Editing Econometrics (section)

==Limitations and criticisms==
{{see also|Criticisms of econometrics}}

Like other forms of statistical analysis, badly specified econometric models may show a [[spurious relationship]] where two variables are correlated but causally unrelated. In a study of the use of econometrics in major economics journals, [[Deidre McCloskey|McCloskey]] concluded that some economists report [[p-value|''p''-value]]s (following the [[Ronald Fisher|Fisherian]] tradition of [[tests of significance]] of point [[null hypothesis|null-hypotheses]]) and neglect concerns of [[type II error]]s; some economists fail to report estimates of the size of effects (apart from [[statistical significance]]) and to discuss their economic importance. She also argues that some economists also fail to use economic reasoning for [[model selection]], especially for deciding which variables to include in a regression.<ref>{{cite journal |title=The Loss Function has been mislaid: the Rhetoric of Significance Tests|last=McCloskey|journal=American Economic Review|date=May 1985|volume=75|issue=2}}</ref><ref>[[Stephen Ziliak|Stephen T. Ziliak]] and [[Deirdre McCloskey|Deirdre N. McCloskey]] (2004). "Size Matters: The Standard Error of Regressions in the ''American Economic Review''", ''Journal of Socio-Economics'', 33(5), pp. [http://faculty.roosevelt.edu/Ziliak/doc/Size%20Matters%20Journal%20of%20Socio-Economics%20Ziliak%20and%20McCloskey.pdf 527-46] {{webarchive|url=https://web.archive.org/web/20100625163446/http://faculty.roosevelt.edu/Ziliak/doc/Size%20Matters%20Journal%20of%20Socio-Economics%20Ziliak%20and%20McCloskey.pdf |date=25 June 2010 }} (press '''+''').</ref>

In some cases, economic variables cannot be experimentally manipulated as treatments randomly assigned to subjects.<ref name="Leamer 31–43">{{cite journal|last=Leamer|first=Edward|title=Let's Take the Con out of Econometrics|journal=American Economic Review|date=March 1983|volume=73|issue=1|pages=31–43|jstor=1803924}}</ref> In such cases, economists rely on [[observational studies]], often using data sets with many strongly associated [[covariate]]s, resulting in enormous numbers of models with similar explanatory ability but different covariates and regression estimates. Regarding the plurality of models compatible with observational data-sets, [[Edward Leamer]] urged that "professionals ... properly withhold belief until an inference can be shown to be adequately insensitive to the choice of assumptions".<ref name="Leamer 31–43"/>