View source for Homoscedasticity and heteroscedasticity

{{Short description|Statistical property}}
[[File:Homoscedasticity.png|thumb|right|Plot with random data showing homoscedasticity: at each value of ''x'', the ''y''-value of the dots has about the same [[variance]].]]
[[File:Heteroscedasticity.png|thumb|right|Plot with random data showing heteroscedasticity: The variance of the ''y''-values of the dots increases with increasing values of ''x''.]]

In [[statistics]], a [[sequence]] of [[random variable]]s is '''homoscedastic''' ({{IPAc-en|ˌ|h|oʊ|m|oʊ|s|k|ə|ˈ|d|æ|s|t|ɪ|k}}) if all its random variables have the same finite [[variance]]; this is also known as '''homogeneity of variance'''. The complementary notion is called '''heteroscedasticity''', also known as '''heterogeneity of variance'''. The spellings ''homos'''k'''edasticity'' and ''heteros'''k'''edasticity'' are also frequently used. “Skedasticity” comes from the Ancient Greek word “skedánnymi”, meaning “to scatter”.<ref>For the Greek etymology of the term, see {{cite journal |first=J. Huston |last=McCulloch |year=1985 |title=On Heteros*edasticity |journal=[[Econometrica]] |volume=53 |issue=2 |page=483 |jstor=1911250 }}</ref><ref name=White>
 {{cite journal
  | doi = 10.2307/1912934
  | title = A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity
  | year = 1980
  | author = White, Halbert
  | journal = Econometrica
  | volume = 48 | issue = 4 | pages = 817–838
  | jstor = 1912934
  | citeseerx = 10.1.1.11.7646
 }}</ref><ref name=Gujarati>
 {{cite book
  | last1=Gujarati | first1=D. N.
  | last2=Porter | first2=D. C.|author2-link=Dawn C. Porter
  | year=2009
  | title=Basic Econometrics
  | edition=Fifth
  | publisher=McGraw-Hill Irwin
  | location=Boston
  | isbn=9780073375779
  | page=400
 }}</ref>
Assuming a  variable is homoscedastic when in reality it is heteroscedastic ({{IPAc-en|ˌ|h|ɛ|t|ər|oʊ|s|k|ə|ˈ|d|æ|s|t|ɪ|k}}) results in [[Biased estimator|unbiased]] but [[Efficiency (statistics)|inefficient]] [[point estimate]]s and in biased estimates of [[standard error]]s, and may result in overestimating the [[goodness of fit]] as measured by the [[Pearson product-moment correlation coefficient|Pearson coefficient]].

The existence of heteroscedasticity is a major concern in  [[regression analysis]] and the [[analysis of variance]], as it invalidates [[statistical hypothesis testing|statistical tests of significance]] that assume that the [[errors and residuals in statistics|modelling errors]] all have the same variance. While the [[ordinary least squares]] estimator is still unbiased in the presence of heteroscedasticity, it is inefficient and inference based on the assumption of homoskedasticity is misleading. In that case, [[generalized least squares]] (GLS) was frequently used in the past.<ref>{{cite book |last=Goldberger |first=Arthur S. |url=https://archive.org/details/econometrictheor0000gold |title=Econometric Theory |publisher=John Wiley & Sons |year=1964 |location=New York |pages=[https://archive.org/details/econometrictheor0000gold/page/238 238–243] |isbn=9780471311010 |author-link=Arthur Goldberger |url-access=registration}}</ref><ref>{{cite book |last=Johnston |first=J. |title=Econometric Methods |publisher=McGraw-Hill |year=1972 |location=New York |pages=214–221 |author-link=John Johnston (econometrician)}}</ref> Nowadays, standard practice in econometrics is to include [[Heteroskedasticity-consistent standard errors]] instead of using GLS, as GLS can exhibit strong bias in small samples if the actual [[skedastic function]] is unknown.<ref name=":0">{{Cite book |last=Angrist |first=Joshua D. |url=https://www.degruyter.com/document/doi/10.1515/9781400829828/html |title=Mostly Harmless Econometrics: An Empiricist's Companion |last2=Pischke |first2=Jörn-Steffen |date=2009-12-31 |publisher=Princeton University Press |isbn=978-1-4008-2982-8 |doi=10.1515/9781400829828}}</ref>

