Mostly Harmless Econometrics: An Empiricist's Companion: Joshua D. Angrist 「ほとんど無害」な計量経済学―応用経済学のための実証分析ガイド ヨシュア・アングリスト 322頁~
ACRONYMS AND ABBREVIATIONS 頭字語と略語 TECHNICAL TERMS 技術的用語
2SLS Two-stage least squares, an instrumental variables (IV) estimator. 2段階最小二乗法,操作変数(IV)法. ACR Average causal response, the weighted average causal response to an ordered treatment. 平均因果的反応,順序付けのなされた処置に対する加重平均因果効果. ANOVA Analysis of variance, a decomposition of total variance into the variance of the conditional expectation function (CEF) and the average conditional variance. 分散分析,総分散の条件付き期待値関数(CEF)の分散と平均条件付き分散への分解. BRL Biased reduced linearization estimator, a bias-corrected covariance matrix estimator for clustered data. バイアス補償線形化推定量,クラスターデータに対するバイアス修正済み共分散行列推定量. CDF Cumulative distribution function, the probability that a random variable takes on a value less than or equal to a given number. 累積密度関数,確率変数が所与の数値以下の値をとる確率. CEF Conditional expectation function, the population average of Yi with Xi held fixed. 条件付き期待値関数,Xiを所与とした Yiの母集団平均. CIA Conditional independence assumption, a core assumption that justifies a causal interpretation of regression and matching estimators. 条件付き独立の仮定,回帰とマッチング推定量の因果的解釈を正当化する中核的仮定. COP Conditional on positive effect, the treatment-control difference in means for a non-negative random variable looking at positive values only. 正の効果の条件付き,正の値のみを見る非負の確率変数の平均の処置群・対照群間差. CQF Conditional quantile function, defined for each quantile τ, the τ-quantile of Yi, holding Xi fixed. 条件付き分位点関数,Xiを所与とし,Yiのτ-分位点の各分位点τについて定義される. DD Differences-in-differences estimator. In its simplest form, a comparison of changes over time in treatment and control groups. 差分の差分推定量,最も単純な形では処置群と対照群の間の時間を追っての比較. GLS Generalized least squares estimator, a regression estimator for models with heteroskedasticity and/or serial correlation. GLS provides efficiency gains when the conditional expectation function (CEF) is linear. 一般化最小二乗法,不均一分散及び,または系列相関を持つモデルの回帰推定量.GLS は条件付き期待値関数(CEF)が緑形のとき効率性の向上を提供する. GMM Generalized method of moments, an econometric estimation framework in which estimates are chosen to minimize a matrix-weighted average of the squared difference between sample and population moments. 一般化積率法,標本と母集団の積率の差の二乗の行列による加重平均値を最小化するように推定値が選択される計量経済学的推定法.
HC0-HC3 Heteroskedasticity consistent covariance matrix estimators discussed by MacKinnon and White (1985). MacKinnon and White(1985) によって議論された不均一分散一致共分散行列推定量. ILS Indirect least squares estimator, the ratio of reduced-form to first-stage coefficients in an instrumental variables (IV) setup. 間接最小二乗推定量,操作変数(IV) 設定に於ける誘導形と第1段階係数の比. ITT Intention to treat effect, the effect of being offered treatment. 処置の意志の効果,処置を提供されることの効果. IV Instrumental variables estimator or method. 操作変数推定量,または,推定法. JIVE Jackknife instrumental variables (IV) estimator. ジャックナイフ操作変数(IV) 推定量. LATE Local average treatment effect, the causal effect of treatment on compliers. 局所的平均処置効果,順応者に対する処置の因果効果. LDVs Limited dependent variables, such as dummies, counts, and non-negative random variables on the left-hand side of regression and related statistical models. 回帰の左辺にあるダミー変数,個数,非負の確率変数といった制限被説明変数と関連する統計モデル. LIML Limited information maximum likelihood estimator, an alternative to two-stage least squares (2SLS) with less bias. 制限情報最尤推定量,バイアスのより少ない,2段階最小二乗法(2SLS)の他の選択肢. LM Lagrange multiplier test, a statistical test of the restrictions imposed by an estimator. ラグランジュ乗数検定,推定量により課せられた制約の統計的検定. LPM Linear probability model, a linear regression model for a dummy dependent variable. 線形確率モデル,ダミー被説明変数のための線形回帰モデル. MFX Marginal effects. In nonlinear models, the derivative of the conditional expectation function (CEF) implied by the model with respect to the regressors. 限界効果,非線形モデルでは,モデルにより示唆される条件付き期待値関数(CEF)の説明変数に関する微分. MMSE Minimum mean squared error, the minimum expected squared prediction error, or the minimum of the expected square of the difference between an estimator and a target. 最小平均二乗誤差,最小期待二乗予測誤差,または,推定量と目標の間の差の二乗の期待値の最小値. OLS Ordinary least squares estimator, the sample analog of the population regression vector. 最小二乗推定量,母集団回帰ベクトルの標本対応物. OVB Omitted variables bias, the relationship between regression estimates in models with different sets of control variables. 脱落変数バイアス,異なるコントロール変数のリストを持つ回帰推定値の間の関係. QTE Quantile treatment effect, the causal effect of treatment on conditional quantiles of the outcome variable for compliers. 分位点処置効果,処置が順応者の結果の条件付き分位点に与える因果効果.
RD Regression discontinuity design, an identification strategy in which treatment, the probability of treatment, or the average treatment intensity is a known, discontinuous function of a covariate. 回帰不連続デザイン,処置,処置の確率,あるいは平均処置強度が共変量の既知の不連続関数である識別戦略. SEM Simultaneous equations model, an econometric framework in which causal relationships between variables are described by several equations. 同時方程式モデル,いくつかの方程式によって変数間の因果的関係が表わされた計量経済学的枠組み. SSIV Split-sample instrumental variables estimator, a version of the two-sample instrumental variables (TSIV) estimator. 分割標本操作変数推定量,2標本操作変数(TSIV) 推定量の1つのバージョン. TSIV Two-sample instrumental variables estimator, an instrumental variables (IV) estimator that can sometimes be constructed from two data sets when either data set alone would be inadequate. 2標本操作変数推定量,いずれか一方のデータのみでは不十分なときに2つのデータから構築し得る操作変数 (IV) 推定量. VIV Visual instrumental variables, a plot of reduced form against first-stage fitted values in instrumental variables models with dummy instruments. 視覚的操作変数,ダミー操作変数を用いた操作変数モデルで誘導形対第1段階予測値の散布図. WLS Weighted least squares, a GLS estimator with a diagonal weighting matrix. 加重最小二乗法,対角加重行列を用いたGLS推定量.