Optical Answer Sheet - Sample Essay

Due to the large number of sample, e-mail is used to distribute the surveys (self-administered questionnaires). Responses are almost instantaneous which shortens time. This is done for higher level of staff such as executive and management levels as they usually own a computer. The lower levels are given OAS (Optical Answer Sheet) which can later be analysed using an optical scanning system (refer to point 7 below) which helps saves time and reduces human error. The cost is low compared to mail surveys. However, wording problems may occur. E.g. Meaning of a question may differ from person to person. It can be difficult to gather a 100% response rate within a stipulated time-frame as the respondents may not response. (Zikmund 2000)

Interviews are used to collect data on why staff feels that they are unhappy which fulfils objective c as above. The questions are structured and closed. E.g. Does poor product design affect your enthusiasm to sell it? Unlike the questionnaire, interviews can help researcher to ask further in-depth on why are they unhappy. It allows probing which seeks a more defined reason. It also allows asking of complex questions which cannot be asked in questionnaires. E.g. do you think your attitude towards work will affect your sales performance? (Zikmund 2000)However, it is time-consuming to conduct the interviews. (Brewerton and Millward 2001)

Personal interviews are not used as it is very time-consuming. Even if given a small sample of 1% (refer to point 6 below), it is too expensive to conduct 350 personal interviews. Therefore, telephone interview is used. It is flexible, and it is easy to conduct interviews in terms of geographic regions. Interviewers do not have to move from place to place to conduct an interview. Stratified sampling is used for questionnaire as the employees are stratified into their job types – sales, manufacturing, marketing, accounts and others. It identifies subgroups in the population and allows every member of the organisation in each department is included. (Collis and Hussey 2003)

The population is defined as the 35000 members in the organisation. The desired sample size would be 30% of the sample or 10,500 staff. They are then sub-divided into subgroups such as sales as mentioned earlier. Disproportionate sampling is used where 50% of the sample is taken from sales. A higher percentage is given as sales. Sales department is one of the contributors towards sales variable. 20% is taken from accounts as they know which factors contribute to costs. 15% each is taken from marketing and others forming the entire sample of 100%. Lastly, using simple random method, the names from each subgroup are randomly selected. (Gay and Diehl 1992)

The process is repeated for interviews, except a smaller sample of 1% is taken due to time- constraints. This sample will be a representation of the whole company. It is chosen at random, unbiased and is large enough to satisfy the needs of the research. (Collis and Hussey 2003) Analysing Data Once the data is collected, the data will be edited. The data will be edited for consistency. E.g. the respondent must have at least worked in the organisation for more than a year. It so happens that the respondent worked for less than a year showing the data is inconsistent. It is then edited for completeness such that all the questions are being answered. (Zikmund 2000)


After coding the questionnaire, the data is processed using a computer. To record the data, highly structured paper questionnaires (OAS), given out earlier, is scanned directly into the computer using an optical scanning system. Email questionnaires data is then entered into the computer manually. However, errors may occur when keying in data manually as the person may misread or key in the data wrongly. (Cooper and Schindler 2001)

When all the data is entered, SPSS (Statistical Package for the Social Sciences) is used to analyse the data. Using the data entered, it can calculate frequencies. Cross-Tabulation will be used to analyse to compare the three variables, mainly sales, costs and attitudes of the staff. SPSS can also calculate the co-relation of the variables. Hypothesis testing is also used to evaluate if there is a relationship between the variables – sales, costs and attitudes, if they are inter-related.

H0 states that sales, costs and attitudes are inter-related. H1 states that sales, costs and attitudes are not inter-related. From the SPSS results, the critical test value, t (parametric test), X2 (chi-square), is obtained. Using the one- and two- tailed tests at 5% level of significance; the hypothesis is then rejected or accepted. (Cooper and Schindler 2001) However, its disadvantage is it does not state why they are inter-related. Interviews and secondary data

Content Analysis is used to analyse interviews. From the sample, 3, “fat and good”, detailed interviews are selected to develop codes such as sales, poor working environment etc. After the codes are modified by another researcher to achieve consistency, the 350 interviews are coded. It then converts text into quantitative data which later uses SPSS to further analyse. However it is very time-consuming, as it needs the researcher to go through the 350 interviews before 3 good interviews are selected. It is also used to analyse secondary data such as annual reports, CEO speech. Codes such as expenses, costs are developed and a relational analysis is used to count frequencies.

There are several ethical issues which need to be considered. Participants must be ensured confidentiality and anonymity throughout the research. E.g. it must be ensured that confidentiality is given with regard to demographic issues. Participant’s anonymous status will be affected if someone were to cross-reference demographic details with other variables such as sales figures as they will be individually identified. (Brewton and Millward 2001)

Hence, there must be informed consent when asking them to do questionnaires or interviews. In terms of avoidance of harm, participants must not be positioned in harm and risk. E.g. Participants may be at risk of losing their job if researcher divulges that they take home large amounts of stationery which increases cost. This also enforces confidentiality and anonymity. (De Laine 2000)Also with regards to intervention and advocacy, the researcher experiences ethical dilemma whether or not to inform the organisation of such an incident. Researchers must maintain honesty and trust as it is likely to manipulate the data in order to achieve a specific result or researchers. (De George 1999)

The research must be done with integrity and quality that it reaches a certain standard. E.g. A researcher can manipulate past research to make a “new” research. Lastly, the ownership of the data belongs to the organisation. It is unethical to keep the data and sell off to competing companies. (Miles and Huberman 1994)


1. Berg, B, 1998, Qualitative Research Methods for the Social Sciences, 3rd edition, Allyn & Bacon, U.S.A.

2. Brewerton, P and Millward, L, 2001, Organisational Research Methods: A Guide for Students and Researchers, Sage, London

3. Blaikie, N, 1993, Approaches to Social Enquiry, Polity Press, Cambridge