Because heteroscedasticity concerns [[Expected value|expectations]] of the second [[Moment (mathematics)|moment]] of the errors, its presence is referred to as [[Statistical model specification|misspecification]] of the second order.<ref>{{cite book |last1=Long |first1=J. Scott |title=Testing Structural Equation Models |last2=Trivedi |first2=Pravin K. |publisher=Sage |year=1993 |isbn=978-0-8039-4506-7 |editor-last=Bollen |editor-first=Kenneth A. |location=London |pages=66–110 |chapter=Some Specification Tests for the Linear Regression Model |author1-link=J. Scott Long |editor2-last=Long |editor2-first=J. Scott}}</ref>

The [[econometrician]] [[Robert Engle]] was awarded the 2003 [[Nobel Memorial Prize for Economics]] for his studies on [[regression analysis]] in the presence of heteroscedasticity, which led to his formulation of the [[autoregressive conditional heteroscedasticity]] (ARCH) modeling technique.<ref>{{Cite journal |last=Engle |first=Robert F. |date=July 1982 |title=Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation |journal=Econometrica |volume=50 |issue=4 |pages=987–1007 |doi=10.2307/1912773 |issn=0012-9682 |jstor=1912773}}</ref>

== Definition ==
Consider the [[simple linear regression|linear regression]] equation <math>y_i= x_i \beta_i + \varepsilon_i,\ i =  1,\ldots, N,</math> where the dependent random variable <math>y_i</math> equals the deterministic variable <math>x_i</math> times coefficient <math>\beta_i</math> plus a random disturbance term <math>\varepsilon_i</math> that has mean zero. The disturbances are homoscedastic if the variance of <math>\varepsilon_i</math> is a constant <math>\sigma^2</math>; otherwise, they are heteroscedastic. In particular, the disturbances are heteroscedastic if the variance of <math>\varepsilon_i</math> depends on <math>i</math> or on the value of <math>x_i</math>. One way they might be heteroscedastic is if <math>\sigma_i^2= x_i \sigma^2</math> (an example of a [[scedastic function]]), so the variance is proportional to the value of <math>x</math>.

More generally, if the variance-covariance matrix of disturbance <math>\varepsilon_i</math> across <math>i</math> has a nonconstant diagonal, the disturbance is heteroscedastic.<ref>Peter Kennedy, ''A Guide to Econometrics'', 5th edition, p. 137.</ref>  The matrices below are covariances when there are just three observations across time. The disturbance in matrix A is homoscedastic; this is the simple case where OLS is the best linear unbiased estimator. The disturbances in matrices B and C are heteroscedastic. In matrix B, the variance is time-varying, increasing steadily across time; in   matrix C, the variance depends on the value of <math>x</math>. The disturbance in matrix D is homoscedastic because the diagonal variances are constant, even though the off-diagonal covariances are non-zero and ordinary least squares is inefficient for a different reason: serial correlation.
:<math>\begin{align}
A &= \sigma^2\begin{bmatrix}
  1 & 0 & 0  \\ 
  0 & 1 & 0  \\ 
  0 & 0 & 1  \\ 
\end{bmatrix} &
B &= \sigma^2\begin{bmatrix}
  1 & 0 & 0  \\ 
  0 & 2 & 0  \\ 
  0 & 0 & 3  \\ 
\end{bmatrix} &
C &= \sigma^2\begin{bmatrix}
  x_1 &   0 &   0  \\
    0 & x_2 &   0  \\ 
    0 &   0 & x_3  \\ 
\end{bmatrix} &
D &= \sigma^2\begin{bmatrix}
       1 & \rho & \rho^2  \\ 
    \rho &    1 &   \rho  \\ 
  \rho^2 & \rho &      1  \\
\end{bmatrix}
\end{align}</math>

== Examples ==
Heteroscedasticity often occurs when there is a large difference among the sizes of the observations.

