H567/02 Section A - Measuring differences: comparing the pair
Gould (1982) and Hancock et al. (2011) show both the possibilities and dangers of turning human differences into scores. Compare the original measurement with the review of it, then judge the contemporary study's contribution and limits.
Similarities in measuring psychological differences
Both prescribed studies examine the link between a measurable indicator and an underlying psychological characteristic.
Gould reviews the use of army-test performance as an indicator of intelligence; Hancock et al. examine linguistic patterns in relation to PCL-R-assessed psychopathy. Neither a correct test response nor a category of words is the characteristic itself. Researchers must justify construct validity, the connection between indicator and intended concept.
Both involve quantitative evidence and group comparisons. The army programme reported mental-age summaries and racial/national rankings; Hancock reports frequencies and corpus proportions for pre-classified offender groups. Both therefore show how numbers can make comparisons concise while concealing individual circumstances. Gould's prescribed work is a later critical review, however, not the original administration of those tests.
Both raise social sensitivity: conclusions about intelligence or psychopathy can affect how people are treated. Gould discusses discriminatory uses of apparently innate rankings. Hancock's language findings could be misused if ordinary word choices were treated as a diagnostic rule. Social consequences do not automatically invalidate research; they increase the importance of accurate, limited interpretation.
Differences in evidence and control
The prescribed studies differ in method, participants and their relationship to the data being evaluated.
| Matched feature | Gould (1982) | Hancock et al. (2011) | Implication |
|---|---|---|---|
| Method of prescribed study | Historical critical review of Yerkes's testing | Empirical interviews followed by statistical content analysis | Gould critiques existing records; Hancock collects and codes new narratives |
| People represented | About 1.75 million men in the earlier US army programme | 52 male Canadian prison volunteers who admitted homicide; 14 classified, 38 controls | Large historical numbers do not fix bias; the contemporary sample is narrower and much smaller |
| Measure | Army Alpha/Beta and follow-up tests used to infer innate intelligence | PCL-R assessment for grouping; Wmatrix and DAL measures of language | Hancock separates a pre-existing assessment from the linguistic outcomes being investigated |
| Consistency | Variable allocation and follow-up in original camps | Standardised prompting, blind interviewers and explicit coding | Contemporary controls reduce certain procedural biases but leave memory and context problems |
A large sample is not automatically a superior measure, and computer analysis is not automatically unbiased. Alpha's cultural knowledge can contaminate its intended construct; a language classifier can consistently misread context. The useful contrast is which source of error each procedure addresses and which remains.
Ethically, military testing and restrictive social uses raise coercion and discrimination concerns. Hancock's participants volunteered and were told the purpose and procedure, but prison power relations, distress in recalling homicide and potential stigma remain. A comparison needs evidence for each side rather than labelling “old unethical, new ethical”.
Continue this lesson.
You’re previewing NovaLearn without an account. Create a free account to read the rest.
Create free account