ANOVA: A comparison test with SPSS Statistics

Over the past few months, our R&D team has worked around the clock to improve our ANOVA algorithm. Slicker and faster, you’ll now be able to complete ANOVA tests in far less time than with SPSS.

XLSTAT disposes more than 50 test cases on ANOVA in order to evaluate the performance of this feature for various scenarios and validate the output.

As part of the latest improvements, we ran a comparison test with SPSS Statistics on a two-way ANOVA followed by a Tukey post-hoc test.

The dataset

The data set contains one dependent variable (Score)and two explanatory variables (Product, Individual). The factor Product includes 4 modalities while the factor Individual includes 1250 modalities.

The Test

The test was performed on a machine with the following catacteristics:

• Windows 10 64 bits
• RAM 8,00 Go
• Intel(R) Core(TM) i7-4810MQ CPU
• XLSTAT 2020.1.1
• Office 2016 32-bits
• IBM SPSS Statistics Build 1.0.0.1327 64-bit

Setting a two-way ANOVA in XLSTAT

Setting a two-way ANOVA in SPSS Statistics

The Results

XLSTAT computed and displayed the results (ANOVA table plus multiple comparisons) in around 30 secs, using this specific machine, whereas it took SPSS 3min30sec to perform the exact same task. The Excel file including the data set can be downloaded on the top of the page. The execution time may change from one machine to another.