A classic example of heteroscedasticity is that of income versus expenditure on meals. A wealthy person may eat inexpensive food sometimes and expensive food at other times. A poor person will almost always eat inexpensive food. Therefore, people with higher incomes exhibit greater variability in expenditures on food.

At a rocket launch, an observer measures the distance traveled by the rocket once per second. In the first couple of seconds, the measurements may be accurate to the nearest centimeter. After five minutes, the accuracy of the measurements may be good only to 100&nbsp;m, because of the increased distance, atmospheric distortion, and a variety of other factors. So the measurements of distance may exhibit heteroscedasticity.

==Consequences==
One of the assumptions of the classical linear regression model is that there is no heteroscedasticity. Breaking this assumption means that the [[Gauss–Markov theorem]] does not apply, meaning that [[Ordinary least squares|OLS]] estimators are not the [[Best linear unbiased estimator|Best Linear Unbiased Estimators (BLUE)]] and their variance is not the lowest of all other unbiased estimators.
Heteroscedasticity does ''not'' cause ordinary least squares coefficient estimates to be biased, although it can cause ordinary least squares estimates of the variance (and, thus, standard errors) of the coefficients to be biased, possibly above or below the true of population variance. Thus, regression analysis using heteroscedastic data will still provide an unbiased estimate for the relationship between the predictor variable and the outcome, but standard errors and therefore inferences obtained from data analysis are suspect. Biased standard errors lead to biased inference, so results of hypothesis tests are possibly wrong. For example, if OLS is performed on a heteroscedastic data set, yielding biased standard error estimation, a researcher might fail to reject a null hypothesis at a given [[Statistical significance|significance]] level, when that null hypothesis was actually uncharacteristic of the actual population (making a [[Type I and type II errors|type II error]]).

Under certain assumptions, the OLS estimator has a normal [[asymptotic distribution]] when properly normalized and centered (even when the data does not come from a [[normal distribution]]). This result is used to justify using a normal distribution, or a [[chi square distribution]] (depending on how the [[test statistic]] is calculated), when conducting a [[hypothesis test]]. This holds even under heteroscedasticity. More precisely, the OLS estimator in the presence of heteroscedasticity is asymptotically normal, when properly normalized and centered, with a variance-covariance [[Matrix (mathematics)|matrix]] that differs from the case of homoscedasticity. In 1980, White proposed a [[consistent estimator]] for the variance-covariance matrix of the asymptotic distribution of the OLS estimator.<ref name="White" /> This validates the use of hypothesis testing using OLS estimators and White's variance-covariance estimator under heteroscedasticity.

Heteroscedasticity is also a major practical issue encountered in [[ANOVA]] problems.<ref>{{cite journal |last1=Jinadasa |first1=Gamage |last2=Weerahandi |first2=Sam |year=1998 |title=Size performance of some tests in one-way anova |journal=Communications in Statistics - Simulation and Computation |volume=27 |issue=3 |pages=625 |doi=10.1080/03610919808813500}}</ref>
The [[F test]] can still be used in some circumstances.<ref>{{cite journal |author=Bathke, A |year=2004 |title=The ANOVA F test can still be used in some balanced designs with unequal variances and nonnormal data |journal=Journal of Statistical Planning and Inference |volume=126 |issue=2 |pages=413–422 |doi=10.1016/j.jspi.2003.09.010}}</ref>

However, it has been said that students in [[econometrics]] should not overreact to heteroscedasticity.<ref name="Gujarati" /> One author wrote, "unequal error variance is worth correcting only when the problem is severe."<ref>{{cite book |last=Fox |first=J. |title=Applied Regression Analysis, Linear Models, and Related Methods |publisher=Sage Publications |year=1997 |location=California |page=306}} (Cited in Gujarati et al. 2009, p.&nbsp;400)</ref> In addition, another word of caution was in the form, "heteroscedasticity has never been a reason to throw out an otherwise good model."<ref name="Gujarati" /><ref>{{cite journal |last=Mankiw |first=N. G. |author-link=N. Gregory Mankiw |year=1990 |title=A Quick Refresher Course in Macroeconomics |journal=[[Journal of Economic Literature]] |volume=28 |issue=4 |pages=1645–1660 [p. 1648] |doi=10.3386/w3256 |jstor=2727441 |doi-access=free}}</ref> With the advent of [[heteroscedasticity-consistent standard errors]] allowing for inference without specifying the conditional second moment of error term, testing conditional homoscedasticity is not as important as in the past.<ref name=":0" />

