{"id":282977,"date":"2023-10-12T14:04:55","date_gmt":"2023-10-12T12:04:55","guid":{"rendered":"https:\/\/climatescience.press\/?p=282977"},"modified":"2023-10-12T14:04:58","modified_gmt":"2023-10-12T12:04:58","slug":"regression-attenuation-only-depends-upon-the-relative-noise-in-x","status":"publish","type":"post","link":"https:\/\/climatescience.press\/?p=282977","title":{"rendered":"Regression attenuation only depends upon the relative noise in \u201cX\u201d"},"content":{"rendered":"\n
\"\"<\/figure>\n\n\n\n

<\/p>\n\n\n\n

This also impacts the analysis of the urban heat island (UHI).<\/em><\/strong><\/p>\n\n\n\n

From Roy Spencer, PhD.<\/a><\/p>\n\n\n\n

October 11th, 2023 by Roy W. Spencer, Ph. D.<\/small><\/p>\n\n\n\n

I\u2019m not a statistician, and I am hoping someone out there can tell me where I\u2019m wrong in the assertion represented by the above title. Or, if you know someone expert in statistics, please forward this post to them.<\/em><\/p>\n\n\n\n

In regression analysis we use statistics to estimate the strength of the relationship between two variables, say X and Y.<\/p>\n\n\n\n

Standard least-squares linear regression estimates the strength of the relationship (regression slope \u201cm\u201d) in the equation:<\/p>\n\n\n\n

Y = mX + b, where b is the Y-intercept.<\/p>\n\n\n\n

In the simplest case of Y = X, we can put in a set of normally distributed random numbers for X in Excel, and the relationship looks like this:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Now, in the real world, our measurements are typically noisy, with a variety of errors in measurement, or variations not due, directly or indirectly, to correlated behavior between X and Y. Importantly, standard least squares regression estimation assumes all of these errors are in Y, and not in X.<\/em><\/strong> This issue is seldom addressed by people doing regression analysis.<\/p>\n\n\n\n

If we next add an error component to the Y variations, we get this:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

In this case, a fairly accurate regression coefficient is obtained (1.003 vs. the true value of 1.000), and if you do many simulations with different noise seeds, you will find the diagnosed slope averages out to 1.000.<\/p>\n\n\n\n

But, if there is also noise in the X variable, a low bias in the regression coefficient appears, and this is called \u201cregression attenuation\u201d or \u201cregression dilution\u201d:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

This becomes a problem in practical applications because it means that the strength of a relationship diagnosed through regression will be underestimated to the extent that there are errors (or noise) in the X variable.<\/em><\/strong> This issue has been described (and \u201cerrors in variables<\/a>\u201d methods for treatment have been advanced) most widely in the medical literature, say in quantifying the relationship between human sodium levels and high blood pressure or heart disease. But the problem will exist in any field of research to the extent that the X measurements are noisy.<\/p>\n\n\n\n

One can vary the relative amounts of noise in X and in Y to see just how much the regression slope is reduced. When this is done, the following relationship emerges, where the vertical axis is the regression attenuation coefficient (the ratio of the diagnosed slope to the true slope) and the horizontal axis is how much relative noise is in the X variations:<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

What you see here is that if you know how much of the X variations are due to noise\/errors, then you know how much of a low bias you have in the diagnosed regression coefficient. For example, if noise in X is 20% the size of the signals in X, the underestimate of the regression coefficient is only 4%. But if the noise is the same size as the signal, then the regression slope is underestimated by about 50%.<\/p>\n\n\n\n

Noise in Y Doesn\u2019t Matter<\/strong><\/p>\n\n\n\n

But what the 3 different colored curves show is that for Y noise levels ranging from 1% of the Y signal, to 10 times the Y signal (a factor of 1,000 range in the Y noise), there is no effect on the regression slope (except to make its estimate more noisy when the Y noise is very large).<\/p>\n\n\n\n

There is a commonly used technique for estimating the regression slope called Deming regression<\/a>, and it assumes a known ratio between noise in Y versus noise in X. But I don\u2019t see how the noise in Y has any impact on regression attenuation<\/strong>. <\/em>All one needs is an estimate of the relative amount of noise in X, and then the regression attenuation follows the above curve(s).<\/p>\n\n\n\n

Anyway, I hope someone can point out errors in what I have described, and why Deming regression should be used even though my analysis suggests regression attenuation has no dependence on errors in Y.<\/p>\n\n\n\n

Why Am I Asking?<\/strong><\/p>\n\n\n\n

This impacts our analysis of the urban heat island (UHI) where we have hundreds of thousands of station pairs where we are relating their temperature difference to their difference in population density. At very low population densities, the correlation coefficients become very small (less than 0.1, so R2 less than 0.01), yet the regression coefficients are quite large, and \u2014 apparently \u2014 virtually unaffected by attenuation, because virtually all of the noise is in the temperature differences (Y) and not the population difference data (X).<\/p>\n\n\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

