<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Aaron Sojourner on Medium]]></title>
        <description><![CDATA[Stories by Aaron Sojourner on Medium]]></description>
        <link>https://medium.com/@aaronsojourner?source=rss-ad45afd38ed7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*z4snp-OXnBp8uzg6zICK4Q.jpeg</url>
            <title>Stories by Aaron Sojourner on Medium</title>
            <link>https://medium.com/@aaronsojourner?source=rss-ad45afd38ed7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 22:50:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@aaronsojourner/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Americans favor labor unions over big business now more than ever]]></title>
            <link>https://medium.com/@aaronsojourner/americans-favor-labor-unions-over-big-business-now-more-than-ever-c082c21c0ca9?source=rss-ad45afd38ed7------2</link>
            <guid isPermaLink="false">https://medium.com/p/c082c21c0ca9</guid>
            <category><![CDATA[unions]]></category>
            <category><![CDATA[polling]]></category>
            <category><![CDATA[labor]]></category>
            <category><![CDATA[organizing]]></category>
            <category><![CDATA[business]]></category>
            <dc:creator><![CDATA[Aaron Sojourner]]></dc:creator>
            <pubDate>Tue, 20 May 2025 16:10:02 GMT</pubDate>
            <atom:updated>2025-05-20T19:53:12.281Z</atom:updated>
            <content:encoded><![CDATA[<p>Aaron Sojourner and Adam Reich</p><p>For decades, Americans were evenly divided in their relative support of labor unions and big business, but that’s no longer the case. Now, Americans are more likely to side with labor than at any time in the past 60 years. For people whose instincts about economic and political conflicts between unions and big business were honed more than a decade ago, it’s time to update your understanding.</p><p>According to newly-released American National Election Studies (ANES) data from late 2024, analyzed and reported here for the first time, Americans feel more positively toward labor unions and more negatively toward big business than any time since ANES began asking the question in 1964. Using consistent methods to allow comparability over time, ANES uses representative samples of Americans and asks them to rate their feelings toward labor unions and big business on a 0–100 scale.</p><h3>Trends broke after 2012</h3><p>Between 1964 and 2012, Americans’ sentiments toward labor unions and big business moved together, surging and dipping in tandem. The figure below shows how average feelings toward each group changed over time. For instance, during mid-1970s stagflation, Americans cooled to both unions and big business.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x_UC5s3DW6BJm43Uq1kQXA.png" /><figcaption>Figure 1</figcaption></figure><p>In 2016, Americans’ sentiments toward labor and business split. Feelings toward big business stayed flat while feelings toward labor unions warmed to nearly a record high.</p><p>In 2020, warmth toward labor unions kept climbing to a record high while sentiment toward big business fell to a record low.</p><p>Just-released data from late 2024 reveal that warmth toward labor unions climbed even higher while sentiment to big business fell even further, setting new records for warmth to labor unions and coolness to big business.</p><p>A simple way to summarize this is to compute the sentiment difference between labor and big business for each individual who answers both questions. If the person rates both labor and business the same, this difference equals zero. If they rate labor 70 and business 30, the difference equals 40. The more positive the difference, the more pro-union the person is relative to big business. A negative difference expresses more pro-business than pro-labor sentiment.</p><p>The trend in the average sentiment difference illustrates just how unprecedented the current situation is. Labor unions have huge sympathy from Americans relative to big business right now, far more than any time on record. The largest gap between 1964 and 2012 was about +5 pro-union in 1974. Now we are at +16.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oOrkNSZdwm8Ugmpuz6r8aA.png" /></figure><h3>Digging deeper</h3><p>Looking at subgroups where the changes are bigger and smaller can offer insight into what might be driving this recent change. While Gallup has published similar high-level trends on <a href="https://news.gallup.com/poll/12751/labor-unions.aspx">unions</a> and <a href="https://news.gallup.com/poll/5248/big-business.aspx">major corporations</a> over the years, the ANES data enable analysis by subgroup. Also, the ANES survey is more recent than the latest Gallup ones.</p><p>The changes for men and for women are similar. This change is not driven by differential attitude changes by gender.