SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 

 View Only
  • 1.  the crosstabulation (chi square) is empty

    Posted Wed June 28, 2023 08:32 AM
      |   view attached

    Hi everyone!

    I have to analyze the results of a questionnaire for my thesis with SPSS and have a problem: in my questionnaire I had 4 conditions and every person was randomly assigned to one of the conditions. Now, to understand if there is a correlation between the conditions, I have to run a Chi-Square analysis of independence for all binary outcomes, but the test doesn't work. When I try to run the Chi-sqaure test I get the warning "The crosstabulation of Control Condition 1 * Condition 2 is empty.". I suppose because as every participant was assigned to a different condition, the responses are in different rows for each condition (see attachment: the 4 conditions are MQ01, MQ02, MQ03 and MQ04).

    Could please someone help me to solve the problem? Thank you!



    ------------------------------
    Elisabetta Podetti
    ------------------------------


  • 2.  RE: the crosstabulation (chi square) is empty

    Posted Wed June 28, 2023 08:49 AM

    The warning message suggests that there is an empty crosstabulation, which can occur if there are no overlapping categories between the two variables being analyzed. In your case, this could be because each participant was assigned to a different condition, resulting in separate rows for each condition and no overlap in responses between conditions.

    To address this problem, you may need to restructure your data in a way that allows for meaningful comparisons between the conditions. 

    I might suggest you perform separate analyses: Instead of running a Chi-Square analysis of independence across all conditions simultaneously, you could perform separate Chi-Square analyses for each pair of conditions. This would involve comparing Condition 1 with Condition 2, Condition 1 with Condition 3, Condition 1 with Condition 4, Condition 2 with Condition 3, Condition 2 with Condition 4, and Condition 3 with Condition 4. This approach allows you to examine the relationships between pairs of conditions individually.

    I hope this clarifies things for you



    ------------------------------
    Youssef Sbai Idrissi
    Software Engineer
    ------------------------------



  • 3.  RE: the crosstabulation (chi square) is empty

    Posted Wed June 28, 2023 08:58 AM
    Edited by Elisabetta Podetti Wed June 28, 2023 08:59 AM

    Thank you very much! I was already thinking of comparing them separately, but how can I restructure the data, before I do the separate comparison? Is there a way to do it? Thank you very much.



    ------------------------------
    Elisabetta Podetti
    ------------------------------



  • 4.  RE: the crosstabulation (chi square) is empty

    Posted Wed June 28, 2023 09:05 AM
    You should be aware of the multiple testing problem with testing everything separately.  Your chance of rejecting the null hypothesis based on many separate tests is higher than the nominal significance level, so you need to make a correction for this.  The STATS PADJUST extension command, which you can install via the Extensions > Extension Hub menu, offers several different ways of making this adjustment.

    It's not clear to me exactly what you are trying to do.  If you could explain further what you are trying to test and how the data are set up, we can probably suggest a better solution than doing a batch of separate tests.

    --





  • 5.  RE: the crosstabulation (chi square) is empty

    Posted Wed June 28, 2023 09:28 AM

    I run a questionnaire to investigate individuals' perceptions of greenwashing in the context of purchasing decisions and their ability to distinguish between greenwashing claims and genuine sustainability attributes. The questionnaire had 4 conditions to which participants entering the questionnaire were randomly assigned. In each condition they had to choose between two t-shirts (1=t-shirt A, 2=t-shirt B) and each conditions had a different type of description for the t-shirts. I want to test if there is a correlation between the conditions, so if the difference between people's choices is statistically significant. Being all conditions categorical and binary outcomes, I supposed I have to run a chi-square test to test it. 

    I hope it is more clear now and you can help me somehow, thank you



    ------------------------------
    Elisabetta Podetti
    ------------------------------