Social survey data are often collected through human interaction, which is prone to error. The data is collected with the assumption that the characteristics and concepts being measured may be precisely defined, and can be obtained through a set of well-defined procedures, and have true scores independent of the survey. Measurement error is then the difference between the value of a characteristic provided by the respondent and the true but unobserved value of that characteristic (Groves, 1989; Kasprzyk, 2005). To capture the true score of a characteristic, a survey question must be “valid” and “reliable”. The validity of a question is an evaluation of whether it measures a construct or variable that it is intended to measure (Carmines & Zeller, 1979). Reliability reflects the amou nt of error inherent in any measurement and hence how replication of the administration would give a different result (Streiner & Norman, 1995). If a survey variable is unreliable, the statistical inferences obtained using that variable is in turn also untrustworthy. Before we can use a survey variable, we must first be able to quantify how reliable it is.