Previous Measures : How can we improve?

Great points, and I definitely agree. Also, was our method for developing a GD scale the best way to go about it? I think there were some good things about the way we did it, but perhaps some things that could be improved as well for future research.

Hi all, I would like to address again the importance of addressing the limitations of this study in order to provide solutions for future research. When considering future research you are considering the possible improvements that you can make to the study in order to increase both the reliability as well as the validity of the study. Below, I have reworded a couple of the limitations that have been collaborated by the class, and included possible resolutions to these limitations which could aid us in future research.

As mentioned, the overlapping limitation of this study was that a self-report was the only procedure of measurement utilized. Although the use of a self-reports provides ease as well as well as access to information that is reserved only by the individual, its limitations should be considered and the study should be revised for future research. Through the revision of the limitations which this study was hindered by, one is able to develop a better and more accurate study through which gender diagnosticity can be examined. The limitations that should be considered in future research are as follows:

1. When questionnaires ask questions which are private for a particular individual, that individual may respond in a “socially desirable” manner instead of truthfully. To answer a questionnaire in a socially desirable manner is also known as faking good. There are however possible solutions to this limitation that could be considered in future research. It is important to note however that anonymous questionnaires don’t often pose the problem of individuals answering truthfully due to the privacy of the individual being assured.

• One way in which one is able to overcome the problem of individuals answering the questionnaire in a socially desirable manner is to impose a “faking good profile” (Larsen & Buss 2010). This profile is constructed to use in comparison to the responses of an individual to determine whether he or she is “faking good.”

2. This study was representative of ethnicity, religious belief, level of education, and personality, but lacked in the representation of gender. The ratio of individuals who reported their genders were 10 males to 50 females. Due to the limitation of equal gender representation this study is only reliable when distributed to a higher population of BOTH females to males. Due to the lack in reliability and due to the study not measuring what it claims to measure, this study cannot be claimed as reliable or valid.

• In order to define this study as both valid and reliable we need to consider two things 1) that the study measures what it claims to be measuring (validity), and 2) that the study will yield the same results when tested on other populations or sample groups (reliability)
• these two points can be resolved by a sample size that is 1) not biased (for example the sample should not consist of the psychology students constructing the study) 2) representative of both females and males (equal or close to equal number of males and females within the sample) 3) randomly selected to control for other factors such as ethnicity, religious beliefs, educational level, etc.

The first step in future research is resolving these as well as other limitations to the current scale. Once the limitations have been resolved the many interesting and important factors that everyone has suggested could be researched in the future. Once you have a reliable and valid scale on which gender diagnosticity can be measured one will be able to add specific measures to investigate, such as how “gender and personality traits may have great influences on mental disorders,” or determining the correlation between “life expectancy[,] gender and personality traits,” and the rest of the suggestions which have been contributed to this page.

Hi everyone, As mentioned before, an aspect of future research that should be considered is how to revise a study in order to raise its validity and reliability, which will give us more accurate results. I have some additional ideas on solutions to fixing the limitations that were mentioned by GennaFerreira.

1. Solutions to reduce social desirability I found the following methods in the following article: Nederhof, A. (1985). Methods of coping with social desirability bias: A review. European Journal of Social Psychology, 15, 263-280.

• Neutral Questions
  • We should try to make questions that are as neutral as possible, so that it may not be seen as something that a person should answer in a specific way (a socially desirable way) by many different social groups.
• The Bogus Pipeline
  • Participants are told that either an attitude or emotion is being analyzed by a machine, or by a polygraph. However, the wires that are attached to the participant are not real. Since the participant does not know this, they will most likely answer the questions honestly.

2. Solutions to having a more valid and reliable study

• The sample size needs to be much larger; having a few hundred will be more effective than having less than a hundred. This will help eliminate individual differences. Moreover, by having a larger sample, we will be able to have a more equal number of females to males.
• As GennaFerreira mentioned, having random sampling will create less differences between individuals. Therefore, we should be using the random sampling technique to include not only Psychology students but the general public as well.
• In order to increase reliability, this questionnaire should be distributed more times to different samples to see if the results are still the same between different groups (ex. between students, parents etc.). We can also ask multiple questions about the same idea, in order to see if participants are answering it in the same way throughout the questionnaire (I believe we did have a few of these throughout the questionnaire).

