Unit 2 - Practice Quiz

A scatter plot uses dots to represent values for two different numeric variables, making it ideal for visualizing the relationship or association between them.

A pattern moving from lower-left to upper-right indicates that as one variable increases, the other variable also tends to increase. This is known as a positive correlation.

When points are randomly scattered on a plot without any discernible pattern, it implies that there is no clear linear relationship or association between the two variables.

The correlation coefficient has a range from -1 (perfect negative linear correlation) to +1 (perfect positive linear correlation), inclusive.

A value of +1.0 signifies a perfect positive linear relationship. This means as one variable increases, the other increases in a perfectly predictable straight line.

A correlation coefficient of 0 indicates the absence of a linear relationship. However, it does not rule out the possibility of a non-linear relationship between the variables.

The correlation coefficient is a unit-free measure. It is independent of the change of origin and scale of the variables, so changing units does not alter its value.

Correlation measures the strength and direction of a relationship, but it does not prove that one variable causes the change in the other. A third, unobserved variable could be influencing both.

Karl Pearson's coefficient is designed to measure the strength and direction of the linear relationship between two continuous or quantitative variables.

Karl Pearson's correlation coefficient is frequently referred to as the Pearson product-moment correlation coefficient (PPMCC).

A positive sign indicates a positive linear relationship (as one variable increases, so does the other), while a negative sign indicates a negative linear relationship (as one variable increases, the other decreases).

Pearson's correlation coefficient specifically measures the strength and direction of a linear relationship. If the relationship is non-linear, this coefficient may not be an accurate measure of the association.

Spearman's rank correlation is a non-parametric test that is used to measure the degree of association between two variables that are measured on an ordinal (rank) scale.

To calculate Spearman's rho, the raw data for each of the two variables must first be converted into ranks, typically from lowest to highest.

A monotonic relationship is one where the variables tend to move in the same relative direction, but not necessarily at a constant rate. Spearman's correlation measures the strength of this monotonic association.

Spearman's correlation is specifically designed for ordinal (ranked) data. It is also a good choice for quantitative data that has a non-linear but monotonic relationship.

Linear regression aims to find the best-fitting straight line that can be used to predict the value of a dependent variable based on the value of an independent variable.

The coefficient is the slope. It represents the estimated change in the dependent variable for every one-unit increase in the independent variable .

The method of least squares is used to find the line of best fit. This method works by minimizing the sum of the squared differences (residuals) between the observed values and the values predicted by the line.

A residual is the error in a prediction for a single data point. It is calculated as the actual observed value () minus the value predicted by the regression line ().

The formula for Karl Pearson’s correlation coefficient (r) is . Here, , so . Also, , so . Plugging in the values: .

Pearson's coefficient measures the strength of a linear relationship. Spearman's coefficient measures the strength of a monotonic relationship (one that is consistently increasing or decreasing). It is suitable for non-linear but monotonic relationships and for ordinal data.

A fundamental property of the linear regression line is that it always passes through the point of means, . Therefore, the means must satisfy the regression equation. Substituting into the equation gives .

The Pearson correlation coefficient (r) measures the strength and direction of a linear relationship. A U-shaped pattern indicates a strong non-linear relationship, but since the initial decrease is cancelled out by the subsequent increase, there is no overall linear trend. Thus, r will be close to 0.

The correlation coefficient is a dimensionless quantity, meaning it is independent of the units of measurement. Changing the scale of the variables (e.g., from meters to centimeters) by multiplying by a positive constant does not change the value of the correlation coefficient.

The slope of the regression line (in this case, 5) represents the predicted change in the dependent variable (Y) for a one-unit increase in the independent variable (X). Therefore, for each extra hour of study, the score is expected to increase by 5 points.

The ranks for Y are in the exact reverse order of the ranks for X. This represents a perfect negative monotonic relationship. Therefore, the Spearman's rank correlation coefficient must be -1.0 without calculation. Alternatively, are (-2, 0, 2). . Using the formula , we get .

The correlation coefficient is the geometric mean of the two regression coefficients, and . The formula is . The sign of must be the same as the sign of the regression coefficients. Since both are negative, must be negative. .

The percentage of variation in Y explained by the linear relationship with X is given by the coefficient of determination, which is . If , then . Expressed as a percentage, this is 49%.

Both regression lines, Y on X and X on Y, must pass through the point of means . For the regression of Y on X, the point is . For the regression of X on Y, the variables are swapped, so the line must pass through the point , which is (19, 10).

The data points all lie on a perfect straight line with a positive slope, specifically . When there is a perfect positive linear relationship between two variables, the Karl Pearson's correlation coefficient is +1.

A correlation coefficient close to -1 indicates a strong negative linear relationship. It does not imply causation. The coefficient of determination means 81% of the variance is explained, not that 90% of points are on the line.

The correct option follows directly from the given concept and definitions.

The formula for the slope of the regression line of Y on X is . Plugging in the given values: .

Heteroscedasticity refers to the situation where the variability (scatter) of a variable is unequal across the range of values of a second variable that predicts it. In this case, the spread of Y changes as X changes, which violates an assumption of standard linear regression.

If two variables are independent, their covariance is 0. Since the correlation coefficient is calculated by dividing the covariance by the product of the standard deviations, a covariance of 0 will result in a correlation coefficient of 0. However, the converse is not always true; a correlation of 0 does not necessarily imply independence.

The question asks for the predicted value, which is found by substituting into the regression equation. . The actual value of 75 would be used to calculate the residual (error), which would be .

The correct option follows directly from the given concept and definitions.

The Pearson correlation coefficient is invariant to changes of origin (adding or subtracting a constant) and scale (multiplying or dividing by a positive constant). Shifting the data by adding or subtracting a constant from all values does not alter the strength or direction of the linear relationship.

