C2 · 50 words
Inferential statistics in Spanish
Sample, estimator, test and error — the terms in which the few are made to speak for the many.
Half this page turns on la muestra, the sample, and on the verb behind it, muestrear. The same noun does two more jobs in ordinary Spanish. It is an exhibition — una muestra de cine iraní, a small festival — and it is a token of something felt: una muestra de cariño, en muestra de respeto. All three come from mostrar, to show: a sample shows what the rest is like, a show shows the work, and a token shows what somebody means. English needs three words to keep them apart.
The words
View
Phrase (47)
la inferencia estadísticathe reasoning that goes from the few to the manyPhraseC2
La inferencia estadística va de la muestra a la población.Statistical inference goes from the sample to the population.
la población estadísticathe whole group one wants to say something aboutPhraseC2
La población estadística se define antes de muestrear.The statistical population is defined before sampling.
el muestreo aleatorio simplethe drawing in which everybody has the same chancePhraseC2
El muestreo aleatorio simple da a todos la misma probabilidad.Simple random sampling gives everybody the same probability.
el muestreo estratificadothe drawing that makes sure each subgroup is inPhraseC2
El muestreo estratificado asegura cada subgrupo.Stratified sampling makes sure of each subgroup.
el muestreo por conglomeradosthe drawing of whole groups at once, which is cheaper in the fieldPhraseC2
El muestreo por conglomerados abarata el trabajo de campo.Cluster sampling makes the fieldwork cheaper.
el marco muestralthe list one actually draws from, which is never quite the populationPhraseC2
El marco muestral omitía a los sin teléfono.The sampling frame left out those without a telephone.
el error de coberturathe gap between that list and the real group, which more interviews do not closePhraseC2
El error de cobertura no se corrige con más encuestas.Coverage error is not corrected by more surveys.
la no respuestathe people who did not answer, who bias the result if their silence has a reasonPhraseC2
La no respuesta sesga si depende del tema.Non-response biases the result if it depends on the subject.
el estimador insesgadothe estimator that is right on averagePhraseC2
El estimador insesgado acierta en promedio.The unbiased estimator is right on average.
la eficiencia del estimadorthe question of how tightly it clusters around the truthPhraseC2
La eficiencia del estimador se mide por su varianza.The efficiency of the estimator is measured by its variance.
la consistencia del estimadorthe property of closing in on the truth as the sample growsPhraseC2
La consistencia del estimador solo se ve al crecer la muestra.The consistency of the estimator shows only as the sample grows.
el error estándarthe spread of the estimate itself, which falls with the square rootPhraseC2
El error estándar baja con la raíz del tamaño.The standard error falls with the square root of the size.
la distribución muestralthe spread of the estimate over repeated samples, not of the dataPhraseC2
La distribución muestral no es la de los datos.The sampling distribution is not the distribution of the data.
el teorema central del límitethe theorem that explains why the bell curve keeps appearingPhraseC2
El teorema central del límite explica por qué aparece la campana.The central limit theorem explains why the bell appears.
la ley de los grandes númerosthe law that says nothing at all about the short runPhraseC2
La ley de los grandes números no promete nada a corto plazo.The law of large numbers promises nothing in the short run.
el intervalo de credibilidadthe interval that may honestly be read as a probabilityPhraseC2
El intervalo de credibilidad sí admite lectura como probabilidad.The credible interval does admit a reading as a probability.
la cobertura nominal del intervalothe coverage the interval claims, here ninety-five per centPhraseC2
La cobertura nominal del intervalo se fijó en el noventa y cinco por ciento.The nominal coverage of the interval was set at ninety-five per cent.
la precisión de la estimaciónthe tightness of an estimate, which falls away in small subgroupsPhraseC2
La precisión de la estimación cae cuando el subgrupo es pequeño.The precision of the estimate falls when the subgroup is small.
la hipótesis nulathe claim that there is no differencePhraseC2
La hipótesis nula afirma que no hay diferencia.The null hypothesis states that there is no difference.
la hipótesis alternativathe claim on the other side, written down before lookingPhraseC2
La hipótesis alternativa se formula antes de mirar.The alternative hypothesis is formulated before looking.
el contraste de hipótesisthe procedure that can reject and never confirmPhraseC2
El contraste de hipótesis solo puede rechazar.Hypothesis testing can only reject.
el valor pthe probability of a result this extreme if the null were truePhraseC2
El valor p supone cierta la hipótesis nula.The p-value assumes the null hypothesis to be true.
el nivel de significaciónthe threshold below which one rejects, which is a conventionPhraseC2
El nivel de significación es una convención, no una ley.The significance level is a convention, not a law.
el error de tipo unothe error of announcing something that is not therePhraseC2
El error de tipo uno afirma lo que no hay.A type one error asserts what is not there.
