Hw#2 exam 2 review multiple choice 1. For a given set of data, the larger the explained variance, Question 1 options: The larger the total variance The larger is the R-squared value The greater the slope of the regression line The greater the total of the sum of the dependent variable values Question 2 (1 point) In logistic regression, the Y intercept Question 2 options: Is equal to the square of (y-bar) Is equal to 1 (one) over the slope of Y Is a value between 0 (zero) and 1 (one) Any value (it can't be determined from the question) Question 3 (1 point) Temperature, where the differences between any two degrees represents the same change (difference) as the difference between any other two degrees, would be an example of Question 3 options: Ratio scale Interval scale Partial correlation Question 4 (1 point) In a traditional regression analysis, the predictors are Question 4 options: The values obtained after subtracting the Y values and the mean of Y The X values The beta coefficients The expected values of Y given any value of X Question 5 (1 point) In multiple regression, beta coefficients Question 5 options: Can be different for each variable Can only be positive values or 0 (zero) Become larger in value as the number of predictor variables increase Question 6 (1 point) Assume a partial correlation is computed (with X, Y and Z). If the correlation between the original X and original Y values remains the same (as compared to the outcome of the partial correlational analysis) then the Question 6 options: Controlled variable (third variable) explains all of the original correlation Correlation between the original X and Y values must be 1.0 Correlation between either X and Z or Y and Z must be 1.0 Influence of the third variable is 0 (zero) Question 7 (1 point) If a regression line improves upon predicting the Y mean value for any X value, then based on the regression line, the Question 7 options: Sum of squares explained increases Sum of square total increases The slope of the regression line increases The beta weights approach 0 (zero) Question 8 (1 point) R-squared (r2 or R2) describes Question 8 options: Explained variance is about 0 (zero) The total sum of squares (sum of squares total) The strength of a regression model Question 9 (1 point) A regression analysis is run and a best fit line is computed. If all of the actual Y values are located on the regression line, Question 9 options: Explained variance is about 0 (zero) Unexplained variance is positive infinity The unexplained variance is about negative infinity None of the above Question 10 (1 point) As compared to a correlation analysis, an advantage of a scientific experiment is Question 10 options: Experiments can be completed in any set of circumstances while correlations are dependent on randomizing participants Experiments produce computed beta/coefficient weights while with correlation these values are only estimated Causality Question 11 (1 point) The Y intercept is Question 11 options: 0 (zero) 1 (one) Equal to the sum of squares total The point where a regression line crosses the Y axis Question 12 (1 point) In a regression analysis, as the predicted outcome (value) of a variable increases in distance from what was actually measured for the variable, Question 12 options: The residual value increases The explained variance increases The mean of the outcome variable increases The modal (the mode) value of the outcome variable approaches 0 (zero) Question 13 (1 point) In a regression analysis, the dependent variable Question 13 options: Reflects what the regression model is trying to predict Has a beta weight between the values of 0 (zero) and 1 (one) Is the confound variable Question 14 (1 point) If the sum of squares for the residuals is 0 (zero) Question 14 options: The residuals will each be 1 (one) standard deviation (of the independent variable) from the regression line The difference between the predicted values of Y and the actual values is also 0 (zero) None of the actual Y values equal any of the predicted values for Y Question 15 (1 point) If the regression line is equal to the mean of the Y values (the intercept is the Y mean and the slope is 0 (zero)) Question 15 options: The regressor values must each also equal 0 (zero) The sum of squares residuals (explained) equals the sum of squares unexplained The slope of the regression line is not computable Question 16 (1 point) In a partial correlation of three variables (for example, X, Y and Z) Question 16 options: The partial correlation is simply the original r value divided by 3 All three variables are considered outcome variables The computed partial correlation can be any value between -1.0 and 1.0 Question 17 (1 point) According to the concept "regression to the mean" Question 17 options: The likelihood of an extreme value occurring is 0 (zero) If an event produces two outcomes, those outcomes are correlated An extreme outcome of any event is likely to be followed by a less extreme outcome A perfect regression line is one that is fit to the mean of the Y values Question 18 (1 point) In the ordinal scale, the difference between any two points of the scale Question 18 options: Is equal to 0 (zero) Is as large as the number of points making up the scale Can vary Is consistent (can't vary) Question 19 (1 point) In the following "rxy.z", the possible confound is Question 19 options: X Y Z r Question 20 (1 point) After a linear regression analysis has been completed and a linear model has been produced, Question 20 options: The slope of the regression line must be equal to the mean of the Y variable The intercept of the regression line is always where Y = 0 (zero) If, for any X, Y is a negative number, the linear model is invalid A dependent variable value can be computed for any value of the independent variable Question 21 (1 point) In correlational analyses (as discussed in class), the term "control" Question 21 options: Refers to randomization of participants into different groups Is synonymous (means the same) with "causality" Refers to factoring out any impact of a confound Is synonymous (means the same) with "independent variable" Question 22 (1 point) In a linear regression analysis, a goal is to Question 22 options: Separate data points into one of two different outcome conditions Find the mean of the Y variable Produce a model that predicts the values of the independent variable Produce a model that predicts the values of the dependent variable Question 23 (1 point) An attempt to predict the probability of correctly placing an observation into one of two different conditions Question 23 options: Is an example of logistic regression Involves grouping values defined by nominal scale measurements Can only occur if the slope of a best fit regression line is a non-zero value Requires partial correlations Question 24 (1 point) Regressor values are Question 24 options: The independent variable values The dependent variable values In any linear models, those values with the smallest residuals Any dependent variable values that are greater than the mean of the Y variable Question 25 (1 point) The (numeric) value of the Coefficient of Determination Question 25 options: Can never be 0 (zero) Can be any value Must be between 0 (zero) and 1 (one) Will always be higher than the value of r Question 26 (1 point) The measure of central tendency for the Ratio scale is Question 26 options: The coefficient of determination The mean The mode 0 (zero) Question 27 (1 point) Unexplained variance can be directly computed using Question 27 options: 1 (one) and R2 (R squared) X values and the mean of X The predicted Y values and the mean of X The Y intercept and the mean of Y () Question 28 (1 point) In measurement scale, the value 0 (zero) can be part of Question 28 options: Ordinal scales Interval scales All of the scales Only the ratio scale Question 29 (1 point) (Y-bar) is Question 29 options: Involved with calculating the regressors 0 (zero) when the slope of a regression line is 0 (zero) The symbol for the mean of the Y variable values Question 30 (1 point) Which of the following involves a bivariate condition Question 30 options: Pearson's r The mean of the Y-axis values Values associated with a measurement based on the interval scale Regression predictors Question 31 (1 point) In correlation, a confound is Question 31 options: Another name for the correlation coefficient A computational error (e.g. r = 5) A possible explanation for a potential association that extends beyond the variables used in the original correlational analysis Any value among the data that is considered an outlier and artificially alters the computed value of the correlational analysis Question 32 (1 point) The sum of squares (sum of squares total, sum of squares residual and sum of squares explained) is squared because Question 32 options: R-squared is also a square To keep the sum of squares between the values of 0 (zero) and 1 (one) Otherwise all of the values involved in the sum of squares would be negative values All of the relevant values would cancel out and add up to a total of 0 (zero) Question 33 (1 point) A regression line Question 33 options: Represents all of the predicted X values Has a slope equal to the explained variance Can't be a horizontal line (parallel to the X axis) Represents the predicted Y values for any value of X