Mediation Analysis in Psychometrics

KL
by Kirsten Lambert2 views·Updated Dec 13, 2022
Create your own

Chapters

Audio TranscriptAuto-generated

  1. Hi,
  2. I am Kirsten lambert and today I am going to be talking with you all about
  3. mediation analysis and specifically we're going to be
  4. talking about mediation analysis in psycho metrics.
  5. Um So let's get started
  6. when looking at data.
  7. Typically researchers are focusing on an independent and a dependent variable,
  8. but not everything is always that simple.
  9. So sad.
  10. Sometimes there is an intervening variable
  11. that is where mediation analysis comes in.
  12. Like I just stated mediation analysis looks at an intervening variable.
  13. So when a researcher is using this type of analysis,
  14. they are using a regression in order to evaluate the
  15. mechanism through which an independent variable known as X,
  16. affects
  17. a dependent variable, which is known as why.
  18. So with that I want to take some time to discuss
  19. some other terms that are interchangeable when using mediation analysis.
  20. So the first one um of these is the predictor.
  21. A predictor is known as a independent variable or X.
  22. The second term is the outcome variable.
  23. This is also known as the dependent variable or why
  24. The final term is the mediator.
  25. The mediator is the additional intervening variable that is collected in the data.
  26. The mediator creates an indirect path between the predictor
  27. and the outcome variable
  28. also known as
  29. the independent and dependent variable, or X and Y.
  30. So let's talk about why it's important or why we should even use mediation analysis.
  31. Um The purpose of using mediation analysis is to better
  32. understand how a predictor and an outcome variable are related.
  33. So in other words, we want to
  34. have helped a researcher or psychometric mission
  35. um be able to understand how the independent and the
  36. dependent variables are related when there isn't an obvious connection
  37. mediation analysis assists researchers in finding
  38. that missing links between the independent
  39. and the dependent variables also known as our predictor and outcome variables.
  40. Now this is so important because it is going to help the
  41. researchers truly understand the reasoning behind
  42. the outcome once they've found that.
  43. So we're gonna break down mediation analysis a little bit more.
  44. When conducting mediation analysis,
  45. an individual conduct a three regression analyses.
  46. Those analyses are the relationship between
  47. the predictor and the outcome variable.
  48. So X. To Y.
  49. The relationship between the predictor and the mediator. So x. two m.
  50. And then finally the relationship of the predictor
  51. and the mediator on the outcome burial.
  52. So X and M. Two Y.
  53. Now this is illustrated in the video next to
  54. the definition of our breakdown
  55. as you can see and then if you're following along
  56. on our hand out it is also in figure one.
  57. So I know the jargon of statistics and psycho metrics can get very um confusing.
  58. So I wanted to help everybody um better understand mediation analysis.
  59. And so I think it's really helpful to be able to use a real life example.
  60. So a common one that is used throughout academic literature
  61. is the one that we see on the screen or
  62. right now.
  63. Um
  64. And that is the relationship between grades and happiness.
  65. So one person might hypothesize that the higher the grade and an individual
  66. has in school may be related to a higher level of happiness.
  67. As one may take a deeper look at this relationship.
  68. one may be able to identify an intervening variable that may affect the happiness.
  69. Now, what's typically seen in again,
  70. the academic literature is this intervening variable known as self esteem,
  71. and you can see that in the second part of that picture and figure
  72. the self esteem of an individual can affect the relationship
  73. between the students grade and their level of happiness.
  74. So, let's take this into a little example here,
  75. let's say we have student A who has great grades, but really poor self esteem,
  76. they may have a lower level of happiness compared to student B.
  77. Student B has a moderate
  78. level of grades, moderate grades, but very strong self esteem.
  79. So when comparing student A and student B,
  80. their levels of happiness could be equal, yet their grades be very different,
  81. and we can see that this can be affected by their self esteem.
  82. So this shows that there is that direct relationship between
  83. grades and the level of happiness.
  84. But when you take that closer look of
  85. all the factors affecting the students happiness,
  86. you can see that there is an intervening
  87. variable which is also known as our mediator.
  88. Once you've completed your mediation analysis.
  89. So you're three regressions, um we can start assessing the data.
  90. And so it's important to start with identifying
  91. the relationship between the predictor and the outcome
  92. variable. So again, that's X to Y.
  93. If there is no significant relationship between these two variables,
  94. the independent the dependent variable.
  95. There is nothing that can be determined as a mediating variable
  96. because if there's no connection between X and Y,
  97. there's not gonna be a third enter um a third variable entering the situation.
  98. However, if there's a significant relationship,
  99. a researcher can continue to assess the next relationship,
  100. that next relationship is to determine the relationship
  101. between the predictor or independent variable and the mediator
  102. without a relationship to the predictor and intervening
  103. variable can not cannot be a mediator.
  104. If there is a relationship,
  105. then a researcher can continue to assess the final relationship
  106. if there is a significant relationship between the 1st and 2nd tests. So again,
  107. the independent independent variable and the
  108. independent variable to the mediator.
  109. Then a researcher will be able to determine how the predictor
  110. and the mediator are both related to the outcome variable.
  111. This is shown
  112. by a weakened relationship between the predictor and the outcome variable.
  113. When there is a strong enough mediating relationship,
  114. the relationship between the predictor and
  115. the outcome variable can even disappear.
  116. This occurs because there is another variable,
  117. the mediator that is impacting the relationship.
  118. So let's discuss how
  119. the use of mediation analysis occurs in psycho metrics.
  120. But first, let's discuss what psycho metrics is
  121. statistics and psycho metrics are often seen as synonymous fields.
  122. Although they use similar tools, these fields are not the same
  123. psycho metrics.
  124. Is the measurement of a human characteristics, such as behavior or health
  125. statistics, focuses
  126. specifically on analyzing numerical data.
  127. So why is it important to understand
  128. mediation analysis as a psychometric shin psychometric.
  129. Asians use many tests and statistics to analyze the data they collect.
  130. Mediation analysis is one of the ways in which they analyze their data.
  131. Without it, they would lock the ability to see all relationships among variables
  132. and that concludes the presentation for mediation analysis in psycho metrics.
  133. Thank you
  134. very much.