should be written with a period (.) as the decimal separator, as per international scientific conventions.

For reference, a sample dataset is available for download in .csv format (using a comma delimiter) at the following link: https://dx.doi.org/10.6084/m9.figshare.21786335. You can use this sample dataset to better understand the .csv file format and practice using the CLC Estimator.

Below you you can find some examples, using the above-mentioned sample dataset, showing how the .csv should look in different software. 

Running the CLC Estimator app offline 

There are two approaches to running the CLC Estimator app offline on your PC.

Important: Please note in both cases, to run the app offline, you need to install Rtools (available only for Windows at the moment), which can be downloaded from: https://cran.r-project.org/bin/windows/Rtools/

Approach 1. Install from R or R Studio

Enter the following code in the R console (install also the dependencies):

install.packages('devtools') #install devtools package

library('devtools') #load devtools package

install_github('leoegidi/clc') #install clc package from GitHub

library('clc') #load clc package

clc() #load clc estimator

Approach 2. Download the package from GitHub and run it in R Studio

1. Enter in our GitHub page: https://github.com/LeoEgidi/clc

2. On the top right of the page, click "Code" and then "Download ZIP".

3. Unpack the zip file.

4. Navigate from "clc-main" folder to "R" folder.

5. Double-click on "app.R" file and open it in R Studio.

6. Install all the dependencies suggested from the pop-up in R Studio.

7. Click on "Run App".