Test Validity and Reliability

True or False

V The validity coefficient cannot be higher than the index of reliability.

V The validity coefficient is affected by the reliability of the criterion.

F The standard error of estimate is the difference between obtained and predicted scores.

F The coefficient of validity is an indicator of the stability of scores.

F The validity coefficient of a test is independent of the homogeneity of the sample.

F The validity coefficient expresses the correlation between two parallel forms of a test.

V The standard error of estimate is the standard deviation of the distribution of estimation errors.

F The validity coefficient of a test should always be less than the reliability coefficient.

V The validity coefficient cannot be higher than the index of reliability.

F The standard error of estimate of a test is the same with respect to different criteria.

F The validity coefficient of a test is independent of the ratio of predictive value.

F The coefficient of alienation increases as the coefficient of validity increases.

V The coefficient of determination increases with decreasing the coefficient of alienation.

V The true prediction error of the score is the difference between the true score and the predicted true score.

F The statistic X refers to the linear regression MCRG/MCRS.

V The maximum value that the correlation of a test with another variable can reach is the index of reliability.

F The multiple correlation coefficient represents the correlation between test scores and criterion scores predicted by the set of predictors.

V The predictive coefficient is the complement of the coefficient of alienation.

V The validity coefficient of a test for a criterion can reach any desired value by sufficiently stretching the test and the criterion.

V The coefficient of determination of a test for a criterion can be greater than the coefficient of validity.

V The procedure known as multitrait-multimethod matrices is used to establish the concurrent validity of a test for a criterion.

V The procedure known as “correlation attenuation with error” can estimate the correlation between true scores of two variables.

F The analysis of items refers to the study of reliability and validity of tests referred to the criterion.

F The validity index of an item is defined as the correlation between scores on the item and test scores.

V The estimation error affects the accuracy of predictions.

F Evidence of content validity is obtained by test-criterion correlation.

V Sociocultural variables of different groups can be sources of bias in testing.

V The validity of an element is expressed in correlation with an external criterion.

V The correlation between the responses of an item in a test and an external criterion is the first estimate of the validity of the item.

V Content validity is fundamental to make any type of generalization from the scores of the subjects in the test.

F Predictive validity refers to the effectiveness of a test to predict the external criterion scores.

F Concurrent validity is a special case of content validity.

F Content, predictive, and construct validity must be interpreted as different and independent types of validity.

F Improving the reliability of a test favorably affects its validity.

F Content validity is evaluated using the multitrait-multimethod matrix.

F Content validity is analyzed using the coefficient of validity.

F The proportion of criterion variance explained by a predictor test is expressed by the coefficient of validity.

V The slope of the regression of the criterion on the test is a function of the coefficient of validity.

F The multiple linear regression equation in standard scores always includes the term b0.

F The multitrait-multimethod matrix is used to analyze discriminant validity.

V The concurrent validity procedure is used to study evidence of validity referring to the internal structure of the test.

V The proportion of variation of a criterion associated with the variation of a test is expressed by the test-criterion correlation coefficient.

F Correlation attenuation estimates the correlation between two variables X and Y by eliminating measurement errors from both.

V Evidence of content validity of a test is usually obtained through indices of agreement between expert judges.

F The sensitivity of a test is obtained as the ratio of the number of subjects diagnosed as sick to the total number of subjects.

V The specificity of a test for a criterion is obtained as the conditional probability.

V The slope of the regression line expresses the change that occurs in the criterion score for a unit change in the test score.

V Convergent validity is a special case of content validity.

F Validity refers to the degree to which a test accurately measures a construct or domain.

V Face validity affects the motivation of subjects to the test.

F Content validity is obtained by test-criterion correlation.

V Predictive validity is part of theoretical validity.

F The repetition of the results obtained in a test constitutes the validity of the measure.

V We understand by content validity the degree to which the items of a test are representative and relevant to the domain that we want to evaluate.

V In the validity of decisions, the sensitivity index refers to the proportion of subjects correctly detected by the test with respect to the total of existing cases according to the experts.

V In predictive validity, the criterion is measured after applying the test.

V In a study of predictive validity, the test measures are taken prior to the criterion measures.

F In the linear regression model of Y on X, it is not assumed that the variances of the errors are equal at different values of X.

F In the multiple linear regression model, the weights of the variables are on the same scale as the corresponding variables.

F In multiple linear regression, the presence of a statistically significant F contrast indicates that all predictor variables have also significant coefficients.

V In the simple linear regression model, the slopes of the line in direct scores and in standard scores are equal.

V In the linear regression model, the scores predicted by the model and the prediction errors are statistically independent.

