29 Correlations Flashcards
Correlation is:
a. A statistic to establish cause and effect.
b. A process of estimating prediction accuracy of one variable for another.
c. An estimate of shared variance between two variables.
d. A method to compare two groups.
c
Rationale: Correlation does not establish cause and effect, nor does it compare two groups or predict outcomes. The correlation coefficient looks at the degree of shared variance between two variables.
Correlation coefficients can provide information on all of the following except:
a. The relationship between two variables that are measured in different units.
b. Whether there is a direct or inverse relationship between two variables.
c. The degree of shared variance between two variables.
d. The relationship between variables that have a quadratic relationship.
d
Rationale: Correlation is based on the assumption that there is a linear relationship between X and Y.
A researcher has established a correlation of –.29 (p = .002) between knee strength and weight in adults. Based on this outcome, the relationship between strength and weight is:
a. Weak because it has a negative correlation.
b. Weak because the size of the coefficient is small.
c. Strong because the correlation is highly significant. d. Strong because it has a negative correlation.
b
Rationale: The correlation of –.29 is low and weak. The negative direction does not reflect the strength of the correlation. The p value cannot be used to interpret the strength of a correlation
A researcher wants to study the relationship between gender and heart rate. correlation statistics should be used?
a. Pearson product-moment correlation coefficient
b. Kendall’s tau
c. Point biserial correlation
d. Spearman correlation
c
Rationale: The point biserial correlation is used when a dichotomy is correlated with a continuous variable.
A researcher has studied the relationship between function and strength in individuals with dementia in a nursing care facility, finding a correlation of .75. Which of the following statements is a reasonable interpretation of this coefficient?
a. The decline in function or strength could be due to the individual’s memory
deficit.
b. Declines in function are not related to declines in strength.
c. A decrease in strength can cause a decline in function.
d. A decrease in function can cause strength to decline.
a
Rationale: Although this is a strong correlation, it is possible that another variables (memory deficit) could contribute to the decline in both variables. Correlation does not imply causation, only a common direction of change.
A researcher is interested in establishing the relationship between quality of life and age. He recruits a sample of 30 patients in a skilled nursing facility. He finds a correlation of –.25 (p = .31). Most likely, the correlation is low because:
a. There is no relationship between quality of life and age.
b. It is not significant.
c. It is a negative relationship.
d. The sample was probably homogenous on age and quality of life.
d
Rationale: One cause for a low correlation can be a lack of variance in the sample. Members of this sample are likely to be similar on age and their perception of quality of life. Samples should be composed of individuals with diverse scores to be able to visualize correlation.
A correlation between two variables X and Y resulted in r = .14 (p = .001). valid conclusion?
a. The correlation of .14 did not occur by chance.
b. There is a strong relationship between X and Y.
c. Variable X causes variable Y.
d. Variable X accounts for 14% of variance in variable Y.
a
Rationale: The correlation r = .14 is weak. However, with a p value of .001, it is unlikely that this correlation was achieved by chance. Correlation does not reflect causation. The square of the correlation coefficient (r 2) accounts for variance, not correlation.
A study has shown a correlatiTonESbeTtwBeAeNnKdiSetE(LcaLloErRie.s/dCaOyM) and weight at
correlation between exercise frequency and weight at r = .67. It would, therefore, be of interest to examine the correlation of diet and weight with the effect of exercise removed. This would require estimating:
a. The r2 for the relationship between diet and weight.
b. The partial correlation of diet and weight with the effect of exercise removed.
c. The zero-order correlation of weight and exercise.
d. Scatterplot of weight and exercise frequency.
b
Rationale: The partial correlation will adjust the correlation between diet and weight with the variance from exercise removed. This would be a first-order correlation. The correlation coefficient is the zero-order correlation.
Which of the following is not a consideration in determining whether the Pearson correlation is an appropriate statistic to study the relationship between two variables?
a. Both variables should be measured at the interval/ratio level.
b. The sample should be large enough to demonstrate at least 80% power.
c. The scatterplot should show linearity.
d. The variables should come from normal distributions.
b
Rationale: The Pearson correlation is appropriate for use with variables that have underlying normal distributions and that are linearly related. Other forms of the Pearson correlation can be used for rank or dichotomous data. Power is not a criterion on which to base the use of a specific test.
Which of the following statements about outliers is not true?
a. Outliers are values that are substantially different from the rest of the data.
b. Outliers have an effect on regression parameters.
c. Outliers are usually large values in the distribution.
d. Outliers can occur if measurement error occurs with a single participant.
c
Rationale: Outliers can be smaller values in a distribution.