For any non-linear model (for instance [[Logit]] and [[Probit]] models), however, heteroscedasticity has more severe consequences: the [[maximum likelihood estimation|maximum likelihood estimates]] (MLE) of the parameters will usually be biased, as well as inconsistent (unless the likelihood function is modified to correctly take into account the precise form of heteroscedasticity or the distribution is a member of the [[Exponential family|linear exponential family]] and the conditional expectation function is correctly specified).<ref>{{cite web |last=Giles |first=Dave |date=May 8, 2013 |title=Robust Standard Errors for Nonlinear Models |url=http://davegiles.blogspot.com/2013/05/robust-standard-errors-for-nonlinear.html |work=Econometrics Beat}}</ref><ref>{{Cite journal |last=Gourieroux |first=C. |last2=Monfort |first2=A. |last3=Trognon |first3=A. |date=1984 |title=Pseudo Maximum Likelihood Methods: Theory |url=https://www.jstor.org/stable/1913471 |journal=Econometrica |volume=52 |issue=3 |pages=681–700 |doi=10.2307/1913471 |issn=0012-9682}}</ref> Yet, in the context of binary choice models ([[Logit]] or [[Probit]]), heteroscedasticity will only result in a positive scaling effect on the asymptotic mean of the misspecified MLE (i.e. the model that ignores heteroscedasticity).<ref>{{cite journal |last1=Ginker |first1=T. |last2=Lieberman |first2=O. |year=2017 |title=Robustness of binary choice models to conditional heteroscedasticity |journal=Economics Letters |volume=150 |pages=130–134 |doi=10.1016/j.econlet.2016.11.024}}</ref> As a result, the predictions which are based on the misspecified MLE will remain correct. In addition, the misspecified Probit and Logit MLE will be asymptotically normally distributed which allows performing the usual significance tests (with the appropriate variance-covariance matrix). However, regarding the general hypothesis testing, as pointed out by [[William Greene (economist)|Greene]], "simply computing a robust covariance matrix for an otherwise inconsistent estimator does not give it redemption. Consequently, the virtue of a robust covariance matrix in this setting is unclear."<ref>{{cite book |last=Greene |first=William H. |title=Econometric Analysis |publisher=Pearson Education |year=2012 |isbn=978-0-273-75356-8 |edition=Seventh |location=Boston |pages=730–755 [p. 733] |chapter=Estimation and Inference in Binary Choice Models |author-link=William Greene (economist) |chapter-url=https://books.google.com/books?id=-WFPYgEACAAJ&pg=PA733}}</ref>