F A test that has a low validity coefficient will also have low reliability.

V A test with a reliability coefficient close to zero is a test with low predictive validity.

F A content validation study of a test indicates whether its scores adequately represent the domain that the test intends to measure.

V A regression line of Y on X with a positive slope results in a positive correlation between X and Y.

V The standard procedure to examine evidence of validity as the internal structure of a test is factor analysis.

F The threat to content validity is called construct irrelevant variance.

V A test that has a low validity coefficient for a criterion will also have low reliability.

V Validation studies refer to all the evidence that can be obtained to make inferences from the scores of the subjects in the test.

F Standardized coefficients in the multiple linear regression model allow determining the relative importance of the variables in the model.

F Degrees of freedom associated with the sum of squares due to regression depend on the number of predictors in the model.

F Validation studies collect the necessary evidence to make inferences from the scores of the subjects in the test.

V Other conditions being equal, the value of the validity coefficient of a test for a criterion is greater when calculated in a group with greater variability.

V By restricting the range of variability of the X scores, the slope of the regression line of Y on X does not vary.

F Concurrent validity and predictive validity are procedures referred to a criterion that differ in the time of collection of the criterion scores.

V Convergent and discriminant validity are procedures to examine the relative validity for a criterion that differ in the time of data collection.

V Convergent validity exists when the measures of the same trait obtained by different methods are correlated.

F All variables that correlate significantly with the criterion individually become part of the multiple linear regression equation of the criterion.

V If we find that rxy = 0.50, we can say that 50% of the variation of Y is associated with the variation of X.

F If we add parallel elements to a test, its overall validity coefficient for a criterion will increase.

V If the reliability of the test and the criterion are perfect, the validity coefficient will always be 1.

F If the correlation coefficient of a test and a criterion is 0.6, we can say that the percentage of criterion variance that we can explain by the predictor is 60%.

V If there were no measurement errors, the validity coefficient of the test would be the correlation between the true scores of the subjects in the test and the true scores in the criterion.

V To demand that two different tests measure the same domain, alternative methods of the test, implies maintaining a set of common items to anchor the effect of both tests.

V Knowing that the validity coefficient of a test for a criterion is 0.8, we can say that the alienation coefficient is 0.6.

V The validity index of an item is called the correlation between the scores obtained in the item and the scores obtained by those same subjects in the criterion.

V The validity coefficient cannot be higher than the reliability index.

V The validity coefficient is affected by the reliability of the criterion.

F The standard error of estimate is the difference between the obtained score and the predicted score.

V The validity coefficient is affected by the reliability of the criterion.

F The validity coefficient is an indicator of the stability of scores.

F The validity coefficient of a test is independent of the homogeneity of the sample.

F The validity coefficient expresses the correlation between two or more forms of a test.

V The standard error of estimate is the standard deviation of the distribution of estimation errors.

F The validity coefficient of a test must always be less than the reliability coefficient.

V The validity coefficient cannot be higher than the reliability index.

F The standard error of estimate of a test is the same with respect to different criteria.

F The validity coefficient of a test is independent of the predictive value coefficient.

F The alienation coefficient increases when the validity coefficient increases.

V The determination coefficient increases when the alienation coefficient decreases.

V The prediction error of the true score is the difference between the true score and the predicted true score.

F The statistic X refers to linear regression.

V The maximum value that the correlation of a test with another variable can reach is the reliability index.

F The multiple correlation coefficient represents the correlation between the criterion scores and the scores predicted by the set of predictors.

V The predictive coefficient is the complementary value of the alienation coefficient.

V The validity coefficient of a test for a criterion can reach any desired value by sufficiently stretching the test and the criterion.

V The determination coefficient of a test for a criterion can be greater than the validity coefficient.

V The procedure known as multitrait-multimethod matrices is used to establish the concurrent validity of a test for a criterion.

V The procedure known as “correlation attenuation with error” allows estimating the correlation between the true scores of two variables.

F The analysis of items refers to the study of the reliability and validity of tests referred to a criterion.

F The validity index of an item is defined as the correlation between the scores obtained in the item and the scores obtained in the test.

V The estimation error affects the accuracy of predictions.

F Evidence of content validity is obtained through the test-criterion correlation.

V The sociocultural variables of different groups can be sources of bias in tests.

V The validity of an element is expressed in correlation with an external criterion.

V The correlation between the responses of an item in a test and an external criterion is the first estimate of the validity of the item.

V Content validity is fundamental to make any type of generalization from the scores of the subjects in the test.

F Predictive validity refers to the effectiveness of a test to predict the scores of the external criterion.

F Concurrent validity is a special case of content validity.