3. Solutions to Experimenter/Confirmation Bias and Content

The overarching solution to confirmation bias/content is to develop questions from different perspectives (as in, consider culture, geographic location, SES, age/cohorts, etc) to provide the most objective, all-encompassing questionnaire. The questions were presumably developed by individual students. Although there was some convergence in ideas, the questionnaire could have been streamlined more efficiently. Instead, we could develop questions using either (1) a panel consensus method, where questions are considered as a group or (2) a structured delphi method, where individuals in a panel develop their own questions, which are then consolidated and refined repeatedly until ideas are stabilized and the questionnaire is where we want it to be (eg. objective, representative of different perspectives, provides a variety of possible responses).

4. Language

Aside from translating the questionnaire to multiple languages, I think either question development method as mentioned above could potentially address the language barrier and English comprehension issue.

Objectivity also plays into the neutrality of the questions, as PhoebeDychinco stated would address the social desirability/negative connotations issue.

Hi Jaimie,

when I read Genna's ideas about "when determining gender diagnosticity and therefore our sample should include an equal, or close to equal sample of BOTH males and females." I was wondering does sexual orientation effect the results of gender diagnosticity? That is to say, the more "masculine" role in Lesbian couple, compared with the more "feminine" role in Gay couple? Is this a related part need to be concerned in future research?

Hi SunnyZHENG,

I was hoping someone would bring this up! I summarized a study that looked at heterosexual and homosexual men and women: http://wiki.ubc.ca/Course:PSYC305/2013ST2/ClassProject/2.2.1_Introduction_-_GD_Studies#Gender-Related_Traits_of_Heterosexual_and_Homosexual_Men_and_Women

The same occupational preferences that distinguish men from women are the same that distinguish heterosexual men from gay men and lesbian women from heterosexual women. So yes, lesbian women are more likely than heterosexual women to be self-ascribed as masculine, while gay men are more likely than heterosexual men to be self-ascribed as masculine. I'm not entirely certain whether or not their responses would skew the overall results, but I do think it is relevant for future research.

I was reading over my notes/text book from my human sexuality class, which states that gay men and lesbian women are overrepresented in certain occupational areas (the creative arts, and professional sports, respectively). Do you think that a GD or homosexuality-heterosexuality diagnosticity (which was also shown to be just as effective in the same study), would be a stronger predictor of over- or under-representation of the homosexual community over other demographic factors such as educational attainment, SES, or background?

Hi Schuolee,

thanks for your reply and guide me to the related previous studies on sexual orientation and gender-related traits. Firstly, I have to say I have read an article that states many women become homosexual due to genetics. There still have gender differences on homosexual heredity, which may be applied to study homosexual community as you stated above.

Ok, so back to the topic. "Previous studies found that gay men are more feminine than heterosexual men, while lesbian women are more masculine than heterosexual women (Dunne, Bailey, Kirk, & Martin, 2000; Haslam, 1997; Lippa, 2000; Pillard, 1991)." - quote from "GD Studies". As seen, previous research studies the extent of masculine and feminine of the same sex/gender, but how about on opposite sex? For instance, we can compare the masculine or feminine behaviours and cognitions of lesbian women with gay men. What do you think?

Hey SunnyZENG,

I didn't consider comparing masculine vs. feminine preferences/behaviours/cognitions of just the homosexual community for our future research!

I should also note that earlier I wasn't sure if the homosexual community's responses skewed overall responses, but I realized from your response that the study DID mention averages of the homosexual community as a whole:

"In absolute terms, the mean GD scores for gay men (0.51) and for lesbians (0.55) indicated that both groups, on average, had neither strongly male- nor female-typical occupational pref- erences. That is, both group means were near the mid- point of possible GD scores (i.e., 0.50)."

So since the differences between the homosexual and heterosexual community are offset by the average scores as a whole, they're not skewed.

Ok, back to your question. I definitely think we can study and compare gay men and lesbian women alone using GD, since their preferences seem to parallel the heterosexual community. But since, as mentioned above, their responses don't actually skew results, so I'm wondering if the findings/results of the GD would essentially be the same, but applied to opposite sexes? What are your thoughts?

yes, this is what I am thinking. In further studies, researchers might examine GD of homosexualities based on opposite sexes because participants’ sexual orientations and gender identities role in society do really effect the reliability of GD in this research. For future research, participants can be equal sample size of Lesbians and gays, maybe the same method to test the masculine and feminine extents of gay men and lesbian women (or do you have a better way to examine?). Do you agree or any advise?

I agree, we know from research (some summarized in this paper) that gays, lesbians, and bisexuals are more likely to experience stigma, bullying, harassment, and as a result negative physical and mental health behaviours and outcomes. Future research could look at the extent to which this is due to gender nonconformity and a GD scale like ours would be perfect to do this.