H567/03 Section B Option 4 - Kroll and Crenshaw (1970) and personality in sport

Background, aim and design

The prescribed research is Kroll and Crenshaw (1970), Multivariate personality profile analysis of four athletic groups. Earlier work had found no clear personality differences between some skill levels within wrestling or karate. The researchers wanted to extend within-sport research to a between-sport comparison. They also identified problems with poorly defined athlete samples and analyses that did not consider profiles adequately.

The aim was to examine whether four sporting groups could be distinguished using an entire profile, rather than testing each trait in isolation. The sports were American football, gymnastics, wrestling and karate. The football group must not be described as association football players.

Data came from a self-report questionnaire, the Cattell 16PF. The groups already existed: the researchers did not allocate athletes to sports or manipulate personality. This is an observational between-groups comparison using self-report measurement, not a randomised experiment proving that sport changes personality.

The study relates to OCR's personality topic in the cognitive area, because it examines structured descriptions of individual tendencies and their relationship to sporting participation. Individual differences, learning histories and social sporting environments offer possible explanations for group patterns. The measured profile does not establish the origin of a trait.

Sample and response screening

The final analysed sample was 387 athletes from the USA. It covered several teams and levels rather than a single local club.

The excellent wrestling group included athletes rated excellent by their coach who had won at least 60% of their season's matches. Defined sporting standards make the groups more interpretable, but college and selected competitive athletes do not represent every recreational participant.

The 16PF Form A was initially administered to 139 footballers, 141 gymnasts, 96 wrestlers and 71 karate participants, a total of 447. Additional testing included the 15-item MMPI lie scale. Responses scoring seven or above on that scale were excluded because their validity was questioned. The remaining numbers sum to 387, so 60 athletes were excluded overall.

Screening attempts to reduce distorted self-reports. It does not prove deliberate dishonesty or remove every source of bias. It also changes sample composition: the football group reduced from 139 to 81, whereas the gymnastics and karate counts did not change. Conclusions concern the retained athletes, and exclusion could affect representativeness.

The inspected OCR guide does not settle sample sex composition; OCR's 2024 report explicitly acknowledges conflicting accounts and accepts either an all-male or mixed-sex description in that assessment. Do not use an uncertain sex breakdown to justify a precise generalisation. Geographic, sporting and competitive-level restrictions are clear and sufficient to evaluate sampling.

Profile measurement and analysis

Each athlete completed Form A of Cattell's Sixteen Personality Factor Questionnaire. Standardised items and scoring produced quantitative profiles across sixteen factors. The tool assessed personality characteristics, not a direct measure of match success or physical skill.

The analysis used a multiple discriminant function to examine whether combinations of profile scores distinguished the sporting groups. “Multivariate” means considering several variables together. A combination may distinguish groups even when no one trait provides a complete sporting identity. The study did not simply sort people into “introvert” and “extravert” types.

The researchers then classified each athlete into the sport with the highest estimated probability of membership from the profile. Correct classification meant that this assigned sport matched the person's actual sport. It was not a prediction that the person would win a match or should change sports.

For the 81 actual footballers, 32 were classified as footballers, 18 as wrestlers, nine as karate participants and 22 as gymnasts. Those counts sum to 81. The football classification rate is $32/81\times100\approx39.5\%$. The substantial misclassification shows profile overlap rather than a unique personality for every football player.

Classification of the people in this dataset is also different from demonstrating accuracy on an independently recruited future sample. A practical talent-selection claim would need suitable validation against relevant outcomes, alongside skill and welfare considerations.

Results and contribution to personality in sport

The overall profiles differed significantly between groups. Six major contributors to the first discriminant factor were shy–venturesome, group-dependent–self-sufficient, trusting–suspicious, less intelligent–more intelligent, reserved–outgoing and affected by feelings–emotionally stable. “Intelligence” here is the historical name of a questionnaire factor, not evidence that one sport produces globally more intelligent people.

Overall correct classification was 50.6%. The sporting-group rates were:

Pairwise analyses found significant discrimination for all pairs except football and wrestling. Their profiles were relatively similar in this comparison, not identical for every individual. A nonsignificant difference is not proof of perfect equality.

Footballers and wrestlers scored highest in group dependence. Wrestling therefore challenged a simple rule that only team-sport participants would be group-dependent. This factor contributed significantly whenever either group was contrasted with the other sports.