Personality Assessment From Archival Data: Using the Q-Sort Technique

Your Name

Rutgers University

This study was designed to test whether the Q-Sort technique is a good measure for a relatively untrained assessor to use for determining an individual’s personality from archival data. The basic procedure of the study was to characterize two people’s personalities using the elements of personality on the California Child Questionnaire (CCQ) cards, basing personality decisions on how the person was described by others. Predictions were made on each of two individuals as to what degree they possessed qualities of the Big Five dimensions of personality. To determine if the Q-Sort was reliable, inter-observer agreement was measured between the coding done by the researcher and the coding done by an expert.

Method

Participant

The untrained evaluator was a female undergraduate psychology major with some exposure to personality theories in other classes. Anonymous descriptions of two children, one boy and one girl, were used to give the researcher a glimpse of their personality with which to base the Q-Sort on.

Materials

The descriptions of the two children came from interviews done with the children prior to this study. For each child, three interviewers wrote very detailed reports of their interaction.

The CCQ cards used to characterize the two children were constructed using 109 labeled index cards, 100 of which were descriptor cards, and 9 of which were category cards. The category cards were labeled with a range from 1 to 9, where 1 was “Extremely Uncharacteristic” and 9 was “Extremely Characteristic”. The descriptor cards contained different aspects of personality, along with an assigned identifying number, ranging from 1-100.

[The Microsoft Excel program was also used to input the data and calculate the correlation coefficients for inter-observer agreement, and also to calculate the degree of the Big Five elements both individuals possessed, according to how the researcher categorized them. ][Note: This is an example of irrelevant information].

Design and Procedure

The researcher started out by reading the three different interviewer’s descriptions of the boy. After reading their descriptions, the researcher then sorted the 100 descriptor cards into the 9 category piles, based on how well or poorly she thought the description on the card matched the personality of the boy. This was done for all 100 cards. The only restriction of the sort was that 8 of the categories could only have 11 cards at the end of the sort, and the middle, neutral category had to have 12 descriptor cards. [After finishing the Q-Sort, the Microsoft Excel program was used to enter all the data into a spreadsheet where column one was the card number and column two represented the category number that the card was placed into.][Note: The preceding is an example of description that should not be put in the report, because it is irrelevant.] The above process was then repeated for the girl.

After the doing the Q-sort, the researcher made predictions, based on the interviewers’ descriptions of the individuals, of whether the boy and girl was high, low, or in the mid-range for the Big Five dimensions of personality: neuroticism, extraversion, openness, agreeableness, and conscientiousness. Predictions were also made as to whether the researcher thought the individual was either high or low in ego-control and ego- resiliency. The researcher then compared her Q-sort to the Q-sort of an expert done on the same individuals, using the Pearson correlation coefficient to determine inter-observer agreement. The interobserver agreement was the measure of the reliability of the Q-Sort technique for measuring personality. A Pearson correlation was also done to determine if the individual was either high or low in ego resiliency or ego control. This was done by correlating the researcher’s Q-sort with the Q-sort for a prototype for an ego-resilient and ego-controlled person.

Results

For the male subject, the inter-observer agreement between the researcher and the expert sort was .30, a result that indicates low reliability. This means that the sorting done by a novice did not strongly correlate with the expert. For the female subject, the inter-observer agreement was .74, a result that indicates reasonably good reliability. For this case, the sorting done by the novice was very similar to the results produced by the expert.

The predictions for the boy for ego-control and ego-resiliency were: ego-control: high and ego-resiliency-low. For these measures, the correlation between the prototype and the data obtained in the experiment for the boy was: -.38 for ego-control and .005 for ego-resiliency. This means that the boy was the opposite of the prototype for ego-control since the result came up negative, and that there was no evidence found that the boy possessed ego-resiliency, since the result came up very close to zero. When compared to the predictions, the data show that the measures obtained from the sort are just about opposite from what was predicted. For the girl, the predictions in ego-control and ego-resiliency are as follows: ego-control-mid, ego-resiliency-high. The correlation between the prototype and the data from the experiment is as follows: -.02 for ego-control and .62 for ego-resiliency. This means that there was a weak, negative correlation between the prototype for ego-control and the data obtained for the girl, showing she was not judged to possess ego-control. For ego-resiliency, the correlation was a strong, positive relationship, meaning that the girl possessed characteristics of an ego-resilient individual. As compared to the predictions, the measure for ego-control showed the opposite of what was predicted, and the ego-resiliency showed the same result as what was predicted. For the Big Five personality dimensions, the predictions for the boy were: neuroticism- high, extraversion- low, openness- low, agreeableness- mid, conscientiousness- high. As categorized by the researcher, the data show for the boy shows: neuroticism- 5.5, extraversion- 2.3, openness- 4.9, agreeableness- 5.6, and conscientiousness- 5.4. The results were not that close to the predictions; the boy came up lower on neuroticism than expected and higher on openness than expected. For the girl the Big Five personality predictions were as follows: neuroticism- low, extraversion- high, openness- high, agreeableness- high, and conscientiousness- high. As categorized by the researcher, the data for the girl shows: neuroticism- 3.5, extraversion- 7.3, openness- 6.7, agreeableness- 6.7, conscientiousness- 7.0. The results for the girl were more close to the predictions than were those for the boy. The results for openness and agreeableness were expected to be higher, however.

The greater agreement of the predictions for the girl than of those for the boy is consistent with the differences in reliability of the evaluation.

Discussion

Based on the results obtained in this study, it can be concluded that the Q-Sort can be used as an assessment of personality by novice evaluators. The results for the Q-sort on the girl, for example, showed good inter-observer agreement and consistency. However, the Q-sort was not as promising for the boy, with low inter-observer agreement and limited consistency. The reason for this may have been that the interviewers’ descriptions for the girl were much clearer and easier to understand than those for the boy. This may have led to some incorrect sorting for the boy, which would explain the low inter-observer agreement. On the whole, the Big Five dimensions came out relatively close to what was predicted. The inter-observer agreements did not show strong correlations, probably because the data obtained was being compared to a sort done by an expert, who would have much more experience than someone who has done this for the first time.

Table 1:Female
Big 5 Dimensions	Predicted	Actual
Neuroticism	low	3.5
Extraversion	high	7.3
Openness	high	6.7
Agreeableness	high	6.7
Conscientiousness	high	7

Table 2: Male
Big 5 Dimensions	Predicted	Actual
Neuroticism	high	5.5
Extraversion	low	2.3
Openness	low	4.9
Agreeableness	mid	5.6
Conscientiousness	high	5.4