<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div> </div> <div dir="ltr"> </div><blockquote type="cite"><div dir="ltr"><div class="WordSection1"> <div align="center"> <table class="MsoNormalTable" border="0" cellspacing="0" cellpadding="0" width="600" style="width:450.0pt"> <tbody> <tr> <td valign="top" style="padding:0cm 0cm 0cm 0cm"> <img border="0" width="600" height="181" style="width:6.25in;height:1.8854in" id="Imagem_x0020_1" src="cid:image006.png@01D4FE8E.D2497E70" alt="cid:image001.png@01D4C861.DA6A4ED0"><o:p></o:p> </td> </tr> </tbody> </table> </div> <o:p></o:p> <div align="center"> <table class="MsoNormalTable" border="0" cellspacing="0" cellpadding="0" width="600" style="width:450.0pt"> <tbody> <tr> <td style="background:white;padding:30.0pt 37.5pt 37.5pt 37.5pt"> Ciência de Dados e Decisão Seminários Acadêmicos<o:p></o:p> Making Sense of Sensitivity: Extending Omitted Variable Bias<o:p></o:p> <o:p></o:p> Palestrante: Carlos Cinelli<o:p></o:p> Universidade: University of California, Los Angeles<o:p></o:p> <o:p></o:p> <o:p></o:p> We extend the omitted variable bias framework with a suite of tools for sensitivity analysis in regression models that: (i) does not require assumptions about the treatment assignment nor the nature of confounders; (ii) naturally handles multiple confounders, possibly acting non-linearly; (iii) exploits expert knowledge to bound sensitivity parameters; and, (iv) can be easily computed using only standard regression results. In particular, we introduce two novel sensitivity measures suited for routine reporting. The robustness value describes the minimum strength of association unobserved confounding would need to have, both with the treatment and the outcome, to change the research conclusions. The partial R2of the treatment with the outcome shows how strongly confounders explaining all the residual outcome variation would have to be associated with the treatment to eliminate the estimated eﬀect. Next, we oﬀer graphical tools for elaborating on problematic confounders, examining the sensitivity of point estimates, t-values, as well as “extreme scenarios”. Finally, we describe problems with a common “benchmarking” practice and introduce a novel procedure to formally bound the strength of confounders based on comparison to observed covariates. We apply these methods to a running example that estimates the eﬀect of exposure to violence on attitudes towards peace.<o:p></o:p> <o:p></o:p> Carlos Cinelli é atualmente doutorando em Estatística na Universidade da California, Los Angeles. Carlos também é mestre em Economia pela Universidade de Brasília, e trabalhou como economista no Banco Central do Brasil. Seu foco de pesquisa é o desenvolvimento de métodos para inferência causal com dados observacionais.<o:p></o:p> <o:p></o:p> <a href="https://tinyurl.com/y6zopkos">https://tinyurl.com/y6zopkos</a><o:p></o:p> Data: 03/05/2019 Horário: 12h00 Local: Sala 308 -3° andar - Insper <img border="0" width="19" height="19" style="width:.1979in;height:.1979in" id="Imagem_x0020_2" src="cid:image007.png@01D4FE8E.D2497E70" alt="cid:image002.png@01D4C861.DA6A4ED0"> Rua Quatá, 300 – Vila Olímpia, São Paulo/SP Estacionamento: Rua Uberabinha, s/n<o:p></o:p> <o:p></o:p> </td> </tr> </tbody> </table> </div> <o:p></o:p> <div align="center"> <table class="MsoNormalTable" border="0" cellspacing="0" cellpadding="0" width="600" style="width:450.0pt"> <tbody> <tr> <td style="padding:0cm 0cm 0cm 0cm"> <a href="http://facebook.com/insper" target="_blank"><img border="0" width="83" height="154" style="width:.8645in;height:1.6041in" id="Imagem_x0020_3" src="cid:image008.png@01D4FE8E.D2497E70" alt="cid:image003.png@01D4C861.DA6A4ED0"></a><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <a href="http://www.youtube.com/insperedu?gl=BR&hl=pt" target="_blank"><img border="0" width="35" height="154" style="width:.3645in;height:1.6041in" id="Imagem_x0020_4" src="cid:image009.png@01D4FE8E.D2497E70" alt="cid:image004.png@01D4C861.DA6A4ED0"></a><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <a href="http://twitter.com/insper" target="_blank"><img border="0" width="35" height="154" style="width:.3645in;height:1.6041in" id="Imagem_x0020_5" src="cid:image010.png@01D4FE8E.D2497E70" alt="cid:image005.png@01D4C861.DA6A4ED0"></a><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <a href="http://www.instagram.com/insperedu" target="_blank"><img border="0" width="35" height="154" style="width:.3645in;height:1.6041in" id="Imagem_x0020_6" src="cid:image011.png@01D4FE8E.D2497E70" alt="cid:image006.png@01D4C861.DA6A4ED0"></a><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <a href="http://www.linkedin.com/company/1134427" target="_blank"><img border="0" width="35" height="154" style="width:.3645in;height:1.6041in" id="Imagem_x0020_7" src="cid:image012.png@01D4FE8E.D2497E70" alt="cid:image007.png@01D4C861.DA6A4ED0"></a><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <img border="0" width="78" height="154" style="width:.8125in;height:1.6041in" id="Imagem_x0020_15" src="cid:image013.png@01D4FE8E.D2497E70" alt="cid:image008.png@01D4C861.DA6A4ED0"><o:p></o:p> </td> <td style="padding:0cm 0cm 0cm 0cm"> <img src="cid:60B59A37-2EE3-4327-B66C-06B8678EA05C" alt="image014.png" id="49F0CCDE-7FA3-451B-B6D0-11E08D8C960C" x-apple-mail-has-final-size="NO" style="opacity: 1 !important;"><a href="http://www.insper.edu.br/" target="_blank"></a><o:p></o:p> </td> </tr> </tbody> </table> </div> <o:p></o:p> <div align="center"> <table class="MsoNormalTable" border="0" cellspacing="0" cellpadding="0" width="600" style="width:450.0pt"> <tbody> <tr> <td style="background:white;padding:7.5pt 7.5pt 7.5pt 7.5pt"> É de conhecimento do participante que eventuais filmagens e/ou fotografias realizadas no evento poderão ser utilizadas em qualquer meio de mídia para uso institucional.<o:p></o:p> </td> </tr> <tr> <td style="background:#C00026;padding:0cm 0cm 0cm 0cm"> <o:p></o:p> </td> </tr> </tbody> </table> </div> <o:p></o:p> <o:p></o:p> </div> </div></blockquote><style></style></body></html>