Microsoft Word doctoral thesis final[1]. doc

 
 
 
 
 

---

 
118
4.4.2.5 Data collection 
 
The process of data collection began in February 2006 and was completed in the first 
week of July of the same year. To reiterate, data were collected by means of two 
methods. One was a pre-coded questionnaire, and the other was the VConf-FGI.  
 
The questionnaire was used for quantitative data collection which included among 
others, personal demographic information, career profile/prior experience; skills, tasks, 
roles and responsibilities; job challenges and strategies for dealing with them, 
perceptions and role expectations. Questionnaires were posted overland to the 
participants towards the end of March 2006 and responses were received during April 
and July 2006. The response rate was 77 per cent. 
 
4.4.2.6  Data analysis  
 
In this study, the SAS/STAT statistical package, version 9.1, was used to analyse the data 
captured from the questionnaire responses. Twenty three questionnaires were returned 
and their responses captured. Data analysis and presentation procedures for the 
quantitative data were employed. Each questionnaire item was regarded as representing a 
biographical characteristic or perception of the respondent on some managerial issue, and 
thus a separate variable. Analyses undertaken on the responses to these variables included 
the following: 
 
•   one- way frequencies: exploratory frequency tables on each and every  
       questionnaire item. 
•  combined frequency tables: combined frequency tables for each section (D-L) 
ranked according to  sum totals, calculated for each item within an aspect;  
 the sum totals were calculated as the sums of frequencies  in  the  two  most                      
favoured adjacent categories in each section 
•  summary tables of sub-item means: standard deviations minimum and maximum 
values for each of the issues covered within sections D-L of the various 
managerial issues 

---

 
119
•  relationships: cross-tabulations of the sub-sets of items within sections D-L of the 
questionnaire with biographical variables of age, years experience as HoD and 
years attached to their particular institution. 
 
The purpose of  each strategy is discussed below. 
 
The analysis was conducted using one-way frequency tables for each question on the 
questionnaire.  As an initial exploratory analysis step, univariate one-way frequency 
tables, categorised  according to the specific options of each question, such as, ‘strongly 
agree’ to ‘strongly disagree’ ; or ‘single’, married’, ‘divorced’,  were initially calculated 
for each variable. This step was undertaken to validate data and correct or remove any 
spurious responses. For example, if a response of ‘7’ should be encountered for any of 
the 5 point rating scale questionnaire items in sections D-L, the particular response for 
the participant can be further investigated, traced back to the participants’ questionnaire 
and checked. A value of ‘7’ falls outside the range of valid responses - which vary 
between ‘1’ and ‘5’ for the five point rating scale questionnaire items.  
 
The one-way frequency tables on the biographical variables, questions 1-33, were 
furthermore calculated as a means of describing the sample population of women in 
managerial positions.  This was done in two stages, first on the entire sample population, 
then frequencies were split into SA and UK participants separately. 
 
Combined frequency tables for sub-sets of items for each of the sections, D-L, of the 
questionnaire were utilised. In each of the sections, D to L, participants’ ratings on 
several sub-issues were required. Each of these sub-issues represented a separate 
variable. The responses to the sub-issues in each section were integrated into a combined 
frequency table for each section/managerial aspect. By studying the response distribution 
within the various tables, a general impression of participants’ perceptions on the various 
managerial issues can be formed. 
 
Summary tables of mean, standard deviations, means and maximum values for each of 
the sub-issues covered within each of the managerial issues were calculated as well as for 

---

 
120
each of the sub-issues covered within each of the managerial issues of sections D-L.  By 
studying the mean response values within a section, sub-issues that deviate from the 
others within the section/aspect can easily be identified as a mean value substantially 
larger/smaller than the other mean values.  These identified mean values indicate that 
participants perceived/rated the particular items/sub-issues differently to the others.  
 
Combined-items frequency tables for each of sections D-L of the questionnaire were 
ranked. Participants had to decide which sub-items/sub-issues within each managerial 
issue  they regarded as more important than others. In an attempt to rank the importance 
of the various aspects, the frequencies of the most favoured adjacent categories within a 
section, usually the positively rated categories, such as, ‘important’ and ‘very important’/ 
or ‘agree’ and ‘strongly agree’ categories were summed for each sub-item within a 
section/managerial issue. It was argued that by combining adjacent ‘positively’ rated 
‘agree’, ‘strongly agree’  or ‘negatively’ rated  disagree/strongly disagree’ categories, 
sub-items within a section could be ranked according to levels, of ‘like’ if positively rated 
or ‘dislike’ if majority of the items were negatively rated. The sum for each sub-item 
within a managerial aspect was ranked in descending order and presented as a separate 
table; such as, 
 
• 
relationships: cross-tabulations of sub-items within sections D-L of the 
questionnaire with biographical variables of age, years experience as HoD and 
years attached to their particular institution 
• 
frequency tables of cross-tabulations present how the biographical variables of 
age, years experience as HOD and years experience at current institutions 
affect/interact with the positively perceived part of each section’s rating scale; the 
cross-tabulations were done in an attempt to determine whether  these selected 
biographical variables influence participants’ perceptions of the various 
managerial aspects 
• 
chi-square tests were calculated from the table of frequencies to determine if there 
was any significance. 
 
 

---

 
121
4.4.2.7.  Reliability and validity 
 
According  to  McMillan  and  Schumacher  (1993:227),  reliability  refers  to  the               
consistency of measurement, which is, the extent to which the results are similar over 
different forms of the same instrument or occasions of data collecting. A highly reliable 
instrument (Cates 1985: 124) can be depended upon to produce the same, or nearly same, 
score when administered twice to the same subject or when administered to two subjects 
of equivalent talent and experience. To ensure consistency of measurement in the current 
investigation, the researcher administered the instruments to subjects of similar 
educational and professional backgrounds. 
 
Validity is the degree to which scientific explanations of phenomena match the realities 
of the world (McMillan & Schumacher 1993:601). A measurement instrument 
(Cates1985:123) is valid if it measures or represents what it claims to measure or 
represent. In the case of this investigation, the survey instrument was used to assess the 
lived experiences of the participants as women HoDs in universities. Thus the items on 
the questionnaire were all related to this phenomenon of interest.  Questionnaires were 
posted overland to all participants. 
 
In quantitative research, reliability refers to the consistency of the instrument and test 
administration in the study (McMillan & Schumacher1993:385). To enhance reliability 
the survey instrument was administered during the same time period to all participants.