{"id":1017,"date":"2013-11-19T03:46:30","date_gmt":"2013-11-19T10:46:30","guid":{"rendered":"http:\/\/www.mii.ucla.edu\/causality\/?p=1017"},"modified":"2013-11-19T03:46:30","modified_gmt":"2013-11-19T10:46:30","slug":"the-key-to-understanding-mediation","status":"publish","type":"post","link":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/2013\/11\/19\/the-key-to-understanding-mediation\/","title":{"rendered":"The Key to Understanding Mediation"},"content":{"rendered":"<p><strong>Judea Pearl Writes:<\/strong><\/p>\n<p>For a long time I could not figure out why SEM researchers find it hard to embrace the &#8220;causal inference approach&#8221; to mediation, which is based on counterfactuals. My recent conversations with David Kenny and Bengt Muthen have opened my eyes, and I am now pretty sure that I have found both the obstacle and the key to making causal mediation an organic part of SEM research.<\/p>\n<p>Here is the key:<\/p>\n<p>Why are we tempted to &#8220;control for&#8221; the mediator M when we wish to estimate the direct effect of X on Y? The reason is that, if we succeed in preventing M from changing then whatever changes we measure in Y are attributable solely to variations in X and we are justified then in proclaiming the effect observed as &#8220;direct effect of X on Y&#8221;. Unfortunately , the language of probability theory does not possess the notation to express the idea of &#8220;preventing M from changing&#8221; or &#8220;physically holding M constant&#8221;. The only operation probability allows us to use is &#8220;conditioning&#8221; which is what we do when we \u201ccontrol for M\u201d in the conventional way (i.e., let M vary, but ignore all samples except those that match a specified value of M).  This habit is just plain wrong, and is the mother of many confusions in the practice of SEM.<\/p>\n<p>To find out why, you are invited to visit: <a href=\"http:\/\/ftp.cs.ucla.edu\/pub\/stat_ser\/r421.pdf\">http:\/\/ftp.cs.ucla.edu\/pub\/stat_ser\/r421.pdf<\/a>,  paragraph starting with &#8220;In the remaining of this note, &#8230;&#8221;, on page 2. <\/p>\n<p>Best,<br \/>\nJudea<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Judea Pearl Writes: For a long time I could not figure out why SEM researchers find it hard to embrace the &#8220;causal inference approach&#8221; to mediation, which is based on counterfactuals. My recent conversations with David Kenny and Bengt Muthen have opened my eyes, and I am now pretty sure that I have found both [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10,16,26],"tags":[],"class_list":["post-1017","post","type-post","status-publish","format-standard","hentry","category-definition","category-general","category-mediated-effects"],"_links":{"self":[{"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/posts\/1017","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/comments?post=1017"}],"version-history":[{"count":0,"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/posts\/1017\/revisions"}],"wp:attachment":[{"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/media?parent=1017"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/categories?post=1017"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/causality.cs.ucla.edu\/blog\/index.php\/wp-json\/wp\/v2\/tags?post=1017"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}