Course:PSYC305/2013ST2/ClassProject/3.3.3 Method - Procedure - Data Analysis

Data analysis

After the questionnaires were completed, data was statistically analyzed for gender differences based on the questions answered by all participants, the gender distribution for all the participants, final items of Gender Diagnosticity and the Big Five.

In gender differences, the calculation for Cohen's d was performed to see how far the effect size was from the mean. Based on Cohen's d, the values were calculated by taking the differences of the mean over the standard deviation (Rice and Harris, 2005). The positive d values indicated the questions that were scored higher among males and the questions scored higher by females were given a negative d value. the d score of ±0.20 indicates a small difference, ±0.50 indicates a medium difference and ±0.80 indicates a large difference (Larsen and Buss, 2010). If the d score is 0 which indicates there is no gender difference for that particular personality trait. However, the d score is not necessarily imply the preference/performance of the trait for any particular individual (Larsen and Buss, 2010). For example, the d score indicates a large sex differences (eg. d=1.54) in "Hobbies - modifying cars", some girls may still love to modify cars as much as boys do.

In terms of final items of Gender Diagnosticity, the factor analysis was performed to see the variation between different genders for each item in gender diagnosticity. Factor analysis was used to describe how closely related the groups of items were (Larsen and Buss, 2010). The factor loading value was used in this case to represent the level of variation each item was explained (Larsen and Buss, 2010). The positive factor loading value indicated that the item was scored higher by males and negative value for factor loading indicated that the item was scored higher by females.

Correlational method was used to analyze Big Five (openness, conscientiousness, extraversion, agreeableness and neuroticism) in order to see the relationship between participants' personality traits and their gender diagnosticity score. Same method was applied to age and the scores on last 25 items in gender diagnosticity. To calculate correlation, significance testing (i.e., p value) was applied to see how significant the value is. Only data of p<0.05 is considered to have significance (Larsen and Buss, 2010). Pearson's r was performed to reveal the linear relationship between Big Five Traits/ age and the gender diagnosticity score, with positive r indicating positive relationship, vice versa. Data of r value close to 0 was also considered to have no significant relationship (Cohen & Cohen, 1975).