UJI VALIDITAS, RELIABILITAS, TINGKAT KESUKARAN DAN DAYA PEMBEDA SOAL TIPE URAIAN

Summary

In this video, Nur Rahmaka, a Mathematics Education student, demonstrates how to analyze essay-type questions using SPSS for a final course assignment. The tutorial covers testing question validity, reliability, discrimination power, and difficulty level. Step-by-step instructions guide viewers through preparing data in SPSS, performing bivariate correlations for validity, calculating Cronbach’s Alpha for reliability, evaluating discrimination power, and determining difficulty levels using average scores. The results show that all questions are valid, reliability is high, most questions have good discrimination, and difficulty levels range from easy to medium. This comprehensive guide helps educators assess and improve test quality effectively.

Takeaways

• 😀 The video demonstrates how to conduct a validity test, reliability test, difficulty level analysis, and distinguishing power analysis on essay-type questions using SPSS.
• 😀 The tutorial begins by explaining how to import essay-type question data into SPSS by copying and pasting from an external source into the application.
• 😀 The variable names in SPSS are adjusted to represent each question (1-10) and a total score, making it easier for further analysis.
• 😀 A validity test is performed by correlating each question with the total score using SPSS's bivariate correlation method. Significant correlations indicate valid questions.
• 😀 Questions with a correlation significance marked with stars (1 star for 5% level, 2 stars for 1% level) are considered valid based on the output.
• 😀 The reliability test uses Cronbach's Alpha to determine the overall consistency of the set of questions. A Cronbach's Alpha value of 0.799 indicates high reliability.
• 😀 The discriminating power of each question is analyzed based on total item statistics in SPSS, with correlations categorized as good, acceptable, or bad depending on their value.
• 😀 The video explains how to categorize discriminating power based on correlation values: above 0.40 is good, 0.30-0.39 is acceptable without revision, and below 0.20 is bad.
• 😀 Difficulty levels of the questions are analyzed by dividing the average score of each question by the highest score (3), with results categorized as easy, medium, or difficult.
• 😀 Each question’s difficulty level is calculated and classified, with most questions falling into the ‘easy’ category, while a few are categorized as ‘medium.’

Q & A

Who created the video and what is its purpose?

The video was created by Nur Rahmaka, a Mathematics Education student, to fulfill the final assignment of the Mathematics Learning Assessment course taught by Mrs. Fauziah Hakim.

What type of questions does the video focus on for testing?

The video focuses on essay-type questions and demonstrates how to conduct validity, reliability, difficulty level, and discriminating power tests using SPSS.

How is the data for essay-type questions entered into SPSS?

The data is copied and pasted into SPSS, with variable names adjusted to 'questions 1 to 10' and the total score labeled as 'total score'. Decimal points are set to zero and the measurement type is set to 'scale'.

What steps are involved in conducting the validity test in SPSS?

To conduct the validity test, click 'Analyze' → 'Correlate' → 'Bivariate', move all question variables to the variable column, select 'Pearson' for correlation, check 'Two-tailed' for significance, and enable 'Flag significant correlations'.

How can you determine if a question is valid based on the SPSS output?

A question is considered valid if the correlation value is significant. SPSS indicates significance with stars: one star for 0.05 significance and two stars for 0.01 significance.

What method is used to test the reliability of the questions?

Reliability is tested using 'Analyze' → 'Scale' → 'Reliability Analysis' in SPSS, with Cronbach's Alpha selected. The value of Alpha indicates reliability, with 0.799 considered high.

How is the discriminating power of each question assessed?

Discriminating power is assessed using the 'Corrected Item-Total Correlation' in the SPSS output. Values below 0.20 are poor, 0.20–0.29 are acceptable with revision, 0.30–0.39 are acceptable without revision, and 0.40–1 are good.

What is the procedure for calculating the difficulty level of questions?

The difficulty level is calculated by dividing the average score of each question by the maximum score. Values closer to 1 indicate easier questions, while lower values indicate medium or difficult questions.

Which questions were categorized as easy and which as medium in difficulty?

Questions 1–7 and 9 were categorized as easy, while questions 8 and 10 were categorized as medium difficulty based on the calculated ratios of average score to maximum score.

Why is it important to test validity, reliability, difficulty, and discriminating power in educational assessments?

Testing these aspects ensures that questions accurately measure student knowledge (validity), produce consistent results (reliability), are appropriately challenging (difficulty), and effectively distinguish between high and low performers (discriminating power), leading to fair and effective assessments.

What SPSS features are repeatedly used throughout the analysis?

The key SPSS features used include 'Data View' for entering data, 'Analyze' menu for correlation, reliability analysis, and descriptive statistics, and output tables to interpret validity, reliability, discriminating power, and difficulty levels.