A correlation coefficient of 0 specifically indicates the absence of a linear relationship. It does not rule out the possibility of a strong non-linear relationship (e.g., a parabolic relationship). While independence implies r=0, r=0 does not imply independence.

Pearson's measures the strength of a linear relationship. Here, the relationship is perfectly quadratic. For values symmetric around 0, an increase in does not consistently lead to an increase or decrease in . For , as increases, decreases. For , as increases, also increases. The positive and negative linear components cancel each other out, resulting in a covariance of 0, and thus a correlation coefficient of 0.

Original equation: . We have and . Substituting these into the original equation: . This simplifies to . Multiplying the entire equation by 5 gives . So the new intercept is and the new slope is .

Spearman's measures the strength of a monotonic relationship. Since is a perfect monotonic increasing function, the ranks of will be in the same order as the ranks of , making . Karl Pearson's measures the strength of a linear relationship. Since is curved, it is not perfectly linear, so its value will be less than 1. Therefore, .

The correlation coefficient is independent of changes in origin and scale. However, it is sensitive to the sign of the scaling factor. is a linear transformation of with a negative slope (-3), and is a linear transformation of with a positive slope (2). Because one of the slopes is negative, the sign of the correlation coefficient is reversed. Therefore, .

In OLS, the residuals are uncorrelated with the predictors and the fitted values. However, since , the covariance . As long as there is some error, is positive, so the correlation between the residuals and the observed values is non-zero.

Rearrange the lines into regression forms. Assume line 1 is x on y: , so . Assume line 2 is y on x: , so . The coefficient of determination is . Since both regression coefficients are negative, must also be negative. Thus, . The other identification of lines would lead to , which is impossible.

A high leverage point is an observation with an extreme value for the independent variable. When this point also falls close to the existing regression line, it doesn't change the slope much, but it greatly reduces the relative sum of squared errors. By adding a distant point that confirms the trend, it pulls the overall correlation coefficient significantly closer to 1 (or -1 if the trend is negative), strengthening the evidence for a linear relationship.

The standard, simplified formula for Spearman's correlation assumes there are no ties in the ranks. When ties are present, as seen in the ranks for variable X (3.5 and 7.5), this formula is no longer exact. A more complex formula involving correction factors for ties should be used, or alternatively, one can calculate the Pearson correlation on the rank values, which naturally handles ties correctly.

The coefficient of determination is . For Model A, . For Model B, . represents the proportion of variance explained, and is also given by . Since Model B has a higher , it explains more variance, meaning it has a smaller proportion of unexplained variance (). As is the same for both, must be smaller for Model B. Therefore, .

The correlation matrix of the three variables must be positive semi-definite. This implies the inequality . Plugging in the values gives , which simplifies to . This quadratic inequality holds for between the roots of the corresponding equation, which are 0.62 and 1. Thus, the minimum possible value for is 0.62.

This scenario describes Anscombe's Quartet. The key takeaway is that summary statistics alone can be extremely misleading. Drastically different data distributions can produce identical statistical properties. It highlights the absolute necessity of visualizing data to understand its underlying structure, check for outliers, and assess the appropriateness of a linear model, which a single correlation value cannot confirm.

Leverage measures how far an observation's x-value is from the mean of the x-values. A low leverage means its x-value is typical. A large residual means the point is a vertical outlier. A point is influential if its removal causes a large change in the model. Points with low leverage, even if they are outliers, are typically not influential because they don't have enough 'lever arm' to drastically change the slope. Their primary effect is to inflate measures of error.

The correct option follows directly from the given concept and definitions.

This scenario is an example related to Simpson's Paradox. When combining groups, the overall correlation can be drastically different from the correlation within each group. Differences in the means of the variables between the groups can create an overall trend that is different from, or even opposite to, the trend within the individual groups. Therefore, the combined correlation is not constrained.

This fan or cone shape in a residual plot is the classic sign of heteroscedasticity, which means the variance of the error terms is not constant across all levels of the independent variable. The linear regression model assumes homoscedasticity (constant variance). This violation suggests that the model's predictions are less reliable for certain ranges of the predictor variable.

The flawed conclusion ignores a confounding variable: the size of the fire. Larger fires require more firefighters and also inherently cause more damage. The size of the fire is the common cause for both variables. The observed correlation is therefore spurious because it is induced by this third variable, not because of a causal link between firefighters and damage.

Calculating the correlation coefficient from the given statistics: . By mathematical definition (due to the Cauchy-Schwarz inequality), the Pearson correlation coefficient must lie in the interval . A value of $1.1$ is impossible. While outliers or non-linearity can affect the value of , they can never push it outside this valid range. The only possible explanation is a mistake in the calculation of the summary statistics.

In a standard OLS regression with an intercept, one of the normal equations ensures that . However, when the model is forced through the origin, this constraint is removed. The model only minimizes , which does not require the sum of residuals to be zero. In fact, will generally not be zero unless the regression line happens to pass through the point of means .

Spearman's measures the strength of a monotonic relationship. The described relationship is a symmetric parabola. As increases from 0 to 5, increases (a positive monotonic relationship). As increases from 5 to 10, decreases (a negative monotonic relationship). Because the relationship is perfectly symmetric and non-monotonic over the full range, the positive and negative rank associations cancel each other out, leading to a Spearman's correlation coefficient close to 0.

The formula for the correlation coefficient is . If is a constant , its variance is 0, and therefore its standard deviation is also 0. Since the denominator of the correlation formula includes , calculating it would involve division by zero. Therefore, the correlation coefficient is undefined. Correlation measures how two variables co-vary; a variable cannot co-vary with a constant.