Abstract The purpose of this study was to develop and assess a verbal, self-report, pencil-and-paper, personality inventory, the Wooster Personality Survey (WooPS). The WooPS was developed and assessed using the criterion- related validity measure of peer nomination. It was predicted that the subjects would show differential response patterns by criterion groups (low trait, high trait) on the appropriate trait subscales, but not on the remaining subscales. It was also hypothesized that items within subscales would have moderate to strong positive correlations, while between scales there would be low correlations. The results of the subjects (N=180) nominated by the researchers did show this differential response pattern on the appropriate subscales and, with only slight exception, not on the remaining subscales. Within-subscale item correlation was moderately to highly positive overall, with minor exceptions being the Optimism and Open- Mindedness subscales. Between-subscale correlation was positive but low for the most part. The discriminant and content validity of the WooPS was fair overall, but needs to be reviewed for future revisions. The issues of reliability, of face validity, and of social desirability bias were also addressed.

Verbal personality inventories have a long-standing history of use in both the research and the applications of the field of psychology (Anastasi, 1988). The purpose of this research was to develop a verbal self-report personality questionnaire which would measure five distinct, general, enduring, and characteristic traits: optimism, self-esteem, niceness, extroversion, and open-mindedness. However, in the development of a questionnaire such as the Wooster Personality Survey (WooPS), several issues regarding test validity needed to be considered. In developing the items for the WooPS, the issue of content validity was raised. Each item on the WooPS was judged by a group of 21 researchers as to its representativeness of the trait in question. Items reflecting a variety of behaviors and attitudes characteristic of each trait were chosen. This procedure had its basis in work done by Church and Katigbak (1989), Paunonen, Jackson, and Keinonen (1990), and Watson and Clark (1991). The work of these researchers also supports the measure utilized by the present study to create a criterion-related validity for the WooPS. These researchers employed the strategy of peer nomination, a type of rating (Anastasi, 1988), in order to establish criterion-related validity for the WooPS. Church and Katigbak (1989), examining a sample of 168 Filipino students, found that external measures (peer nomination) and self-report measures were positively correlated. Examining both verbal and non- verbal personality inventories, Paunonen, Jackson, and Keinonen (1990) reported similar findings. There was a mean correlation of r = .21 between rater reports and self-reports of personality information; however, at the highest levels of acquaintanceship between raters and subjects (ratees), these correlations have a mean of r = .40 for both verbal and non-verbal scales.

Watson and Clark (1991) also examine this issue of self- rating/peer-rating correlation. In their sample of 150 students, self- rating/peer-rating agreement correlations ranged from r = .19 to r = .41, with a mean coefficient of .27. Watson and Clark also discussed the issue of acquaintance effect, as was found by Paunonen, Jackson, and Keinonen (1990). Watson and Clark found that higher convergent correlations resulted from the best acquainted subject-rater dyads. Based on this research into peer nomination and self-report, it was decided that the test of the criterion-related validity of the WooPS would be peer nomination. The WooPS appeared to be strong in both content and face validity; however, it was necessary to check subjects' answers to the WooPS items against the criterion of peer nomination. If the WooPS was truly a well-developed personality survey, well-known friends and acquaintances nominated by the researchers as being either high or low in a given trait would answer appropriately to those items on the WooPS related to that trait, thus presenting evidence of the criterion-related validity of the WooPS.

Therefore, when analyzing the responses made by researcher (peer) nominated subjects to the items on the WooPs, this study predicted several outcomes. The first hypothesis was broken down into two related sections. The first part stated that there would be differential response patterns by criterion groups (low trait, high trait) on the appropriate scales. For example, there would have been significant differences in the Optimism subscale scores for low optimism and high optimism nominated subjects. The second half of this hypothesis stated that there would not be significant differences in response patterns by criterion groups on any of the remaining four subscales for which they were not nominated. For example, Niceness subscale items should not have distinguished high optimism subjects from low optimism subjects. In this way, the WooPs would have demonstrated discriminant validity, differentiating between the various subscales.

A second hypothesis was also proposed which stated that items within a given WooPS subscale would correlate positively. This would have evidenced the content validity of the WooPS subscale items. Related to this hypothesis, it was also stated that, between subscales, there would be low item correlations.

Subjects The subject group for this study consisted of 210 participants, of whom 180 were included in the final data analysis. The majority of the subjects were college students enrolled at a small, Midwestern, liberal arts college. However, the subject group also included relatives and older, non-student colleagues of the researchers. The subjects ranged in age from 17-72. One-hundred and twenty-one of the subjects were female, and 59 of the subjects were male. Subjects were recruited by the researchers to participate. The subjects were not paid for participation, but volunteered their time to the researchers.