el error de tipo dosthe error of missing something that isPhraseC2
El error de tipo dos pasa por alto lo que sí hay.A type two error overlooks what is.
la potencia del contrastethe chance of catching a real effect, which grows with the samplePhraseC2
La potencia del contraste depende del tamaño de muestra.The power of the test depends on the sample size.
el tamaño del efectothe measure of how big the difference is, not merely whetherPhraseC2
El tamaño del efecto dice cuánto, no solo si.The effect size says how much, not only whether.
el cálculo del tamaño muestralthe working out beforehand of how many are neededPhraseC2
El cálculo del tamaño muestral se hace antes, no después.The sample size calculation is done beforehand, not after.
la corrección por comparaciones múltiplesthe adjustment that stops many tests producing cheap findingsPhraseC2
La corrección por comparaciones múltiples evita hallazgos baratos.Correction for multiple comparisons avoids cheap findings.
la exploración indebida de datosthe searching that always turns something upPhraseC2
La exploración indebida de datos encuentra siempre algo.Improper exploration of the data always finds something.
el registro previo del análisisthe writing down of the plan in advance, which ties one's handsPhraseC2
El registro previo del análisis ata las manos a propósito.Pre-registration of the analysis ties one's hands on purpose.
la replicabilidad del resultadothe coming out the same way when somebody else tries itPhraseC2
La replicabilidad del resultado es la prueba más dura.The replicability of the result is the hardest test.
el contraste de la t de Studentthe test that compares two meansPhraseC2
El contraste de la t de Student compara dos medias.Student's t-test compares two means.
el análisis de la varianzathe method that compares more than two groupsPhraseC2
El análisis de la varianza compara más de dos grupos.The analysis of variance compares more than two groups.
la prueba no paramétricathe test that assumes no particular shape of distributionPhraseC2
La prueba no paramétrica no supone normalidad.The non-parametric test assumes no normality.
el supuesto de normalidadthe assumption of a bell shape, checked with a plotPhraseC2
El supuesto de normalidad se comprueba con un gráfico.The assumption of normality is checked with a plot.
el valor atípicothe point far from the rest, to be examined and not deletedPhraseC2
El valor atípico se examina, no se borra.The outlier is examined, not deleted.
la estimación robustathe estimate that survives a few extreme pointsPhraseC2
La estimación robusta aguanta unos pocos valores extremos.Robust estimation withstands a few extreme values.
la regresión lineal múltiplethe model that holds several variables constant at oncePhraseC2
La regresión lineal múltiple controla varias variables a la vez.Multiple linear regression controls several variables at once.
el coeficiente de determinaciónthe share of variation explained, which proves no causePhraseC2
El coeficiente de determinación no prueba causa.The coefficient of determination does not prove cause.
el residuo del modelothe leftover between what the model said and what happenedPhraseC2
El residuo del modelo se dibuja contra el valor previsto.The residual of the model is plotted against the predicted value.
la inferencia bayesianathe approach that starts from a belief and updates itPhraseC2
La inferencia bayesiana parte de una creencia previa.Bayesian inference starts from a prior belief.
la distribución a priorithe belief held before the data, which must be stated openlyPhraseC2
La distribución a priori debe declararse abiertamente.The prior distribution must be declared openly.
la distribución a posteriorithe belief after the data have been taken inPhraseC2
La distribución a posteriori combina datos y creencia.The posterior distribution combines data and belief.
el factor de Bayesthe ratio that weighs one model against another directlyPhraseC2
El factor de Bayes compara dos modelos directamente.The Bayes factor compares two models directly.
la simulación estadísticathe imitating of the process on a computer instead of solving itPhraseC2
La simulación estadística sustituye a la teoría difícil.Statistical simulation replaces the difficult theory.
la incertidumbre declaradathe uncertainty printed alongside every published numberPhraseC2
La incertidumbre declarada acompaña a toda cifra publicada.The declared uncertainty accompanies every published figure.
Noun (f) (2)
la homocedasticidadthe assumption that the spread is the same everywhereNoun (f)C2
La homocedasticidad falla cuando el error crece con el valor.Homoscedasticity fails when the error grows with the value.
la multicolinealidadthe trouble caused when two predictors carry the same informationNoun (f)C2
La multicolinealidad hace inestables los coeficientes.Multicollinearity makes the coefficients unstable.
Noun (m) (1)
el remuestreothe drawing again from one's own data to see how much it wobblesNoun (m)C2
El remuestreo estima el error sin fórmula cerrada.Resampling estimates the error without a closed formula.
This list belongs to History, research and data — nine lists and 450 words that can also be practised as a whole.
Practise this list
From these 50 words, as many per round as you like — up to all of them: first what is due today, then new ones. First you choose, then you fill the gap, then you write it yourself. Mistakes come back again shortly.