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-Bi9gV9r0_PEGPVZos3EaQ.png" /></figure><p>In contrast, recent trends do differ by education group. Until 2012, Americans with the most formal education (Bachelor+) were the most pro-business. By 2024, they became the most pro-union. On the other hand, Americans with the least formal education (Less than High School) were historically among the most pro-union but recently became the least pro-union. High school graduates, some college, and bachelor plus all expressed record-high pro-union sentiment in 2024. Even those without a high school degree remained near their record high.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-52lqSGd3rl-CymruTvlfA.png" /></figure><p>People who identify with different political ideologies experienced similar changes since 2012. Liberals, conservative, moderates, and “Don’t Knows” are more pro-union than ever.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uYMx2gMJp5cAtIl2zwIa7w.png" /></figure><p>Pro-union shifts occurred among both union members and nonunion workers, stronger among nonunion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oRK_7bAQJU0t3-4sFd2pNA.png" /></figure><p>Different race and ethnic groups had similar trends, with all groups becoming more pro-union in rough parallel since 2012.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6pHT2nN2KhQc-lFkOnzRyA.png" /></figure><p>Between 2012 and 2024, sentiment moved most in a pro-union direction in Alaska, Delaware, and Vermont, with average difference shifts over 40 points pro-union in each of these states. Substantial increases also occurred in some surprising places, in the South (Tennessee, Mississippi, and Louisiana) and in the Southwest and Mountain West (Arizona, Colorado, and Utah). Sentiment became more pro-business only in Wyoming and Montana.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mgtmQEGnsHD9ulilyJaTBQ.png" /></figure><p>There are several possible explanations for the public’s increasing support for unions. First, broad campaigns by low-wage worker organizations such as the <a href="https://en.wikipedia.org/wiki/Fight_for_$15">Fight for $15</a>, <a href="https://en.wikipedia.org/wiki/Starbucks_unions">Starbucks Workers United</a>, and <a href="https://en.wikipedia.org/wiki/Amazon_Labor_Union">Amazon Labor Union</a> have become increasingly visible. Second, <a href="https://www.washingtonpost.com/politics/2020/03/02/two-populist-movements-sanders-trump/">populist politics</a> surged in Sanders’s presidential campaigns and the rise of <a href="https://en.wikipedia.org/wiki/Squad_(U.S._Congress)">the Squad</a> in Congress. President Trump sought to soften the GOP’s conventional labor policy agenda with labor-friendly rhetoric during his first term and 2024 campaign. Then-President Biden’s clear <a href="https://www.govexec.com/management/2024/05/bidens-labor-report-card-historian-gives-union-joe-higher-grade-any-president-fdr/397002/">pro-union positions</a> made union support more prominent inside the Democratic coalition. Third, <a href="https://realtimeinequality.org/">wealth concentration</a> and <a href="https://www.cfr.org/backgrounder/us-inequality-debate">dissatisfaction</a> with it has continued to rise. Fourth, some bosses’ reactions to the pandemic — layoffs and forcing workers to take on more risks —made clear to many Americans that their boss would use power against them. This seems to have left Americans <a href="https://www.shrm.org/executive-network/insights/covid-19-likely-to-cause-spike-union-organizing">looking for more control</a> over their job and working conditions.</p><p>The rise in public sympathy has not yet translated into a higher share of workers forming unions. While an increasing share of nonunion workers <a href="https://journals.sagepub.com/doi/abs/10.1177/0019793918806250">want a union in their workplace</a>, <a href="https://www.epi.org/publication/unlawful-employer-opposition-to-union-election-campaigns/">unlawful</a> management opposition to workers’ organizing campaigns remains fierce.</p><p>The National Labor Relations Act treats white-collar, strategic, and repeat violators of U.S. workers’ rights the way we treat children who stole something for the first time, requiring them only to give back what they took and say they are sorry. The law, written 90 years ago, provides no possibility of fines or penalties — only back pay — and so provides <a href="https://www.piie.com/publications/working-papers/2021/do-us-firms-have-incentive-comply-flsa-and-nlra">little incentive for compliance</a>.</p><p>Regardless of the cause of the split, in fights between organized labor and organized capital, Americans are likely to support labor more now than any time on record.</p><p><strong>Also posted at EPI: </strong><a href="https://www.epi.org/blog/americans-favor-labor-unions-over-big-business-now-more-than-ever/">https://www.epi.org/blog/americans-favor-labor-unions-over-big-business-now-more-than-ever/</a></p><p><strong>About the authors:</strong> Sojourner is a labor economist at the <a href="http://www.upjohn.org/">W.E. Upjohn Institute for Employment Research</a>. Reich is a professor of sociology at Columbia University and co-director of the <a href="https://laborlabcu.org/">Columbia Labor Lab</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c082c21c0ca9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Differential Non -Response by Nativity May Distort U.S. Labor Force Statistics]]></title>