This also impacts the analysis of the urban heat island (UHI). From Roy Spencer, PhD. October 11th, 2023 by Roy W. Spencer, Ph. D. I\u2019m not a statistician, and I am hoping someone out there can tell me where I\u2019m wrong in the assertion represented by the above title. Or, if you know someone expert […]<\/p>\n","protected":false},"author":121246920,"featured_media":282985,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_coblocks_attr":"","_coblocks_dimensions":"","_coblocks_responsive_height":"","_coblocks_accordion_ie_support":"","_crdt_document":"","advanced_seo_description":"","jetpack_seo_html_title":"","jetpack_seo_noindex":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"This also impacts the analysis of the urban heat island (UHI).","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[1],"tags":[691823558,691823561,691823557,691821156,691823559,691823560],"class_list":{"0":"post-282977","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","6":"hentry","7":"category-uncategorized","8":"tag-errors","9":"tag-regression","10":"tag-relative-noise","11":"tag-urban-heat-island-uhi","12":"tag-x-variable","13":"tag-y-variable","15":"fallback-thumbnail"},"jetpack_publicize_connections":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/10\/image-304.png?fit=1000%2C600&ssl=1","jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/paxLW1-1bC9","jetpack-related-posts":[{"id":349393,"url":"https:\/\/climatescience.press\/?p=349393","url_meta":{"origin":282977,"position":0},"title":"A Demonstration: Low Correlations Do Not Necessarily Lead to Low Confidence in Data Regressions","author":"uwe.roland.gross","date":"10\/29\/2024","format":false,"excerpt":"In a recent post I used our new Urban Heat Island (UHI) warming estimates at individual U.S. GHCN stations having at least 120 years of data to demonstrate that the homogenized (adjusted) GHCN data still contain substantial UHI effects. Therefore, spurious warming from UHI effects is inflating reported U.S. warming\u2026","rel":"","context":"In \"GHCN data\"","block_context":{"text":"GHCN data","link":"https:\/\/climatescience.press\/?tag=ghcn-data"},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0urbanheatisland.jpg?fit=1200%2C648&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0urbanheatisland.jpg?fit=1200%2C648&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0urbanheatisland.jpg?fit=1200%2C648&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0urbanheatisland.jpg?fit=1200%2C648&ssl=1&resize=700%2C400 2x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0urbanheatisland.jpg?fit=1200%2C648&ssl=1&resize=1050%2C600 3x"},"classes":[]},{"id":348954,"url":"https:\/\/climatescience.press\/?p=348954","url_meta":{"origin":282977,"position":1},"title":"Urban Heat Island Effects Have Not Yet Been Removed from Official GHCN Warming Trends","author":"uwe.roland.gross","date":"10\/26\/2024","format":false,"excerpt":"Our paper (co-authored by John Christy and Danny Braswell) on computing the urban heat island (UHI) effect as a function of population density (PD) is now in the final stages of review after a 3rd round of edits, and I\u2019m hopeful it will be accepted for publication soon.","rel":"","context":"In \"GHCN Warming Trends\"","block_context":{"text":"GHCN Warming Trends","link":"https:\/\/climatescience.press\/?tag=ghcn-warming-trends"},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0Screenshot-2024-10-26-083344.png?fit=1200%2C851&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0Screenshot-2024-10-26-083344.png?fit=1200%2C851&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0Screenshot-2024-10-26-083344.png?fit=1200%2C851&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0Screenshot-2024-10-26-083344.png?fit=1200%2C851&ssl=1&resize=700%2C400 2x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2024\/10\/0Screenshot-2024-10-26-083344.png?fit=1200%2C851&ssl=1&resize=1050%2C600 3x"},"classes":[]},{"id":255147,"url":"https:\/\/climatescience.press\/?p=255147","url_meta":{"origin":282977,"position":2},"title":"Urbanization Effects on GHCN Temperature Trends, Part IV: UHI Effects on Tmax and Tmin","author":"uwe.roland.gross","date":"04\/28\/2023","format":false,"excerpt":"This is part 4 of my series on quantifying Urban Heat Island (UHI) effects on surface air temperatures as reported in the monthly GHCN datasets produced by NOAA.","rel":"","context":"In \"GHCN temperature trends\"","block_context":{"text":"GHCN temperature trends","link":"https:\/\/climatescience.press\/?tag=ghcn-temperature-trends"},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/04\/00UHI-May-June-July-Tmax-Tmin-2.jpg?fit=864%2C960&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/04\/00UHI-May-June-July-Tmax-Tmin-2.jpg?fit=864%2C960&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/04\/00UHI-May-June-July-Tmax-Tmin-2.jpg?fit=864%2C960&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/04\/00UHI-May-June-July-Tmax-Tmin-2.jpg?fit=864%2C960&ssl=1&resize=700%2C400 