Very aggressive and intimidates the entire family”; (2) = “He is aggressive with his family”; and (1) = “He tries to tell his family what to do.” As noted, using a Case-Theme Scale Coding Matrix aids in summarizing the respective machismo code values resident within each machismo thematic category. Under this process, the team supervisor compares and summarizes the newly created thematic AZD-8835 site variable scale values generated independently by the two independent raters as encoded onto their Independent Solution Table (not shown). Then in the “Round Table 2,” the “scale coding round table review,” the coordinating supervisor and the independent raters compare and discuss these independently generated scale codes to research consensus in generating an Optimal Solution Table (see Table 1). Intercoder reliabilities–Within each thematic category, for initial frequency or intensity ratings, one can calculate an interrater reliability coefficient or a Cohen’s kappa coefficient that assesses intercoder agreement beyond chance (Fleiss, 1981), as an index of initial levels of intercoder agreements. However, the final scale code ratings move beyond these preliminary ratings based on the results of the round table discussions to generate an optimal solution. In summary, scale coding dimensionalizes a thematic category, converting it from a thematic category, which has the codes, 1 = mentioned and 0 = not mentioned, into a thematic variable (which has code values of 0, 1, 2, or 3). Once dimensionalized, each newly created thematic variable has distributional properties similar to those of a measured variable, which is defined by scores describable as Likerttype scaling.11 When dimensionalized, and if treating coded values as a Likert-type scale, a thematic variable can then be used as a conventional measured variable and incorporated into conventional correlation, regression, or other multivariate data analyses. A thematic variable may also be used as a moderator variable, one that encodes conditional effects. A moderator variable is “a variable that modifies the form or strength of the relation between an independent and a dependent variable” (MacKinnon, 2008, p. 275). A moderator variable that is derived from qualitative text analyses may operate as a “discovered” conditional effect, one that was not previously anticipated during the design stages of a given research study (Yoshikawa et al., 2008) but one that as a discovered variable can aid in describing new and important conditional and interactive effects.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript11According to measurement theory, it can be argued that for frequency scale coding, Codes 0, 1, 2, and 3 represent numerical counts, and for intensity scale coding these values represent categories on an I-CBP112 web ordinal scale. Treating these values as an interval-level Likerttype scale introduces the assumption of equal intervals between the values of 0, 1, 2, and 3, as these numbers would represent equal increments on the counts and on the levels of intensity for the frequency and intensity scale coding, respectively. We recognize that some investigators may question the assumption of equal interval levels as needed to satisfy parametric measurement assumptions for the use of interval scaling and the use of parametric statistics. For both modes of scale coding, frequency and intensity, we take this parametric approach wherein we indicate to our research assistants that the exempla.Very aggressive and intimidates the entire family”; (2) = “He is aggressive with his family”; and (1) = “He tries to tell his family what to do.” As noted, using a Case-Theme Scale Coding Matrix aids in summarizing the respective machismo code values resident within each machismo thematic category. Under this process, the team supervisor compares and summarizes the newly created thematic variable scale values generated independently by the two independent raters as encoded onto their Independent Solution Table (not shown). Then in the “Round Table 2,” the “scale coding round table review,” the coordinating supervisor and the independent raters compare and discuss these independently generated scale codes to research consensus in generating an Optimal Solution Table (see Table 1). Intercoder reliabilities–Within each thematic category, for initial frequency or intensity ratings, one can calculate an interrater reliability coefficient or a Cohen’s kappa coefficient that assesses intercoder agreement beyond chance (Fleiss, 1981), as an index of initial levels of intercoder agreements. However, the final scale code ratings move beyond these preliminary ratings based on the results of the round table discussions to generate an optimal solution. In summary, scale coding dimensionalizes a thematic category, converting it from a thematic category, which has the codes, 1 = mentioned and 0 = not mentioned, into a thematic variable (which has code values of 0, 1, 2, or 3). Once dimensionalized, each newly created thematic variable has distributional properties similar to those of a measured variable, which is defined by scores describable as Likerttype scaling.11 When dimensionalized, and if treating coded values as a Likert-type scale, a thematic variable can then be used as a conventional measured variable and incorporated into conventional correlation, regression, or other multivariate data analyses. A thematic variable may also be used as a moderator variable, one that encodes conditional effects. A moderator variable is “a variable that modifies the form or strength of the relation between an independent and a dependent variable” (MacKinnon, 2008, p. 275). A moderator variable that is derived from qualitative text analyses may operate as a “discovered” conditional effect, one that was not previously anticipated during the design stages of a given research study (Yoshikawa et al., 2008) but one that as a discovered variable can aid in describing new and important conditional and interactive effects.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript11According to measurement theory, it can be argued that for frequency scale coding, Codes 0, 1, 2, and 3 represent numerical counts, and for intensity scale coding these values represent categories on an ordinal scale. Treating these values as an interval-level Likerttype scale introduces the assumption of equal intervals between the values of 0, 1, 2, and 3, as these numbers would represent equal increments on the counts and on the levels of intensity for the frequency and intensity scale coding, respectively. We recognize that some investigators may question the assumption of equal interval levels as needed to satisfy parametric measurement assumptions for the use of interval scaling and the use of parametric statistics. For both modes of scale coding, frequency and intensity, we take this parametric approach wherein we indicate to our research assistants that the exempla.