            <link>https://medium.com/@aaronsojourner/how-differential-non-response-by-nativity-may-distort-u-s-labor-force-statistics-9b24e19894a3?source=rss-ad45afd38ed7------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b24e19894a3</guid>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[economics]]></category>
            <category><![CDATA[economy]]></category>
            <category><![CDATA[labor]]></category>
            <category><![CDATA[labor-market]]></category>
            <dc:creator><![CDATA[Aaron Sojourner]]></dc:creator>
            <pubDate>Sun, 18 May 2025 14:05:06 GMT</pubDate>
            <atom:updated>2025-05-18T14:05:06.178Z</atom:updated>
            <content:encoded><![CDATA[<p>By Jing Cai, Brad Hershbein, and Aaron Sojourner</p><h3>High-level summary</h3><ul><li>Survey response rates appear to be falling faster for foreign-born than for native-born Americans, which can bias estimates of population sizes and labor force statistics under current official statistical procedures.</li><li>Geographic areas with higher foreign-born shares start out with higher average Current Population Survey (CPS) response rates than other areas in 2016. However, their response rates decline over time to roughly equal the rates of areas with lower foreign-born shares by 2024. This decline is almost entirely driven by falling initial response rates.</li><li>The Census Bureau and the Bureau of Labor Statistics adjust sample weights on the basis of age, sex, race, and ethnicity, but <em>not </em>foreign-born share. This means changing response rates by areas with a greater share of Black residents should have minimal effects on both aggregate statistics (which we find), like the labor force participation and employment rates, or counts, like total labor force or employment. However, this need not be the case for areas with changing CPS response by foreign-born share.</li><li>When we adjust sample weights to account for changing response rates in areas with a greater share of foreign-born residents, we find that labor force participation rates and employment rates have grown slightly faster than the official numbers indicate, although not enough to make a substantive difference.</li><li>However, this underestimation problem may continue to grow if areas with many foreign-born — which have high workforce participation rates — become increasingly undercounted in the CPS.</li></ul><h3><em>Introduction</em></h3><p>In recent years, overall response rates to the U.S. Current Population Survey (CPS) have declined, a <a href="https://www.bls.gov/cps/methods/response_rates.htm">fact</a> understood by the U.S. Bureau of Labor Statistics (BLS), Census Bureau, and many CPS users. However, we know much less about whether response rates have been declining evenly across the board or whether declines have been sharper for certain groups of the population. This distinction matters, because without proper statistical corrections or other survey adjustments, changing response rates across different groups may distort our understanding of how the size of the labor force and key labor force statistics, such as the labor force participation rate (LFPR) and employment rate, are changing.</p><p>We report new evidence that, since 2016, CPS response rates in areas with more foreign-born residents have declined faster than response rates in areas with more native-born residents, and that these differentially changing response rates have led to slight — but increasing — underestimation of the growth in the share of the working age population that is foreign-born.¹ In turn, because the foreign-born are <a href="https://fred.stlouisfed.org/graph/?g=1FFUj">more likely to work</a> than the native-born, changing survey response rates affect the numbers and rates of Americans who are in the labor force and employed. However, the good news is that we find that these positive adjustments are (currently) quite small, less in magnitude than the change from one month to the next that could be due to sampling error alone. This survey response rate issue is distinct from, but compounds with, the <a href="https://www.hamiltonproject.org/publication/post/higher-new-census-population-estimates-will-affect-the-employment-report/">recent upward revision</a> in total population estimates in implying that population, labor force, and employment counts have been growing faster than recently thought.</p><h3><em>CPS Response Rates</em></h3><p>Studying CPS response rates by demographic group is not straightforward. You may wonder, for example, how one can measure the nativity or race of <em>non-respondents</em> to construct subgroup-specific response rates? By definition, their nonresponse means their nativity and race are unreported. We cannot solve this directly.