== Correction ==
There are several common corrections for heteroscedasticity. They are:
* A stabilizing transformation of the data, e.g. [[Logarithmic scale|logarithmized]] data. Non-logarithmized series that are growing exponentially often appear to have increasing variability as the series rises over time. The variability in percentage terms may, however, be rather stable.
* Use a different specification for the model (different ''X'' variables, or perhaps non-linear transformations of the ''X'' variables).
* Apply a [[weighted least squares]] estimation method, in which OLS is applied to transformed or weighted values of ''X'' and ''Y''. The weights vary over observations, usually depending on the changing error variances. In one variation the weights are directly related to the magnitude of the dependent variable, and this corresponds to least squares percentage regression.<ref>{{cite journal |author=Tofallis, C |year=2008 |title=Least Squares Percentage Regression |journal=Journal of Modern Applied Statistical Methods |volume=7 |pages=526–534 |doi=10.2139/ssrn.1406472 |ssrn=1406472|url=https://digitalcommons.wayne.edu/jmasm/vol7/iss2/18 }}</ref>
* [[Heteroscedasticity-consistent standard errors]] (HCSE), while still biased, improve upon OLS estimates.<ref name="White" /> HCSE is a consistent estimator of standard errors in regression models with heteroscedasticity. This method corrects for heteroscedasticity without altering the values of the coefficients. This method may be superior to regular OLS because if heteroscedasticity is present it corrects for it, however, if the data is homoscedastic, the standard errors are equivalent to conventional standard errors estimated by OLS. Several modifications of the White method of computing heteroscedasticity-consistent standard errors have been proposed as corrections with superior finite sample properties.
* [[Wild bootstrapping]] can be used as a [[Resampling (statistics)|Resampling method]] that respects the differences in the conditional variance of the error term. An alternative is resampling observations instead of errors. Note resampling errors without respect for the affiliated values of the observation enforces homoskedasticity and thus yields incorrect inference.
* Use [[MINQUE]] or even the customary estimators <math display="inline">s_i^2 = (n_i - 1)^{-1} \sum_j \left(y_{ij} - \bar{y}_i\right)^2</math> (for <math>i=1,2,...,k</math> independent samples with <math>j=1, 2, ..., n_i</math> observations each), whose efficiency losses are not substantial when the number of observations per sample is large (<math>n_i > 5</math>), especially for small number of independent samples.<ref>{{cite journal |author=J. N. K. Rao |date=March 1973 |title=On the Estimation of Heteroscedastic Variances |journal=Biometrics |volume=29 |pages=11–24 |doi=10.2307/2529672 |jstor=2529672 |number=1}}</ref>

==Testing==
[[File:Hsked residual compare.svg|thumb|right|Absolute value of residuals for simulated first order heteroscedastic data]]
Residuals can be tested for homoscedasticity using the [[Breusch–Pagan test]],<ref>{{Cite journal|last1=Breusch|first1=T. S.|last2=Pagan|first2=A. R.|date=1979|title=A Simple Test for Heteroscedasticity and Random Coefficient Variation|url=https://www.jstor.org/stable/1911963|journal=Econometrica|volume=47|issue=5|pages=1287–1294|doi=10.2307/1911963|jstor=1911963|issn=0012-9682}}</ref> which performs an auxiliary regression of the squared residuals on the independent variables. From this auxiliary regression, the explained sum of squares is retained, divided by two, and then becomes the test statistic for a chi-squared distribution with the degrees of freedom equal to the number of independent variables.<ref>{{Cite web|last=Ullah|first=Muhammad Imdad|date=2012-07-26|title=Breusch Pagan Test for Heteroscedasticity|url=https://itfeature.com/correlation-and-regression-analysis/ols-assumptions/breusch-pagan-test|access-date=2020-11-28|website=Basic Statistics and Data Analysis|language=en-US}}</ref> The null hypothesis of this chi-squared test is homoscedasticity, and the alternative hypothesis would indicate heteroscedasticity. Since the Breusch–Pagan test is sensitive to departures from normality or small sample sizes, the Koenker–Bassett or 'generalized Breusch–Pagan' test is commonly used instead.<ref name=":1" />{{Additional citation needed|date=November 2020}} From the auxiliary regression, it retains the R-squared value which is then multiplied by the sample size, and then becomes the test statistic for a chi-squared distribution (and uses the same degrees of freedom). Although it is not necessary for the Koenker–Bassett test, the Breusch–Pagan test requires that the squared residuals also be divided by the residual sum of squares divided by the sample size.<ref name=":1">{{cite web|last1=Pryce|first1=Gwilym|title=Heteroscedasticity: Testing and Correcting in SPSS|url=http://reocities.com/Heartland/4205/SPSS/HeteroscedasticityTestingAndCorrectingInSPSS1.pdf|url-status=live|archive-url=https://web.archive.org/web/20170327080250/http://www.reocities.com/Heartland/4205/SPSS/HeteroscedasticityTestingAndCorrectingInSPSS1.pdf|archive-date=2017-03-27|access-date=26 March 2017|pages=12–18}}</ref> Testing for groupwise heteroscedasticity can be done with the [[Goldfeld–Quandt test]].<ref>{{Cite journal |last=Baum |first=Christopher F. |date=2006 |title=Stata Tip 38: Testing for Groupwise Heteroskedasticity |journal=The Stata Journal: Promoting Communications on Statistics and Stata |language=en |volume=6 |issue=4 |pages=590–592 |doi=10.1177/1536867X0600600412 |s2cid=117349246 |issn=1536-867X|doi-access=free }}</ref>

