Canonical Correlation using R

00:02:41
https://www.youtube.com/watch?v=7TKvgpe3YOQ

الملخص

TLDRThe Stat Tree module introduces Canonical Correlation Analysis, a statistical method to explore relationships between two sets of continuous variables. At least four continuous variables are required for conducting this test. Using the PEVS dataset, the module guides users on selecting relevant variables and running the analysis in RStudio. It details how to interpret the output, emphasizing the significance figure's importance in research reporting. The tutorial also highlights how to document findings properly, including additional statistics that should be reported if criteria are met, and instructions for reviewing output samples.

الوجبات الجاهزة

  • 📊 Understand Canonical Correlation to analyze relationships between variable sets.
  • 📋 At least four continuous variables are necessary for proper testing.
  • 📚 Use PEVS dataset: educational vs entertainment responses.
  • 📝 Learn to run Canonical Correlation in RStudio with provided code.
  • ⚙️ First check for significance figures in test results.
  • ✔️ Report findings if significance criteria are met.
  • 📉 If not significant, report the lack of difference.
  • 🔍 View output examples for better understanding.
  • 🎯 Include Wilks’ Lambda and coefficients in reports.

الجدول الزمني

  • 00:00:00 - 00:02:41

    The Canonical Correlation module explains the process for testing relationships between two sets of continuous variables, requiring at least four variables to perform the test. Using the Preferred Entertainment Viewing Survey as the dataset, the demonstration employs educational responses as one variable set and entertainment responses as another. In RStudio, users must specify variables from the PEVS dataset and run the Canonical Correlation test by entering provided code into a Script box and executing it. The significance figure from the output is crucial; if it meets set criteria, results must be documented in research papers, including F score figures, Wilks' Lambda value, squared canonical correlations, and standardized coefficients. If significance criteria are not met, results are not reported, but a statement about the lack of significant differences must be included.

الخريطة الذهنية

فيديو أسئلة وأجوبة

  • What is Canonical Correlation used for?

    It tests the relationship between two sets of continuous variables.

  • How many continuous variables are required to run the Canonical Correlation?

    You need at least four continuous variables.

  • What dataset is used in this demonstration?

    The Preferred Entertainment Viewing Survey (PEVS) dataset.

  • Which variables are selected for analysis in the demo?

    'History' and 'Learn' for educational responses, 'Entertain' and 'Rewind' for entertainment responses.

  • What should be included in the report if the significance figure is met?

    Report significance figures, F score figures, Wilks’ Lambda, squared canonical correlations, and standardized coefficients.

  • What if the significance figure does not meet the criterion?

    State that no significant difference or relationship was found.

  • How can I see the output for Canonical Correlation?

    Press the [OUTPUT] button in the Stat Tree module.

عرض المزيد من ملخصات الفيديو

احصل على وصول فوري إلى ملخصات فيديو YouTube المجانية المدعومة بالذكاء الاصطناعي!
الترجمات
en-US
التمرير التلقائي:
  • 00:00:03
    Welcome to the Stat Tree video module for Canonical Correlation.
  • 00:00:08
    To test whether there is a relationship
  • 00:00:10
    between two sets of continuous variables,
  • 00:00:13
    the Canonical Correlation would be appropriate.
  • 00:00:16
    You will need at least four continuous variables in order to run this test.
  • 00:00:21
    In looking at the Preferred Entertainment Viewing Survey,
  • 00:00:24
    all of the scale questions three through seven
  • 00:00:27
    would be appropriate for the continuous variables for our demonstration.
  • 00:00:31
    For this demonstration we will use educational related responses
  • 00:00:35
    for one set of variables,
  • 00:00:37
    and entertainment related responses for the other set of variables.
  • 00:00:42
    With the PEVS dataset loaded into the global environment in RStudio,
  • 00:00:46
    you would choose history and learn for the first set of variables,
  • 00:00:50
    and entertain and rewind for the second set.
  • 00:00:54
    To run the Canonical Correlation test for multiple association,
  • 00:00:58
    create a new Script box.
  • 00:01:00
    In the Script box, type in the following code.
  • 00:01:04
    This code can be copied from the new browser window
  • 00:01:07
    when you press the [SCRIPT] button on the sending page
  • 00:01:09
    from the Stat Tree module.
  • 00:01:12
    Select the code.
  • 00:01:13
    Now press the Run icon.
  • 00:01:16
    R will generate an output in the Console window.
  • 00:01:19
    As with all statistical tests,
  • 00:01:22
    the first result to look for is the significance figure
  • 00:01:24
    associated with the test score.
  • 00:01:27
    If the significance figure meets your preset criterion,
  • 00:01:30
    you will need to report your test results in your research paper.
  • 00:01:34
    You will find that significance figure here.
  • 00:01:38
    In this test, we have met our significance criteria for the relationship
  • 00:01:42
    between Educational and Entertainment reasons
  • 00:01:44
    for watching shows and movies.
  • 00:01:47
    Here are the test results.
  • 00:01:49
    You would report these figures,
  • 00:01:51
    as well as F score figures for any canonical correlations
  • 00:01:55
    which meet the significance criterion.
  • 00:01:58
    Additionally, the Wilks’ Lambda figure,
  • 00:02:01
    the squared canonical correlations for the significant functions,
  • 00:02:04
    and a table of standardized coefficients and canonical loadings
  • 00:02:08
    for each variable in the analysis should be reported.
  • 00:02:12
    If the significance figure does not meet the criterion,
  • 00:02:15
    test results do not need to be reported in your research paper.
  • 00:02:20
    However, you will need to report
  • 00:02:21
    that the test failed to find a significant difference
  • 00:02:24
    or relationship in the data.
  • 00:02:27
    To see what this output looks like,
  • 00:02:28
    along with a reporting example,
  • 00:02:31
    press the [OUTPUT] button on the sending page
  • 00:02:33
    from the Stat Tree module.
الوسوم
  • Canonical Correlation
  • RStudio
  • Statistical Analysis
  • Continuous Variables
  • Preferred Entertainment Viewing Survey
  • Significance Figure
  • Wilks’ Lambda
  • Research Reporting
  • Output Interpretation
  • Educational Responses