multiTree
multiTree is a java-based computer program for the analysis of multinomial processing tree models. multiTree provides parameter estimation, hypothesis testing, checks for identifiability, bootstrapping, and power analysis. For details, see
Moshagen, M. (2010). multiTree: A computer program for the analysis of multinomial processing tree models. Behavior Research Methods, 42, 42鈥�54.
Download and Install
Note that multiTree has been written almost 20 years ago, so there are some quirks when running multiTree on modern operation systems. See below.
multiTree for Mac
Depending on the Mac you own, you either need the Intel (pre 2021) or the ARM (most models since 2021; M series) version of both Java and multiTree. If you own an Intel-based Mac, the (JRE) is usually pre-installed. If not, just download and install the latest version . If you own a more modern ARM Mac, you (only) require the latest (JDK). Make sure to install the macOS ARM64 JDK.
As Apple does not consider me a trusted developer, you need to launch multiTree using the terminal. To do so, open the terminal, move to the directory containing multiTree.jar (change directories using 鈥渃d directoryname鈥�, and type 鈥渏ava -jar -XstartOnFirstThread multitree.jar鈥�. This may give some warnings you can just ignore.
multiTree for Unix
Make sure a current (JRE) is installed on the target machine. multiTree comes in a 32bit and a 64 bit version. The correct version depends on whether the installed JRE is 32bit or 64bit.
Extract the file and run multiTree.sh.
multiTree for Windows
Make sure a current (JRE) is installed on the target machine. multiTree comes in a 32bit and a 64 bit version. The correct version depends on whether the installed JRE is 32bit or 64bit.
After downloading, launch the installer and follow the on-screen instructions.
The multiTree user interface looks disarranged on (most modern) high resolution displays. This can be changed by applying a proper scaling to the UI elements as follows: Right-click on multitree.exe, select Properties, select the Compatibility tab, and then select 鈥淒isable display scaling on high DPI settings鈥� (set to System).
Basic usage
Details are given in Moshagen, M. (2010). multiTree: A computer program for the analysis of multinomial processing tree models. Behavior Research Methods, 42, 42鈥�54.
Using multiTree involves three steps:
- Define model in terms of model equations
- Enter data
Save file, estimate model and interpret the results.
1. Model equations
The model equations are entered in 鈥淓quations鈥� tab and the follow this structure:
Treelabel {whitespace} Categorylabel {whitespace} Equation
For instance, the following equations define the two-high threshold (2HT) model:
oldItems Oold d
oldItems Oold (1-d)*g
oldItems Onew (1-d)*(1-g)
newItems Nnew d
newItems Nold (1-d)*g
newItems Nnew (1-d)*(1-g)
Things to note:
- Write one line for each branch
- Literally use the parameter or its complement, e.g. (1-d)*g Don鈥檛 be smart and merge equations (e.g., g-dg)
- multiTree is case-sensitive, so a parameter 鈥淒鈥� is different from a parameter 鈥渄鈥�
- You can assign nearly arbitrary label for the trees, categories and parameters, but don't use any of the following characters: ( ) * -
- Use distinctive labels for each tree and each category.
- A single category may only occur in a single tree. For example, an 鈥淥ld鈥� response for a 鈥淣ew鈥� item must have a different label than an 鈥淥ld鈥� response for an 鈥淥ld鈥� item. If you have the same model structure for different conditions /groups, duplicate the model and pre- or postfix the trees, categories and parameters accordingly, e.g. 鈥渙ldItems_c1 鈥� and 鈥渙ldItems_c2 鈥�.
- A single parameter may occur in different trees. If you want these to differ, use proper labeling, e.g. 鈥渄1鈥� and 鈥渄2鈥�.
2. Data
The data are entered in the 鈥淒ata鈥� tab and follow this structure:
Title (mandatory)
Categorylabel {whitespace} Count
For instance, here are some data for the 2HT defined above:
Example data
Oold 72
Onew 5
Nnew 19
Nold 58
Things to note:
- The first line contains the title. This is required.
- Write one line for each category.
- Use the same category labels as used in the model equations.
- Multiple data sets can be defined by clicking 鈥�+鈥�
3. Save file, estimate model, and interpret results.
After having entered the equations and the data, the file should be saved. This triggers validation processes and populates the 鈥淧arameters鈥� tab with the model parameters. There, restrictions on the parameters can be placed (such as g = .5). To estimate the model and view the results, click the green arrow. For some notes on evaluating the results, consult the annotated output.