Due to the standard use of [[Heteroskedasticity-consistent standard errors|heteroskedasticity-consistent Standard Errors]] and the problem of [[Pre-test]], econometricians nowadays rarely use tests for conditional heteroskedasticity.<ref name=":0" />

=== List of tests ===
Although tests for heteroscedasticity between groups can formally be considered as a special case of testing within regression models, some tests have structures specific to this case.

{{Columns-start}}
; Tests in regression
* [[Goldfeld–Quandt test]]
* [[Park test]]<ref>{{cite journal |author=R. E. Park |year=1966 |title=Estimation with Heteroscedastic Error Terms |journal=Econometrica |volume=34 |issue=4 |pages=888 |doi=10.2307/1910108 |jstor=1910108}}</ref>
* [[Glejser test]]<ref>{{cite journal |last=Glejser |first=H. |year=1969 |title=A new test for heteroscedasticity |journal=[[Journal of the American Statistical Association]] |volume=64 |issue=325 |pages=316–323 |doi=10.1080/01621459.1969.10500976}}</ref><ref>{{Cite journal |last1=Machado |first1=José A. F. |last2=Silva |first2=J. M. C. Santos |year=2000 |title=Glejser's test revisited |journal=[[Journal of Econometrics]] |volume=97 |issue=1 |pages=189–202 |doi=10.1016/S0304-4076(00)00016-6}}</ref>
* [[Harrison–McCabe test]]
* [[Breusch–Pagan test]]
* [[White test]]<ref name="White" />
* [[Cook–Weisberg test]]
{{column}}
; Tests for grouped data
* [[F-test of equality of variances]]
* [[Cochran's C test]]
* [[Hartley's test]]
* [[Bartlett's test]]
* [[Levene's test]]
* [[Brown–Forsythe test]]{{columns-end}}

==Generalisations==

===Homoscedastic distributions===

Two or more [[normal distribution]]s, <math>N(\mu_1,\Sigma_1), N(\mu_2,\Sigma_2), </math> are both homoscedastic and lack [[serial correlation]] if they share the same diagonals in their [[covariance matrix|covariance]] matrix, <math>\Sigma_1{ii} = \Sigma_2{jj},\ \forall i=j.</math> and their non-diagonal entries are zero. Homoscedastic distributions are especially useful to derive statistical [[pattern recognition]] and [[machine learning]] algorithms. One popular example of an algorithm that assumes homoscedasticity is Fisher's [[linear discriminant analysis]].
The concept of homoscedasticity can be applied to distributions on spheres.<ref>Hamsici, Onur C.; Martinez, Aleix M. (2007) [http://jmlr.csail.mit.edu/papers/volume8/hamsici07a/hamsici07a.pdf "Spherical-Homoscedastic Distributions: The Equivalency of Spherical and Normal Distributions in Classification"], ''Journal of Machine Learning Research'', 8, 1583-1623</ref>