2x"},"classes":[]},{"id":248390,"url":"https:\/\/climatescience.press\/?p=248390","url_meta":{"origin":282977,"position":3},"title":"Urbanization Effects on GHCN Temperature Trends, Part III: Using Population Density, 1880-2015","author":"uwe.roland.gross","date":"03\/18\/2023","format":false,"excerpt":"Thus, previous investigations that found little difference in temperature trends between urban and rural sites don\u2019t really prove anything.","rel":"","context":"Similar post","block_context":{"text":"Similar post","link":""},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/03\/image-676.png?fit=1200%2C675&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/03\/image-676.png?fit=1200%2C675&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/03\/image-676.png?fit=1200%2C675&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/03\/image-676.png?fit=1200%2C675&ssl=1&resize=700%2C400 2x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/03\/image-676.png?fit=1200%2C675&ssl=1&resize=1050%2C600 3x"},"classes":[]},{"id":378625,"url":"https:\/\/climatescience.press\/?p=378625","url_meta":{"origin":282977,"position":4},"title":"SHOCK CLIMATE REPORT! Urban Heat Islands Responsible for 65% of Global Warming","author":"uwe.roland.gross","date":"05\/16\/2025","format":false,"excerpt":"A new study from the University of Alabama in Huntsville addresses the question of how much the Urban Heat Island (UHI) effect is responsible for the higher temperatures at weather stations across the world. Dr. Roy Spencer and Dr. John Christy have spent several years developing a novel method that\u2026","rel":"","context":"In \"climactic \u201cgolden age\u201d\"","block_context":{"text":"climactic \u201cgolden age\u201d","link":"https:\/\/climatescience.press\/?tag=climactic-golden-age"},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2025\/05\/0Screenshot-2025-05-16-165630.png?fit=1200%2C671&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2025\/05\/0Screenshot-2025-05-16-165630.png?fit=1200%2C671&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2025\/05\/0Screenshot-2025-05-16-165630.png?fit=1200%2C671&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2025\/05\/0Screenshot-2025-05-16-165630.png?fit=1200%2C671&ssl=1&resize=700%2C400 2x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2025\/05\/0Screenshot-2025-05-16-165630.png?fit=1200%2C671&ssl=1&resize=1050%2C600 3x"},"classes":[]},{"id":280654,"url":"https:\/\/climatescience.press\/?p=280654","url_meta":{"origin":282977,"position":5},"title":"Summer warming 1895-2023 in U.S. cities exaggerated by 100% from the urban heat island effect","author":"uwe.roland.gross","date":"09\/27\/2023","format":false,"excerpt":"We are now getting close to finalizing our methodology for computing the urban heat island (UHI) effect as a function of population density, and will be submitting our first paper for publication in the next few weeks. I\u2019ve settled on using the CONUS (Lower 48) U.S. region as a demonstration\u2026","rel":"","context":"In \"1895-2023\"","block_context":{"text":"1895-2023","link":"https:\/\/climatescience.press\/?tag=1895-2023"},"img":{"alt_text":"","src":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/09\/01895-2023-CONUS-JJA-time-series-for-blog-post.jpg?fit=1200%2C675&ssl=1&resize=350%2C200","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/09\/01895-2023-CONUS-JJA-time-series-for-blog-post.jpg?fit=1200%2C675&ssl=1&resize=350%2C200 1x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/09\/01895-2023-CONUS-JJA-time-series-for-blog-post.jpg?fit=1200%2C675&ssl=1&resize=525%2C300 1.5x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/09\/01895-2023-CONUS-JJA-time-series-for-blog-post.jpg?fit=1200%2C675&ssl=1&resize=700%2C400 2x, https:\/\/i0.wp.com\/climatescience.press\/wp-content\/uploads\/2023\/09\/01895-2023-CONUS-JJA-time-series-for-blog-post.jpg?fit=1200%2C675&ssl=1&resize=1050%2C600 3x"},"classes":[]}],"_links":{"self":[{"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/posts\/282977","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/users\/121246920"}],"replies":[{"embeddable":true,"href":"https:\/\/climatescience.press\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=282977"}],"version-history":[{"count":4,"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/posts\/282977\/revisions"}],"predecessor-version":[{"id":282986,"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/posts\/282977\/revisions\/282986"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/climatescience.press\/index.php?rest_route=\/wp\/v2\/media\/282985"}],"wp:attachment":[{"href":"https:\/\/climatescience.press\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=282977"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/climatescience.press\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=282977"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/climatescience.press\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=282977"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}