What we found regarding self-efficacy March 28, 2011 Comments Off

Self-Efficacy

Self-efficacy is defined as a person’s believe in their ability to successfully accomplish a particular task. As I was engaged in my literature search I found many research papers that highlighted gender differences in self-efficacy related to computers. Most relevant to my area was a study that found the even after a full term business course that focused on teaching business students how to use Excel and other software programs, the females still had lower belief in their ability to successfully preform the more complex tasks using Excel (Busch 1995).

Self-efficacy is important in many different ways, but one of the most important components is that persons with low self-efficacy are likely to give up when faced with an obstacle, which in the Busch study suggests that females would be less likely to persist through more complex tasks using Excel when the going got tough. A person with high self-efficacy, on the other hand, is more likely to persist through a challenge and not attribute the failure to their own inability.  Self-efficacy is consider task related, a person may be highly efficacious when it comes to conflict resolution among feuding parties, but believe that they are unable to accomplish a computer task. (Of course if one person has never been trained in conflict resolution or to operate a computer then asking their self-efficacy regarding the task has little meaning.)

Study

When we designed our studies, therefore, we required that all participants be familiar with spreadsheets and had used them to calculate. That is, they must have used formulas before. This was required because a large part of the study’s task was to edit formulas in spreadsheets.  We set another limit that we did not want participants to be in an engineering field where they would have encountered many different programming environments, as many students do during their engineering education.  We were interested in studying and understanding end-user programmers, the business student, forestry student, etc., whom use spreadsheets as part of their daily tasks of accomplishing other goals for their respective fields.

We ran the study as follows (roughly): background questionnaire (including self-efficacy questions), tutorial of the spreadsheet environment, 2 tasks where they should find and fix errors in the spreadsheet formulas, questionnaire assessing understanding of spreadsheet environment and perceptions of the their performance.

Results

We found a number of interesting things:

* females had lower self-efficacy about the task than the males did as they entered the study

* females self-efficacy was predictive of how well they used the features

* males self-efficacy was not tied to any measure of performance

* females and males understood the features equally as well

* however, females did not use these features because they worried it would take them too long to learn them

Some points I want to dive into here.  First, the females: they spent more of their time focused on the feature they had used before coming to the study.  Since the spreadsheet environment was new to them (it’s a research prototype), we are certain that 2 of the 3 “feature areas” were totally new to them.  The only area they were familiar with was editing/changing formulas.  They did this significantly earlier, and more often, than the males.   Of the other “feature areas,” one we taught and used extensively during the tutorial and the other we told of its existence, but did not explain how it worked or other details.  Interestingly, despite their low usage of these features, their understanding of the features did not differ, statistically, from the males. These features are known from other studies to significantly help in this specific task (see references in paper linked at the bottom of this post).

There are a few different interpretations that might be made from this.  Perhaps females are less likely to start using new and unfamiliar features to them in a time of “stress” when they have a task to accomplish and they cannot be sure the new tools given to them will be helpful.  An alternative explanation is that the tools are simply not what the females would like to use, despite the evidence that they “work” to help others’ solve the task. I will dive into what avenue we explored in another post.

What about the males?  Self-efficacy theory suggests that self-efficacy, at least to some degree should predict performance. However, we found no measure of males’ performance that was predicted by their self-efficacy.  Although we didn’t explore this topic in the paper, I am curious.  Did the males lie about their self-efficacy on the task?  While of course this is possible, they were filling the questionnaire out on their own and their names were in no way associated with their work. Another explanation could be that social expectations suggest that males should be good at things on the computer.  OR, that they were over confident about their ability to solve problems with spreadsheets. (If others have thoughts about why self-efficacy was predictive for the females, but not for the males in the same manners, I’d love to hear your hypotheses!)

You can find the research paper that dives into the details here (published at CHI in 2005): http://hciresearcher.com/research/p869-beckwith.pdf

 

What is it? November 16, 2010 Comments Off

There are numerous places (i.e., Gender HCI’s home page, Wikipedia, etc.) to get an overview of Gender HCI, a field of HCI my former advisor, Margaret Burnett, and I started using with regard to my PhD work.  HCI (Human Computer Interaction), roughly, is the study of how best to design technology taking into consideration the humans who will be interacting with it.  Gender HCI is understanding how males and females interact with technology, ensuring that a particular design supports both equally. 

History

Studying gender differences with respect to technology is not new. As the Wikipedia entry on Gender HCI details, researchers in the 1980s found differences in how males and females interacted with computers.  The research on gender differences with respect to technology use took off in the 1990’s.  By 2003, when I started, the majority of this research was focused in two primary areas: the dismal numbers of women majoring in computer science and gender differences in attitudes towards (and sometimes the subsequent use of) computers. 

Why is understanding gender differences important?

In recent years women have started to outnumber men at colleges and universities.  Women also comprise nearly 50% of the workforce.  Given the importance of computers in today’s workplace, inefficient computer use by nearly half the workplace population could have a very significant effect on the bottom line.  This is one potential consequence of being unaware of gender differences in computer use. 

Furthermore, these gender differences do not appear to be disappearing with those who have been exposed to computers through their whole lives.  Studies of children have found that not only do boys use computers for more time, but teachers allow girls to “give-up” when solving computer-based problems earlier than the boys.  These findings, along with others, paint a disappointing picture for the hope that the gender differences we see in today’s adults will change in a generation. 

For these reasons, and more, I believe we must take gender differences into account as we design software.  It is yet another tool to help make all computer users more effective in their daily lives.

What it means for industry

From an industry perspective, designing software which accounts for gender differences has potential monetary benefits as well.  As females take more and more leadership positions and are in roles where they decide which software to purchase, their own experience with software which fits their problem-solving styles is likely to have a large influence on their decision.  Furthermore, in areas outside technology, companies who consider the needs of their female audience have designed tools, for example, which work better for both genders.  From this same perspective, companies designing software and accounting for gender difference may just improve the experience for all that use the software.

There are a great variety of topics related to Gender HCI which range from gender differences in areas unrelated to computer science, to HCI as a general topic, and onto the conditions under which software is developed. Each of these together comprise important factors into the field of Gender HCI, and are areas which I will cover in this blog.

About

Topics related to Gender HCI: covering areas from general gender differences to women in computer science, interaction design, end-user programming and beyond.