===Multivariate data===
The study of homescedasticity and heteroscedasticity has been generalized to the multivariate case, which deals with the covariances of vector observations instead of the variance of scalar observations. One version of this is to use covariance matrices as the multivariate measure of dispersion. Several authors have considered tests in this context, for both regression and grouped-data situations.<ref>{{Cite journal | last1 = Holgersson | first1 = H. E. T. | last2 = Shukur | first2 = G. | doi = 10.1080/00949650410001646979 | title = Testing for multivariate heteroscedasticity | journal = Journal of Statistical Computation and Simulation | volume = 74 | issue = 12 | pages = 879 | year = 2004 | hdl = 2077/24416 | s2cid = 121576769 | hdl-access = free }}</ref><ref>{{Cite journal
| last1 = Gupta | first1 = A. K.
| last2 = Tang | first2 = J.
| doi = 10.1093/biomet/71.3.555
| title = Distribution of likelihood ratio statistic for testing equality of covariance matrices of multivariate Gaussian models
| journal = Biometrika
| volume = 71
| issue = 3
| pages = 555–559
| year = 1984
| jstor=2336564
}}</ref> [[Bartlett's test]] for heteroscedasticity between grouped data, used most commonly in the univariate case, has also been extended for the multivariate case, but a tractable solution only exists for 2 groups.<ref>{{Cite book | last1 = d'Agostino | first1 = R. B. | last2 = Russell | first2 = H. K. | doi = 10.1002/0470011815.b2a13048 | chapter = Multivariate Bartlett Test | title = Encyclopedia of Biostatistics | year = 2005 | isbn = 978-0470849071 }}</ref> Approximations exist for more than two groups, and they are both called [[Box's M test]].

==See also==
* [[Heterogeneous#Statistics|Heterogeneity]]
* [[Spherical error]]
* [[Heteroskedasticity-consistent standard errors]]

==References==
{{Reflist|30em}}

==Further reading==
Most statistics textbooks will include at least some material on homoscedasticity and heteroscedasticity. Some examples are:
* {{cite book |last1=Asteriou |first1=Dimitros |last2=Hall |first2=Stephen G. |title=Applied Econometrics |publisher=Palgrave MacMillan |year=2011 |edition=Second |isbn=978-0-230-27182-1 |pages=109–147 |url=https://books.google.com/books?id=6qYcBQAAQBAJ&pg=PA109 }}
* {{cite book |first1=Russell |last1=Davidson |first2=James G. |last2=MacKinnon |title=Estimation and Inference in Econometrics |location=New York |publisher=Oxford University Press |year=1993 |isbn=978-0-19-506011-9 |pages=547–582 |url=https://books.google.com/books?id=Ot6DByCF6osC&pg=PA547 }}
* {{cite book |last=Dougherty |first=Christopher |year=2011 |title=Introduction to Econometrics |pages=280–299 |location=New York |publisher=Oxford University Press |isbn=978-0-19-956708-9 |url=https://books.google.com/books?id=UXucAQAAQBAJ&pg=PA280 }}
* {{cite book |last1=Gujarati |first1=Damodar N. |author-link=Damodar N. Gujarati |last2=Porter |first2=Dawn C.|author2-link=Dawn C. Porter |title=Basic Econometrics |location=New York |publisher=McGraw-Hill Irwin |edition=Fifth |year=2009 |pages=365–411 |isbn=978-0-07-337577-9 |url=https://books.google.com/books?id=6l1CPgAACAAJ&pg=PA365 }}
* {{cite book |last=Kmenta |first=Jan |author-link=Jan Kmenta |title=Elements of Econometrics |location=New York |publisher=Macmillan |year=1986 |edition=Second |isbn=978-0-02-365070-3 |pages=[https://archive.org/details/elementsofeconom0003kmen/page/269 269]–298 |url=https://archive.org/details/elementsofeconom0003kmen |url-access=registration }}
* {{cite book |last1=Maddala |first1=G. S. |author-link=G. S. Maddala |last2=Lahiri |first2=Kajal |pages=211–238 |title=Introduction to Econometrics |location=New York |publisher=Wiley |year=2009 |edition=Fourth |isbn=978-0-470-01512-4 |url=https://books.google.com/books?id=vkQvQgAACAAJ&pg=PA211 }}

==External links==
* {{YouTube|id=TfuqBxRgRTU&list=PLD15D38DC7AA3B737&index=3#t=47s|title=Econometrics lecture (topic: heteroscedasticity)}} by [[Mark Thoma]]

{{statistics}}

[[Category:Statistical deviation and dispersion]]
[[Category:Regression analysis]]

[[ja:ARCHモデル#分散不均一性]]