The aim of this study is to further support this theory and investigate it using hyperthical sistuations to see if it only occurs in bussiness or if there is a general inclination to appoint women to risky leadership positions and men to stable ones. This will be investigated by presenting participants with a job position for a festival with either declining or increasing success and a applicant for the job who is either male or female, making these the independent variables. The dependant variables wil be how suitable the participants see the applicant as for the job and if they believe it to be a good opportunity for them. Based on the research gathered above, the following hypotheses were derrived:
- A woman will be seen as more suitable for a risky leadership position than a man.
- A man will be seen as more suitable for a non-risky leadership position than a woman.
Method
Participants
The participants were first year psychology students in Exeter University and were selected by their attendance to a practical lesson and the conditions were randomly allocated. The participants were predominantly female and most were between the ages of 18 and 20.
Design
The independent variables were gender and how successful the festival was (in order to indicate how risky the job would be). These were manipulated by information given to the participant on the first few pages of the questionnaire. The dependant variable was how suitable the participant felt the applicant was for the job, which was tested by a relevant question on the questionnaire.
Apparatus
The participants were given a pack containing a CV, an advert for a job helping run a festival, a newspaper article and a questionnaire (see appendix). The CVs differed only in gender for the applicant and the newspaper article in one condition told of how successful the festival in the job advertisement was and in the other how unsuccessful it was. The questionnaire contained a total of eleven questions, two of which were relevant to the results being collected (asking whether the applicant is suitable for the position, whether it is a good opportunity for them with a manipulation check of whether the participant would recommend it to a friend) and one of which was to ensure the independent variable of the festival’s success had been manipulated correctly (asking whether the festival was doing well). The other questions were randomly chosen to prevent demand characteristics caused by the participants guessing the aim of the study. They were required to answer each question on a scale from 1 (do not agree at all) to 7 (agree completely) to produce analysable, quantitative data for easier comparisons between conditions.
Procedure
The study took place in a practical room where outside noise would be kept to a minimum to prevent this being an extraneous variable. The participants were asked to complete the questionnaire in their own time after reading the information and then the results were collected and analysed.
Results
The results from each of the three questions were collected separately. The first question asked whether the participant felt the applicant was suitable for the job, and the results collected are shown in the table and the graph below.
Table to Show Mean Answers to Whether The Applicant is Suitable For the Job
Table 1
Graph 1
At first glance it is apparent that hypothesis 2 (that men will be seen as more suitable for non-risky leadership positions) is not supported as the trend is in the opposite direction to that predicted as the average score for the female applicant in the successful (and therefore non-risky) festival condition was higher than the average score for the male applicant. However, hypothesis 1 (that women will be seen as more suitable for risky leadership positions) does seem to be supported as the mean score for the suitability of the female applicant is quite a bit higher than that of the male applicant.
For the sake of this experiment, a statistical test will not be performed on this data but a difference of 1.00 between the means will be deemed as a statistically significant result. Therefore, despite the correct direction being evident in hypothesis1, it is not significant enough to suggest the difference wouldn’t have occurred by chance.
The second question of relevance asked whether the participant felt the job was a good opportunity for the applicant. This question was intended to measure the roughly same thing as the first question, as both are intended to show how gender affects the decisions made on jobs.
Table to Show Mean Answers to Whether the Job is a Good Opportunity
Table 2
Graph 2
The results collected from the second dependant variable, as shown in the graph (2) and table (2) above, seem to be consistent with hypothesis 2, as the average score is higher for the male applicant than female applicant in the condition where the festival’s success was improving. However, in the declining condition, the results contradict hypothesis 1, as the mean score for males is higher in this condition also. Once again, no statistical test will be conducted and there is not a difference of 1.00 or higher between the conditions so these results will not be deemed as statistically significant also.
The third set of results collected was from the question asking whether the participant would recommend the job to a friend. This was intended to work as a manipulation check, as it would show whether the participant saw the declining success of the festival as indication that the job is risky. The results are shown in table 3 and in graph 3.
Table to Show Mean Answers to Whether the Participants Would Recommend the Job to a Friend
Table 3
Graph 3
Graph 4
As predicted there is minimal difference between the conditions of male and female, as the gender of the applicant should not have an affect on how the participant views the job itself. As gender is not of interest in this case, a graph has been constructed combining these two conditions to view the difference between improving and declining festival success easier (Graph 4). The improving success condition is much higher, as predicted. These results do prove statistically significant (improving condition mean = 10.76, declining condition mean = 9.56) as there is a difference of more than 1.00.
Discussion
The aim of this study was to investigate the theory of the ’glass cliff’ by seeing if participants were more likely to appoint a female candidate to a riskier job or more likely to appoint a male to a stable one. Neither of the results collected from the dependant variables appear to be statistically significant, that is neither had a difference of 1.00 between any of the conditions. This means the experimental hypotheses have been rejected and the null accepted,
A worrying feature of the results, however, is that the trends in the first and second question seem to be in opposite directions (the mean score for the female applicants in both the improving and declining conditions outweigh the mean score for the male in the first question, and vice versa for the second) despite the intention that they measure roughly the same thing. This would suggest that they do not, or that the participants have misunderstood what the question is asking, making the results potentially invalid.
If this is not the case, the implications of this research could be that the ‘glass cliff’ is either not as prevalent outside the companies tested in Haslam and Ryan’s study or that it has limitations, as these results do not support those collected in that investigation.
One issue could have been that the ‘glass cliff’ theory only happens with leadership positions, and the job outlined in the handout could have been seen as less central to the festivals overall success and therefore less risky for the applicant alone. This could be tested by conducting the same experiment but including another condition with the job advertisement referring to a managing director, for example. However, the results from question to check the manipulation of the festival’s success did appear to be statistically significant, which suggests that inadequate manipulation of this variable is not to blame.
The number of female participants far outweighed the number of male ones and this may have affected the result. Women may not be so inclined to hire depending on gender for example. This could be tested by dividing the participants into two conditions depending on their gender, as well as the applicant’s gender. This could correspond with the ‘glass cliff’ as the ‘glass ceiling’ theory dictates that that women are underrepresented in leadership positions in general and so it could be the fact that males are making the decisions to appoint women to risky jobs as they are more likely to be in a position of authority enabling them to do this.
In conclusion, these results do not support the previous research into this area and so could indicate limitations or exceptions such as certain types of jobs that are unaccounted for in this research. However, there are a number of methodological flaws that could have accounted for this discrepancy such as the uneven number of males and females in the sample. Taking both into consideration, it does seem likely that some of the factors of the ‘glass cliff’ theory may not be generalisable to situations outside a business environment and may be restricted to specific types of occupations.
References
Ryan, M.K., & Haslam S. A. (2005). The Glass Cliff: Evidence that women are over-represented in precarious leadership positions. British Journal of Management, 16, 81-90
Ryan M. K & Haslam S.A. (in press). The Glass Cliff: Theories that explain and sustain the precariousness of women’s leadership positions. In B. Schyns & J. R. Meindl (Eds), Implicit leadership theories: Essays and explorations. Greenwich, CT: Information Age Publishing.
Atwater, L.E. & Van Fleet, D. D. (1997). Another ceiling? Can males compete for traditionally female jobs?, Journal of Management. Retrieved November 29th 2005, from http://www.findarticles.com/p/articles/mi_m4256/is_n5_v23/ai_20231484.