Research Materials

The research materials consisted of the Wooster Personality Survey (WooPS), a 20-item, pencil-and -paper, self-administered inventory of personality traits. The WooPS was developed by the researchers for the purpose of this study. One version of the WooPS is presented in Appendix A.

In order to develop the WooPS, the researchers began by discussing traits that they believed were isolated in nature, general, enduring, characteristic, and functional. In the end, the researchers selected five traits they felt would be desirable in a "date". These traits were "Optimism", "Self-Esteem", "Niceness", "Extroversion", and "Open- Mindedness". The researchers then divided themselves into five groups. Each group examined one of the five traits and developed four simple statements regarding that trait. In order to discourage any response sets on the part of the subjects, two of the statements were designed so that they would be endorsed, if a subject possessed the given trait. The remaining two statements would not be endorsed if the subject possessed the trait. The statements were created so that they could be responded to with either an endorsement of "True" of "False" depending on whether or not the subject felt that a given statement described her/his behavior and/or attitudes most of the time.

The researchers as a whole then judged the statements in order to ensure that they were clearly worded and appeared to measure the trait in question. Certain statements measuring the trait "Niceness" were modified to be less absolute in nature. For example, the statement "I would always help somebody who asked for directions." was modified to state, "I would usually help somebody who asked for directions." After all 20 statements were approved by the group, each researcher was allowed to develop his or her own version of the WooPS with the stipulation that no two items from the same trait subscale could appear consecutively. All versions of the WooPS were identical in content; they varied only in the item order.

Procedure

Each researcher selected 10 subjects to participate in the study. These subjects were selected based on peer nomination by each individual researcher. Each of the 10 subjects exemplified either a high or low presence of one of the 5 traits. For example, each researcher selected both a participant they believed to have high self-esteem and one who possessed low self-esteem.

Each potential participant was approached by the researcher and asked to participate in filling out a questionnaire examining personality traits. The participants were assured that their answers would be kept confidential and anonymous. Each subject was then instructed to carefully read each statement and respond "True" if the statement described his/her attitudes and/or behavior most of the time, or "False" if the statement did not. The WooPS was administered individually with each subject taking no more than 10 minutes to complete the inventory and return it to the researcher. Upon completion of the WooPS, each subject was thanked for his/her participation. The subjects were not debriefed due to the sensitive nature of the participant selection. The results of the WooPS were then tallied. The possible responses for each item were coded as to whether or not they were in the direction of the trait being measured. For example, if a subject answered "True" to the statement, "In time the economy will get better." this response was coded with a "1" to indicate that the subject responded in the direction of the trait "Optimism". However, if the subject had responded "False" to the same question, this response would have been coded with a "0" to indicate that the subject did not endorse the optimistic answer choice. The highest possible score for a trait was a "4"; the lowest possible score on a given trait was "0".

Results were calculated for 180 subjects. Using the technique described above, subjects were scored on the trait for which they were nominated, along with being scored on the other four traits, as to whether or not they were high or low in each trait. Table 1 presents the mean scores for each trait for those subjects who were peer-nominated as either high or low in that trait.

------------------------------------------------

Insert Table 1 about here

------------------------------------------------

In order to ascertain whether or not these differences in trait scores were significant, analysis of variance tests were performed on these data. A main effect was found for the trait "Optimism", F(1,34) = 20.39, p < .05. Subjects nominated as high in optimism scored significantly higher on the Optimism subscale than those subjects nominated as low in optimism. A main effect in the same direction was also found for the trait "Self- Esteem", F(1,34) = 6.15, p < .05. Similar effects were also found for the trait "Niceness", F(1,34) = 11.52, p < .05. A main effect was found for the trait "Extroversion", F(1,34) = 18.86, p < .05. Subjects nominated as extroverted scored significantly higher on the Extroversion subscale than those subjects nominated as introverted. A main effect in the same direction was also found for the trait "Open-Mindedness", F(1,34) = 11.14, p < .05. These analyses evidenced that there were differential response patterns by criterion groups (high trait, low trait) on the appropriate subscales.

However, there should not be significant differences in the response patterns by criterion groups on subscales other than the trait subscale for which the subjects were nominated. For the most part, this was the case. For example, the difference in mean Self-Esteem scores for subjects nominated on the trait "Optimism" was not significant, F(1,33) = 1.10, p > .05. Self-Esteem subscale scores did not differentiate between high and low optimism subjects.

