Measurement, Error, and Reliability in Quantitative Research

15 question-and-answer flashcards reviewing key ideas about measurement, error, and reliability from the lecture notes.

What is the definition of measurement used in the lecture?

Assigning numbers to properties of objects or events according to specific rules (an operational definition).

According to classical test theory, how is an observed score (Xo) decomposed?

Xo = Xt + Xe, where Xt is the true score and Xe is measurement error.

Which two broad types of measurement error were distinguished?

Systematic (constant) error and random error.

How does systematic error differ from random error?

Systematic error adds a constant bias to all scores, whereas random error varies unpredictably around zero and only affects variability.

What is the expected (mean) value of random error across many measurements?

Zero; the average of the random error distribution is 0.

How is reliability defined in terms of variances?

Reliability = True variance / Observed (total) variance.

How can reliability be expressed using correlations?

It equals the square of the correlation between the true score and the observed score (r²Xt,Xo).

What effect does random error have on the mean of observed scores?

None; because the mean of random error is zero, the sample mean remains an unbiased estimate of the true mean.

What effect does random error have on the variance of observed scores?

It inflates observed variance by adding error variance to true-score variance.

According to the attenuation principle, how does reliability limit correlations between two observed variables X and Y?

The observed correlation rXo,Yo cannot exceed the square root of reliability for X or for Y (whichever is lower).

If reliabilities for two measures are 0.8 and 0.7, what is the theoretical upper bound on their observed correlation?

√(0.8 × 0.7) ≈ 0.74.

Why does a large within-subject standard deviation signal lower reliability?

Greater spread in repeated measurements on the same person indicates larger random error, reducing the proportion of true variance.

Which instrument in the example study is used to quantify social anxiety?

The LSAS (Liebowitz Social Anxiety Scale) questionnaire.

Which computerized task in the example study assesses behavioral regulation in social situations?

The Emotional Stroop task implemented on a computer.

In the stated research question, which variable is suggested to moderate the link between social anxiety and behavioral regulation?

Biological sex (gender).