Skip to main content icon/video/no-internet

Validity, Face and Content

Validity refers to a condition in which statements or conclusions made about reality are reflective of that reality. A number of forms of validity exist in social science research. One form is generalizability, which represents the extent to which a statement or conclusion applies to populations or settings not included within the context of a specific study. Causal validity is concerned with statements or conclusions drawn about hypothesized effects of variables on outcomes under study. Measurement validity relates to the degree to which measures designed to capture concepts do in fact measure desired concepts. In other words, how accurately a measure in a study assesses what researchers believe the measure captures is reflective of the amount of measurement validity present within the study design. Thus, establishing whether a measure has gauged what it was designed to measure is an essential first step for determining validity of research findings. Measurement validity can be established using four distinct subtypes: face validity, content validity, criterion validity, and construct validity.

This entry focuses on the first two subtypes of measurement validity—face validity and content validity—specifically how they relate to one another as well as other forms of measurement validity (criterion validity and construct validity). It also provides examples of various approaches commonly used to assess face validity and content validity. Last, this entry discusses potential challenges faced by researchers interested in determining face validity and content validity of their research.

Measurement Validity

Measurement validity represents the degree to which a measure constructed to assess a specific concept actually measures the concept in question. A valid measure of a concept will be strongly associated with other measures of the concept previously shown to display high validity. Similarly, a valid measure should also significantly correlate with other concepts known to be strongly related to the concept being measured. In addition, a valid measure should not be associated with distinctly unrelated concepts. The validity of a measure is called into question when the measure is not associated with known valid measures of a concept, fails to correlate with measures of related concepts, or is found to be associated with measures of unrelated concepts.

Consider the following example. A researcher is testing the effects of playing violent video games on aggressive behavior among children. The researcher examines the validity of the measure for violent video-game playing and finds that it is associated with previously established measures of violent video-game playing and measures of other types of violent media consumption, including a measure assessing violent-film exposure. In addition, the researcher considers whether the measure of violent video-game playing is associated with measures of news consumption and confirms that the study’s measure is not associated with this unrelated concept. Based on this, the researcher can conclude that the study’s measure of violent video-game playing exhibits high measurement validity.

Although the process for establishing the validity of a measure may be straightforward, developing highly valid measures can be a challenging endeavor. When a measure does not meet criteria of validity and thus fails to measure the concept it was designed to measure, a measurement error has occurred. Measurement error may occur as a result of three factors: idiosyncratic individual errors, generic individual errors, or method factors. Idiosyncratic individual errors arise when a relatively small subset of individuals respond to a measure in somewhat unsystematic and unexpected ways. For example, a respondent may fail to understand the meaning of a question about road safety or misinterpret specific wording regarding driving laws, which leads the respondent to offer answers that deviate from how the respondent would answer with a clearer interpretation of the question. Similarly, a research participant may provide answers to questions about road safety immediately following a near collision that are very different from responses to these questions following a less eventful day.

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Sage Recommends

We found other relevant content for you on other Sage platforms.

Loading