However, there were certain exceptions to this general case. Scores on the Niceness subscale were significantly different for subjects nominated on the trait "Self-Esteem", F(1,34) = 4.75, p < .05. Subjects nominated as high in self-esteem scored significantly higher on the Niceness subscale than those subjects nominated as low in self-esteem. Scores on the Self-Esteem subscale were significantly different for subjects nominated on the trait "Extroversion", F(1,34) = 5.13, p < .05. Scores on the Open-Mindedness subscale were also significantly different for "Extroversion" nominated subjects, F(1,34) = 4.35, p < .05. Subjects high in extroversion were significantly higher in self-esteem and more open-minded (as measured by the questions in these subscales) than subjects nominated as low in extroversion.

A Pearson's correlation matrix was then generated for the item scores in each subscale. There should be moderate to high positive correlations between the items in each subscale. The items in the Optimism subscale had a range of correlation coefficients (r) from -.05 to .16. Correlation coefficients for the items in the Self-Esteem subscale ranged from .05 to .40. The range of coefficients in the Niceness subscale was .10 to .46. Correlation coefficients for the items in the Extroversion subscale ranged from .18 to .41. The items in the Open-Mindedness subscale had a range of correlation coefficients from .01 to .27. Overall, within subscales, WooPS items showed positive, moderate to high correlations, with the Optimism and Open-Mindedness subscale items evidencing the lowest within-subscale correlations.

A Pearson's correlation matrix was also generated for the mean subscale scores. This was done in order to analyze the correlations of items between subscales. There should be low correlations between each subscale and the remaining four subscales. This matrix is presented in Table 2.

------------------------------------------------

Insert Table 2 about here

------------------------------------------------

The lowest correlation was between the Niceness and Extroversion subscales (r=.08). The highest correlations were found between the Self- Esteem and Extroversion subscales (r=.29), and between the Niceness and Open-Mindedness subscales (r=.37). All correlations between subscales were fairly low, positive correlations.

In this sample (N=180) of well-acquainted peers of the researchers, it was found that scores on the Wooster Personality Survey (WooPS) evidenced differential response patterns by criterion groups (low trait, high trait) on the appropriate subscales. Each trait subscale was found to reliably discriminate between subjects who had been peer nominated as low in the respective trait and those subjects who had been nominated as high in that trait. This finding gave support to the discriminant validity of the WooPS. This finding of differential response patterns in the expected direction for each trait subscale also lent further support to the strength of peer nomination (peer rating) as a viable means of criterion-related validity. Those subjects nominated by the researchers as either high or low in a given trait performed in the direction of those nominations. Their actual performance on the WooPS was similar to the performance that had been predicted by the researchers. In this manner, the WooPS also showed acceptable criterion-related validity.

Related to this hypothesis, it was also found that, for the most part, any differences in response patterns by criterion groups on trait subscales for which the subjects had not been nominated were not significant. For example, mean scores on the Self-Esteem subscale were not significantly different for those subjects rated as either high or low in the trait "Optimism". The Self-Esteem subscale scores did not differentiate between low optimism and high optimism subjects. Aside from evidencing further the discriminant validity of the WooPS, this finding added support to the isolated nature of each trait. "Self-Esteem" and "Optimism" were traits which appeared to be fairly independent of one another. However, there were a small number of exceptions to this finding. The first of these is that the Niceness subscale differentiated between subjects nominated on the trait "Self-Esteem". Those nominated as high in self-esteem received reliably higher Niceness scores than those subjects nominated as low in self-esteem. This finding may reflect a tendency for those subjects low in self-esteem, who may have less respect for themselves, to have less respect for others. Conversely, those high in self-esteem, who feel good about and respect themselves, have a heightened concern for others which is expressed through behaviors similar to those found on the Niceness subscale.

The Self-Esteem and Open-Mindedness subscales were also shown to differentiate between those subjects nominated on the trait "Extroversion". Subjects peer-rated as high in extroversion self-rated as higher on both the Self-Esteem and the Open-Mindedness subscales of the WooPS. This may suggest that the trait of "Extroversion" is not an independent trait, but is strongly influenced by one's degree of self- esteem. Those higher in self-esteem may be more confident in their opinions and abilities than those low in self-esteem. This confidence allows them to be more outgoing, to express their beliefs and feelings, to see themselves as valuable to others and confident in approaching them; in other words, to be more extroverted than those low in self-esteem. However, the relationship between extroversion and open-mindedness is not as apparent. Further research needs to be done examining the relationship between these two traits in order to attempt an explanation of the relationship between the two as evidenced by the WooPS. The second hypothesis proposed by this study predicted that WooPS items within-subscales would correlate positively, lending support to the content validity of the WooPS. Overall, items within-subscales showed positive, moderate to high correlations. The lowest correlations were between items in the Optimism subscale and between items in the Open-Mindedness subscale. In further revisions of the WooPS, it will be necessary to examine and revise the items contained in these subscales in order to increase the correlations of items within each subscale, thereby increasing the content validity of these subscales.