</p><p>Instead, we study response rates in local geographies known to have different resident populations. The Census Bureau produces reliable estimates of the share of residents who are native-born versus foreign-born and by race and ethnicity from the ongoing American Community Survey and makes these available at various levels of geography, from neighborhoods to states. From 2016 forward, we link each household invited to respond to the CPS (the location of which is known) to their local community’s nativity share and race share, and we then estimate response rates by community type.² We look at trends in response rates for communities with a high share of native-born residents separately from those with a low share, for instance.</p><p>Figure 1 communicates one of our main findings: response rates have fallen more in communities with above-median share of their population foreign-born than in other communities since 2016. Each invited household is asked to respond to eight waves of the CPS.³ In 2016, the average response rate among households in communities with a high foreign-born share was 74%. In these same communities, the average response rate declined 13 percentage points, to 61%, by 2024. Communities with low foreign-born shares started with a 71% response rate in 2016 but ended at 60%, implying a smaller 11 percentage point decline. The gap shrank in 2020, when COVID-19 erupted, before closing in 2023 and remaining very small since.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*pflPaKbv7VLeF6OrBtNRHA.png" /><figcaption>Figure 1</figcaption></figure><p>The faster decline in response rates in areas with more foreign-born is driven by differential changes in <em>initial</em> response rates — the first wave of the survey (Figure 2). Trends in the response rate to subsequent waves — conditional on having responded to an earlier wave — are similar (Figure 3).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*C27VMYRfQLgdu39Ai-XbvQ.png" /><figcaption>Figure 2</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*c_UL6YR0XIQccCdjYE7u9g.png" /><figcaption>Figure 3</figcaption></figure><h3><em>Adjusting for Differential CPS Response</em></h3><p>This issue can distort our understanding of the labor market if unrecognized and unaddressed. Specifically, faster declines in response by households with foreign-born members will lead to underestimation in the growth of the foreign-born population, labor force, and employment level, because their true population will be increasingly under-represented in the estimated population based on respondents. This occurs because the statistical weighting procedure used in the CPS to make the survey sample aggregate to the whole population accounts for <em>overall</em> response and certain demographic factors, but <a href="https://www.census.gov/programs-surveys/cps/technical-documentation/methodology/weighting.html"><em>not</em> the nativity status of the household</a>.⁴</p><p>We attempt to correct for this issue in Figure 4. Specifically, for each respondent household observation after 2016, we adjust its CPS weight by multiplying it by a correction factor that is designed to bring the household group’s overall response rate back to its 2016 average while preserving the total count of people. This inflates the weight for households from more recent years in areas with a greater share of the foreign-born and reduces the weight for more-recent native-born households.</p><p>Figure 4 shows that this adjustment implies a larger share of the (working-age) foreign-born population than what is obtained using official estimates. Whether we adjust weights based on being in different halves of the foreign-born geographic distribution (red line), different fifths (green line, nearly invisible), or different tenths (yellow line), the picture is the same: a foreign-born share of the population about 0.15 percentage points higher by the end of 2024 than the official weights (blue line) imply. This translates into nearly 400,000 additional foreign-born working-age people (and nearly 400,000 fewer native-born).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*_QpIJNFnCK_DAMfm-fwbyg.png" /><figcaption>Figure 4</figcaption></figure><p>Because the foreign-born population are more likely to work than the native-born, this adjustment also affects trends in the size of the labor force — but not by much. As shown in Figure 5, the change to the labor force is minimal, an increase of about 110,000 (0.07%), less than the monthly volatility in this count that is likely due to sampling error.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*TSgUZ8Nll8zSqA6r3uzuBA.png" /><figcaption>Figure 5</figcaption></figure><p>Because the change in the labor force is small, and population is held fixed even with the adjusted weights by foreign-born share of the area, the labor force participation rate also changes very little, by less than 0.1 percentage points. Employment growth and changes in the employment-population ratios similarly are minimally affected.