Core principle 4: Theory-guided construct measurement. Methods used to measure theoryrelated
or derived processes, outcomes, and exogenous factors (e.g., contextual and environmental)
directly associated with the specified program theory or model varied widely (e.g., from secondary
or extant data to interviews with program participants to document analysis to large-scale self-report
surveys). Very often, samples of theoretical constructs, and their respective targets of generalization,
were obtained from poorly devised measures intended to represent broader, more complex, latent
constructs. It was not uncommon for single indicator measures, which likely account for little of the
variance in the target latent construct, to be the primary means of construct measurement. Mole,
Hart, Roper, and Saal (2009), for example, used simple, easily obtained measures of sales and
employee growth as proxy indicators of small business productivity as opposed to operationally
defining productivity and using the operational definition to support direct theoretical construct
measurement. Others, such as Weitzman, Mijanovich, Silver, and Brecher (2009), however, used
very refined, sometimes standardized, measures of theoretically derived latent constructs such as
neighborhood quality of life, to evaluate a citywide health initiative. Noticeably, few of the cases
reviewed reported reliability coefficients and even fewer reported validity coefficients or other
information pertaining to the precision and accuracy of information as related to samples of latent
or observed constructs or their qualitative counterparts (e.g., trustworthiness, dependability, and
confirmability).