4.4 Exercises#

Dichotomous IRT#

We will now solve a few exercises on Item Response Theory with Dichotomous Data.

1️⃣ Dimensionality assessment#

Instructions:

  • Download and load the dataset

  • Test unidimensionality on the addition items of the zareki dataset.

  • What conclusion can you draw from the results of the analysis?

  • Are there any items that seem to violate unidimensionality? If so, remove them and run the analysis again. What has changed?

# Data handling

# Princals

# EFA tetrachoric correlation

# ICA

2️⃣: The Rasch model#

Instructions:

  • Analyze the item parameters of a Rasch model on all the mathematical items in the zareki dataset

    • Assess general model fit

    • Assess misfit at the item level

    • Clean the dataset and re-fit the Rasch model

    • Interpret the result

  • Plot the item response functions (ICC)

# Data handling

# FIT testing

# Cleaning and comparison

# ICC

3️⃣: 2-PL model#

Instructions:

  • Replicate the analysis above, but this time use a 2-PL model

  • Extract item difficulties and item discrimination into a python pandas.dataframe

  • Create 2 subplots.

    • 1st subplot: barplot showing item difficulties

    • 2nd subplot: barplot showing item discrimination

    • add xlabels and ylabels and a title

# 2-PL model

# Item parameters extraction

# Plotting