How masculine are you? This might seem a fairly simple question, especially if you’re asked to fill out a simple ten-question survey investigating traits such as “aggressive” and “forceful”.
But it may not be as straightforward as you think: your self-reported masculinity levels are not only a product of those personality traits, but also the method by which you’re surveyed.
Along with colleagues we recently published a paper in the journal Personality and Individual Differences which shows that people describe themselves as less masculine when completing an online questionnaire than they do when filling out a pen-and-paper version.
Online vs offline tests
Increased widespread access to the internet has led researchers to use the internet to collect data. It makes sense - benefits of conducting research online include:
- greater numbers of participants from different locations
- reduced costs (no printing involved)
- fewer missing responses (as internet data collection programs can ensure individuals answer all questions)
- fewer errors when entering data.
But it appears that when filling out online questionnaires people are less likely to tailor their responses to what is considered socially desirable than they are when writing with pen and paper.
This can be attributed to people feeling more anonymous online.
Researchers are also aware of potential problems when conducting research online. Some of these problems include higher non-response rates and lack of experimental control. For example when someone is filling out a questionnaire online there is no way of controlling their environment. A respondent might be distracted by television or their kids and this could affect the result.
There is also concern that the structure of online surveys may not be equivalent with their offline counterpart. Specifically, does administering a questionnaire online affect the way it performs as a psychological survey?
Research suggests it is important to not assume equivalence between online questionnaires and their offline counterparts. As these measures are being completed in different environments, this may produce different responses, so research should seek to establish equivalence with all online and offline measures before interpreting results.
Testing gender roles
We sought to establish the online and offline equivalence of the Bem Sex Role Inventory (BSRI), a self-report test that measures or shows levels of masculinity and femininity. Although the Bem Sex Role Inventory is almost 40 years old, researchers consider it to be a good measure of gender roles.
Gender roles are considered to be socially defined appropriate behaviour for men and women. As contemporary research continues to use the Bem Sex Role Inventory to assess how people rate their gender characteristics, it is important to establish the equivalence of this measure online and offline.
We also decided to use the Bem Sex Role Inventory short-form instead of the full version, as the short-form has been found to be a more accurate measure of gender roles. Due to better accuracy, the short-form is more likely to be used in gender roles research.
We recruited 244 participants to complete the Bem Sex Role Inventory short-form online, and 128 participants to complete the Bem Sex Role Inventory short-form offline by pen and paper. The short-form contains ten items classified as masculine traits (such as “independent” and “assertive”) and ten items classified as feminine traits (such as “sensitive” and “understanding”).
We were interested in whether participants’ masculinity and femininity scores would be the same online and offline. As online and offline equivalence has been established for a range of other surveys, we anticipated the scores would be equivalent.
We found no difference between online and offline femininity scores. But interestingly, masculinity scores were higher offline compared to online. These results were consistent for both men and women.
Anonymity and freedom
Why do people score higher in masculinity offline compared to online? We wondered if when people completed the measure offline, they may have been responding in a more socially desirable way.
Previous research has found that online responding increases feelings of anonymity, and people are more likely to be candid with their responses.
If people completing the measure offline were answering in a socially desirable manner, this means they were over-representing traits they believe to be socially desirable. This is interesting, as our findings suggest that masculine traits such as “defends own beliefs”, “independent”, and “assertive” are considered socially desirable in both sexes.
It is also possible that conceptualisations of masculinity may differ online and offline. If this is true, then this is an important result in itself, as the concept of masculinity changes in different environments. This result raises the question of whether people behave differently in online environments compared to offline environments.
If people do behave differently, this has potential consequences for researchers who do not establish equivalence of online and offline questionnaires. It cannot just be assumed that online scores will reflect offline scores.
Results of our study suggest that further research should explore how people’s behaviour changes in different environments - specifically, how does the filter of online anonymity change our virtual character?