F Content, predictive, and construct validity must be interpreted as different and independent types of validity.

F Improving the reliability of a test favorably affects its validity.

F Content validity is evaluated using the multitrait-multimethod matrix.

F Content validity is analyzed using the validity coefficient.

F The proportion of criterion variance explained by a predictor test is expressed using the validity coefficient.

V The slope of the regression of the criterion on the test is a function of the validity coefficient.

F The multiple linear regression equation in standard scores always includes the term b0.

F The multitrait-multimethod matrix is used to analyze discriminant validity.

V The concurrent validity procedure is used to study evidence of validity referring to the internal structure of the test.

V The proportion of variation of a criterion associated with the variation of a test is expressed using the test-criterion correlation coefficient.

F Correlation attenuation estimates the correlation between two variables X and Y by eliminating measurement errors from both.

V Evidence of content validity of a test is usually obtained through indices of agreement between expert judges.

F The sensitivity of a test is obtained as the ratio of the number of subjects diagnosed as sick to the total number of subjects.

V The specificity of a test for a criterion is obtained as the conditional probability.

V The slope of the regression line expresses the change that occurs in the criterion score for a unit change in the test score.

V Convergent validity is a special case of content validity.

F Validity refers to the degree to which a test accurately measures a construct or domain.

V Face validity influences the motivation of the subjects to the test.

F Content validity is obtained through the test-criterion correlation.

V Predictive validity is part of theoretical validity.

F The repetition of the results obtained in a test constitutes the validity of the measure.

V We understand by content validity the degree to which the items of a test are representative and relevant to the domain that we want to evaluate.

V In the validity of decisions, the sensitivity index refers to the proportion of subjects correctly detected by the test with respect to the total of existing cases according to the experts.

V In predictive validity, the criterion is measured after applying the test.

V In a study of predictive validity, the test measures are taken prior to the criterion measures.

F In the linear regression model of Y on X, it is not assumed that the variances of the errors are equal at different values of X.

F In the multiple linear regression model, the weights of the variables are on the same scale as the corresponding variables.

F In multiple linear regression, the presence of a statistically significant F contrast indicates that all predictor variables also have significant coefficients.

V In the simple linear regression model, the slopes of the line in direct scores and in standard scores are equal.

V In the linear regression model, the scores predicted by the model and the prediction errors are statistically independent.

F A test that has a low validity coefficient will also have low reliability.

V A test with a reliability coefficient close to zero is a test with low predictive validity.

F A content validation study of a test indicates whether its scores adequately represent the domain that the test intends to measure.

V A regression line of Y on X with a positive slope results in a positive correlation between X and Y.

V The standard procedure to examine evidence of validity as the internal structure of a test is factor analysis.

F The threat to content validity is called construct irrelevant variance.

V A test that has a low validity coefficient for a criterion will also have low reliability.

V Validation studies refer to all the evidence that can be obtained to make inferences from the scores of the subjects in the test.

F Standardized coefficients in the multiple linear regression model allow determining the relative importance of the variables in the model.

F Degrees of freedom associated with the sum of squares due to regression depend on the number of predictors in the model.

F Validation studies collect the necessary evidence to make inferences from the scores of the subjects in the test.

V Other conditions being equal, the value of the validity coefficient of a test for a criterion is greater when calculated in a group with greater variability.

V By restricting the range of variability of the X scores, the slope of the regression line of Y on X does not vary.

F Concurrent validity and predictive validity are procedures referred to a criterion that differ in the time of collection of the criterion scores.

V Convergent and discriminant validity are procedures to examine the relative validity for a criterion that differ in the time of data collection.

V Convergent validity exists when the measures of the same trait obtained by different methods are correlated.

F All variables that correlate significantly with the criterion individually become part of the multiple linear regression equation of the criterion.

V If we find that rxy = 0.50, we can say that 50% of the variation of Y is associated with the variation of X.

F If we add parallel elements to a test, its overall validity coefficient for a criterion will increase.

V If the reliability of the test and the criterion are perfect, the validity coefficient will always be 1.

F If the correlation coefficient of a test and a criterion is 0.6, we can say that the percentage of criterion variance that we can explain by the predictor is 60%.

V If there were no measurement errors, the validity coefficient of the test would be the correlation between the true scores of the subjects in the test and the true scores in the criterion.

V To demand that two different tests measure the same domain, alternative methods of the test, implies maintaining a set of common items to anchor the effect of both tests.

V Knowing that the validity coefficient of a test for a criterion is 0.8, we can say that the alienation coefficient is 0.6.

V The validity index of an item is called the correlation between the scores obtained in the item and the scores obtained by those same subjects in the criterion.