twoValuesAccumulatorByAddAllPartitionedPairedStats.populationCovariance());
    assertDiagonalLinearTransformation(
        manyValuesAccumulator.leastSquaresFit(),
        manyValuesAccumulator.xStats().mean(),
        manyValuesAccumulator.yStats().mean(),
        manyValuesAccumulator.xStats().populationVariance(),

   * product-moment correlation coefficient</a> of the values. The count must greater than one, and
   * the {@code x} and {@code y} values must both have non-zero population variance (i.e. {@code
   * xStats().populationVariance() > 0.0 && yStats().populationVariance() > 0.0}). The result is not
   * guaranteed to be exactly +/-1 even when the data are perfectly (anti-)correlated, due to
   * numerical errors. However, it is guaranteed to be in the inclusive range [-1, +1].

    assertThrows(IllegalStateException.class, () -> emptyAccumulator.populationVariance());
    assertThrows(
        IllegalStateException.class,
        () -> emptyAccumulatorByAddAllEmptyIterable.populationVariance());
    assertThrows(
        IllegalStateException.class, () -> emptyAccumulatorByAddAllEmptyStats.populationVariance());
    assertThat(oneValueAccumulator.populationVariance()).isEqualTo(0.0);

   * product-moment correlation coefficient</a> of the values. The count must greater than one, and
   * the {@code x} and {@code y} values must both have non-zero population variance (i.e. {@code
   * xStats().populationVariance() > 0.0 && yStats().populationVariance() > 0.0}). The result is not
   * guaranteed to be exactly +/-1 even when the data are perfectly (anti-)correlated, due to
   * numerical errors. However, it is guaranteed to be in the inclusive range [-1, +1].

        twoValuesAccumulatorByAddAllPartitionedPairedStats.populationCovariance());
    assertDiagonalLinearTransformation(
        manyValuesAccumulator.leastSquaresFit(),
        manyValuesAccumulator.xStats().mean(),
        manyValuesAccumulator.yStats().mean(),
        manyValuesAccumulator.xStats().populationVariance(),

      assertThat(actualStats.populationVariance()).isEqualTo(0.0);
      assertThat(actualStats.min()).isWithin(ALLOWED_ERROR).of(expectedStats.min());
      assertThat(actualStats.max()).isWithin(ALLOWED_ERROR).of(expectedStats.max());
    } else {
      assertThat(actualStats.mean()).isWithin(ALLOWED_ERROR).of(expectedStats.mean());
      assertThat(actualStats.populationVariance())
          .isWithin(ALLOWED_ERROR)

        TWO_VALUES_PAIRED_STATS.leastSquaresFit(),
        TWO_VALUES_PAIRED_STATS.xStats().mean(),
        TWO_VALUES_PAIRED_STATS.yStats().mean(),
        TWO_VALUES_PAIRED_STATS.xStats().populationVariance(),
        TWO_VALUES_PAIRED_STATS.populationCovariance());
    // For datasets of many double values, we test many combinations of finite and non-finite
    // x-values:
    for (ManyValues values : ALL_MANY_VALUES) {

        TWO_VALUES_PAIRED_STATS.leastSquaresFit(),
        TWO_VALUES_PAIRED_STATS.xStats().mean(),
        TWO_VALUES_PAIRED_STATS.yStats().mean(),
        TWO_VALUES_PAIRED_STATS.xStats().populationVariance(),
        TWO_VALUES_PAIRED_STATS.populationCovariance());
    // For datasets of many double values, we test many combinations of finite and non-finite
    // x-values:
    for (ManyValues values : ALL_MANY_VALUES) {

   * Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
   *
   * @throws IllegalStateException if the dataset is empty
   */
  public final double populationVariance() {
    checkState(count != 0);
    if (isNaN(sumOfSquaresOfDeltas)) {
      return NaN;
    }
    if (count == 1) {
      return 0.0;
    }

   * Double#NEGATIVE_INFINITY}, or {@link Double#NaN}) then the result is {@link Double#NaN}.
   *
   * @throws IllegalStateException if the dataset is empty
   */
  public double populationVariance() {
    checkState(count > 0);
    if (isNaN(sumOfSquaresOfDeltas)) {
      return NaN;
    }
    if (count == 1) {
      return 0.0;
    }