I'm new here, and somewhat of a lay-user in terms of statistics, and hoping this question will be at the appropriate place.

I'm running the mentioned analysis in SPSS on 3 iterations of an experiment I ran. Most compared metrics are shown in graphs similar to this, with the expected ranking across the 3 iterations. 

 However, in some cases, for some metrics, I'm getting graphs that show 5 bars instead of the expected 3. 

This is what I'm using to generate the analysis for all metrics:

NPTESTS

/RELATED TEST(T0_2_time T1_2_time T2_2_time T3_2_time)

/MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE

/CRITERIA ALPHA=0.05 CILEVEL=95.

As I can't find any information regarding this, I'm hoping someone can help me understand what this means/represents.

Many thanks.

For anyone stumbling upon this in the future, I was able to find that this is a visualization "quirk" from SPSS with no impact on the test itself. It is due to having several duplicate values in the dataset, particularly when at the extremes of the data range. When multiple cases in the dataset share the same value, SPSS assigns them the same rank, resulting in clusters of tied ranks. These clusters are split into additional bars to reflect the frequency of tied values, which can result in additional bars appearing in the graph.

I tested introducing some randomized variation to the dataset's duplicate values to remove duplicates (just for testing purposes) and the extra bars did indeed go away.

Original Message:
Sent: Sat December 14, 2024 09:31 AM
From: Alexandre Clément
Subject: Related-samples friedman's two-way analysis of variance by ranks extra bars in graph

 I'm new here, and somewhat of a lay-user in terms of statistics, and hoping this question will be at the appropriate place.

I'm running the mentioned analysis in SPSS on 3 iterations of an experiment I ran. Most compared metrics are shown in graphs similar to this, with the expected ranking across the 3 iterations. 

 However, in some cases, for some metrics, I'm getting graphs that show 5 bars instead of the expected 3. 

This is what I'm using to generate the analysis for all metrics:

NPTESTS

/RELATED TEST(T0_2_time T1_2_time T2_2_time T3_2_time)

/MISSING SCOPE=ANALYSIS USERMISSING=EXCLUDE

/CRITERIA ALPHA=0.05 CILEVEL=95.

As I can't find any information regarding this, I'm hoping someone can help me understand what this means/represents.

Many thanks.

------------------------------
Alexandre Clément
------------------------------