Should Gender be Measured as a Uni-dimensional or Multidimensional Construct?

When we describe gender differences and sex roles conceptions, we often relate them in the respect of dimensions. At the earliest stage of gender studies, men and women were classified as containing either one of the two major dimensions, masculinity and femininity. Masculinity contained traits that reflected assertiveness, boldness, dominance, self-sufficiency, and instrumentality; whereas femininity contained traits that reflected nurturance, expression of emotions, and empathy. It was thought that when one scores high on either dimension, the other dimension must be low, and thus it was called a Uni-Dimensional study.

The 1970s feminist movement sought to emphasize the idea that gender is not a single dimension, but instead, that masculinity and femininity are independent dimensions (Larsen & Buss, 2010). Expanding the narrow dimension to include androgynous and undifferentiated in addition to masculine and feminine allows for a more thorough assessment. Working with a more multidimensional approach to measure gender may also assist researchers in applying their gender identity model to other cultures. The less dimensional a measure is, the more subject it may be to measuring stereotypes, specifically Western ones.

Our class worked within the Bayesian approach and used a Gender Diagnosticity Questionnaire in which it provided us with a better approach to gender differences rather than merely looking at it as either masculine or feminine, in terms of a multidimensional construct. Throughout the questionnaire we had questions that addressed masculine or feminine actions in which participants were free to choose from, giving them the opportunity to best reflect their preferences regardless of their gender stereotypes. So individuals could score either high or low on one of the dimensions, and also on both dimensions, representing the multidimensional construct.

One-Dimensional Construct In 1922, Terman and Miles noticed that there were considerable gender differences of intellectually gifted children in the play and activity preferences. They started to carry out research and eventually established a multifaceted masculinity-femininity test (Lippa & Connelly, 1990). Terman and Miles believed that there are different reactions to several personality items between men and women. For example, men and women displayed different responses when they were asked whether they preferred to take bath or shower (Larsen & Buss, 2010). Personality researchers also asserted that the differences could be described by a single personality continuum, with masculinity at one side and femininity at the other side, and all people could be found on this single masculinity-femininity dimension. In this unidimensional approach, researchers declared that an individual who scored high on masculinity was presumed to score low on femininity, and vice visa (Larsen & Buss, 2010).

Multidimensional Construct With the rise of the feminist movement in the early 1970s, many researchers began to challenge the original assumption and started unfolding the premise that masculinity and femininity were two independent dimensions. Thus, one could be high on both masculinity and femininity dimensions or be low in both dimensions. An individual could be stereotypically masculine, which is high in masculinity and low in femininity. An individual could also be stereotypically feminine, which is high in femininity and low in masculinity (Larsen & Buss, 2010).In 1973, Constantinople asserted that rather than the traditional unidimensional scales by Terman and Miles which overly related to several demographic elements, such as social status and age, and unduly defined in respects of gender stereotypes, masculinity-femininity scales should be multidimensional measures (Lippa & Connelly, 1990).

By using multidimensional scales, gender roles can be described as masculine, feminine, androgynous and undifferentiated.

Masculinity: A dimension consists of items indicating “assertiveness, boldness, dominance, self-sufficiency, and instrumentality” (Larsen & Buss, 2010).

Femininity: A dimension consists of items indicating “nurturance, expression of emotion, and empathy” (Larsen & Buss, 2010).

Androgynous: A person who is high on both masculine and feminine characteristics (Lippa & Connelly, 1990).

Undifferentiated: A person who is low on both masculine and feminine characteristics (Lippa & Connelly, 1990).

Examples of how the Multidimensional Construct Works

Multidimensional construct has been wildly used in gender differences studies. Mark J Krejci had used the multidimensional scaling analysis (MDS) in an attempt to verify whether there were gender differences in their performance in relation to the God image (1998). MDS is a methodology which allows the experimenters to collect data by asking participants to determine the similarities between different stimuli rather than asking individuals to rate along experimenter-derived rating scales (Krejci, 1998). Therefore, as Krejci pointed out, MDS allows minimum experimenter contamination which may create by the participants when they use rating scale to answer questions (1998). In his study, with the assistance of under graduate students from a church-affiliated liberal arts college, participants were recruited with a Christian’s background-- group one had 128 females and 87 males with an average age of 37.2 and group two had 84 females and 54 males with an average 33.03 (Krejci, 1998). Both of the two groups were asked to complete a card sort of 27 words (judge, protector, savior, mother, master, redeemer, creator, father, gentle, stern, distant, demanding, patient, infinite, loving, forgiving, perfect, caring, just, spirit, deliverer, divine, comforting, glorious, supporting, merciful and gracious) which are used to describe God on three dimensions—nurturing-judging, controlling-saving, and concrete-abstract (Krejci, 1998). The first group of participants were asked to read and arrange similar items of God into as many piles as they wanted, and the second group of people were asked to order the 27 word of God images in terms of how well the image represented their image of God (Krejci, 1998).

A total of 215 valid survey were analyzed, and as the results showed, a three dimension solution best reflected the collected date from both groups for both gender—it not only indicated similarity and differences between the two genders of each three identified dimensions, but also, it revealed the similarity and differences between the two groups (Krejci, 1998). Therefore, Krejci concluded that the two matrices could be analyzed at the same time based on the consistent of the dimensional organization of the stimuli between the two groups (1998). Comparing male and female groups, the date analysis also provides subject weights which indicate the relative salience of each dimension to the group. As the data indicates, only the second dimension (Dimension 1, male=6844, female=6483; Dimension 2, male=5399, female=6160; Dimension 3, male=4051, female=3839) appeared slightly higher weighted by women than men (Krejci, 1998). The weight pattern shows that in the three dimensions, only the judging-nurturing dimension is the most salient for both groups, and the controlling-saving dimension and concrete-abstract dimension showed similar weighted for women versus men (Krejci, 1998). At last, Krejci(1998) used MDS analysis found out that, although, men and women are similar in terms of their two God schemas, which was organized into categories such as judging-nurturing and concrete-abstract, the salience of these dimensions does show date that indicate gender differences on controlling-saving of God image. The results confirmed that a three- dimension solution is the best way for analyzing the date for men and women.