</p><h3><em>Conclusion</em></h3><p>This would seem to be good news: although the CPS is missing a growing share of responses in areas with higher foreign-born shares of the population, this surveying difficulty has not substantially affected important economic indicators. At least, not yet by the end of 2024. However, the problem may grow in the future for three reasons. First, the <a href="https://www.bls.gov/web/empsit/cps-pop-control-adjustments.pdf">recently revised total population estimates</a>, from the beginning of this year, increased the total civilian population by roughly 3 million, but they do not correct for changing survey response rates across different areas of the country, so the issue we identify has likely been scaled up. Second, and more important, perhaps, is that the decline in response rates to the CPS in heavily foreign-born areas may accelerate under policies that increase the risk of deportation of immigrants. If more immigrants choose not to respond to the CPS, and they are still more likely to work than the native-born, then underestimation of labor force and employment growth will become more severe. Third, this growth underestimation could be compounded further under cuts to the CPS sample size if the sample design is not adjusted to account for changing response rates. Although a plan to reduce the CPS sample size was <a href="https://www.bls.gov/cps/notices/2024/cps-funding.htm">paused</a> last fall, the fix was temporary, and without adequate funding, cuts may once again be on the table later this year.</p><p>The Census Bureau has been exploring the possibility of switching to <a href="https://www.census.gov/library/working-papers/2021/acs/2021_Rothbaum_01.html">more recently developed weighting techniques</a>, which could help address these problems. However, they (and the Bureau of Labor Statistics) will require sufficient resources to make these changes and ensure that our labor market statistics rely on cutting-edge methods and remain as accurate as possible.</p><p><strong>About the authors: </strong>Cai is a senior analyst and Hershbein and Sojourner are senior economists at the <a href="https://www.upjohn.org/">W.E. Upjohn Institute for Employment Research</a>.</p><h3><strong>Technical Appendix</strong></h3><p>The CPS reports for all sampled households (whether they respond or not) the geographic area in which the household resides, whether the household responded at all to the survey, and, if it did, the demographics and labor market information of each household member and whether the household responded to each of (up to) seven subsequent interview attempts. This information allows us to measure whether each invited household initially responded to the CPS, an initial response rate among all invited households in each area (see endnote 1), and, for initiators in each community type, a response rate for subsequent invitations.</p><p>¹ We define the foreign-born as individuals who are either non-citizens or naturalized citizens.</p><p>² The CPS identifies the location of sampled households at various levels of geography to protect privacy and confidentiality. We observe county of residence for households from approximately the 280 most populous counties (in which almost two-thirds of the U.S. population live). For these households we assign shares at the county level. Another sixth of the population live in a metro area that is identified in the CPS but for which the county is not observed; for these households we assign shares based on the balance of the metro area, after netting out shares in identifiable counties. The last sixth or so of the population lives outside a metro area, and we observe only their state in the CPS. For these households, we assign shares based on the balance of the state, after netting out shares in identifiable counties and metro areas. In total, we observe over 480 locations. We rank these locations based on foreign-born or Black share and divide them into two (or more) groups in each case, based on the midpoint (quantile) of the total share of the population (that is, we use the population-weighted median (quantile) as the dividing point(s)).</p><p>³ See <a href="https://cps.ipums.org/cps/sample_designs.shtml">https://cps.ipums.org/cps/sample_designs.shtml</a>.</p><p>⁴ We conducted a similar exercise using the local share of residents who are Black. Although we do find a small decrease in the <em>subsequent</em> response rate among households in areas with a greater share of Black residents, the overall decrease in response rate is much smaller than it is for areas with a greater share of foreign-born residents. Both for this reason, and because CPS population weights explicitly adjust to independent totals of the Black population, there are essentially no effects of this changing response rate on estimates of the population, labor force, and employment.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b24e19894a3" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>