Examining item correlations between the subscales, the Self- Esteem and Extroversion subscales had strong positive correlations, as compared to the correlations between the remaining subscales. The only stronger correlation was between the Niceness and Open-Mindedness subscales. The correlation between the Self-Esteem and Extroversion subscales was most likely due to the role self-esteem may play in extroversion, as described above. The strong correlation between the Niceness and Open-Mindedness subscales may be a result of the structure of open-mindedness. People who are open-minded may have a greater sense of respect, concern, and tolerance for the various members of the human race and all their differences. This heightened respect and tolerance for others may be evidenced in behaviors typical of the trait "Niceness". In future revisions of the WooPS, items on these subscales would have to be reworked so that they test more exclusively for the appropriate trait in question.

The WooPS also evidenced face validity. That is, it appeared to be testing what it was indeed testing. However, a problem arises with this apparent face validity. As Paunonen, Jackson, and Keinonen (1990) remind researchers and practitioners who utilize both verbal and non- verbal personality tests, the response bias of social desirability is one of the "certain pervasive problems that tend to plague" personality inventories (p.501). It is entirely possible that the scores of those subjects nominated as low in one of the five traits, although in the direction expected, were not accurately low due to the response bias of social desirability. Each of the five traits examined in this study is a trait which appears to be valued by others and by society; therefore, subjects may have felt a need or desire to appear more socially acceptable. The results of this study shed a favorable light on the future of the WooPS. However, the issues raised concerning discriminant and content validity must be addressed. Previous research on similar measures should help provide insight into these areas. One of the problems with the WooPS may have been that the choice of traits to be investigated was not researched thoroughly in order to determine the relative independence of the traits. Measures of reliability also need to be performed on the data generated by the WooPS. The WooPS needs to be not only valid, but reliable, in order to be an effective research or clinical tool.

However, care needs to be taken in utilizing the WooPS, as with any other psychometric instrument. It needs to be remembered that the results of the survey are based on only a selection of all the possible attitudes and behaviors representative of a given trait. Situational factors also need to be considered. A given subject may behave in one manner in a given situation and in a completely different manner in a contrasting situation. In other words, the results of a personality inventory such as the WooPS should not be utilized as a sole criterion in making research or clinical judgments. The use of a given psychometric measure should always coincide with other diagnostic techniques, such as additional psychometric instruments or a diagnostic interview.

Anastasi, A. (1988). Psychological testing (6th ed.). New York: Macmillan.

Church, A. T., & Katigbak, M. S. (1989). Internal, external, and self-report structure of personality in a non-Western culture: An investigation of cross-language and cross-cultural generalizability. Journal of Personality and Social Psychology, 57(3), 857-872.

Paunonen, S. V., Jackson, D. N., & Keinonen, M. (1990). The structured nonverbal assessment of personality. Journal of Personality, 58(3), 481-502.

Watson, D., & Clark, L. A. (1991). Self-versus peer ratings of specific emotional traits: Evidence of convergent and discriminant validity. Journal of Personality and Social Psychology, 60(6), 927-940.

Age:_______ Sex:_______

In time the economy will get better. T F

I like myself. T F

I would usually help somebody who asked for directions.T F

When entering a party, I am able to introduce myself to a group of strangers. T F

I believe there is never a situation where the law should be broken. T F

I see half a glass of water as half empty. T F

I tend not to share my ideas and opinions with others.T F

If someone in front of me drops their groceries, I just ignore them. T F

When in class, I voluntarily offer my knowledge. T F

Life should be lived in a certain way. T F

If you treat everyone with dignity and respect, they will treat you the same. T F

When criticized, I don't let it affect the way I perceive myself. T F

When stopped in traffic, I usually let people merge in front of me. T F

During class debate, I sit back quietly and keep my opinions to myself. T F

Every culture has a right to flourish in the United States.T F

I feel as if I will never succeed. T F

When I make a mistake, I put myself down. T F

I usually do not put my change in charity boxes at stores.T F

Given the choice of going to a party with my friends or staying home alone to watch a movie, I would choose to stay alone. T F

The feminist movement is just as much a human movement as one that deals primarily with women. T F

Table 1 Mean Trait Subscale Scores as a Function of Peer Nominated Criterion Groups.

______________________________________________________