Leeds University Business School LUBS 5911 Data Collection – II Dr. Surender Munjal PhD, CA, CMA, M.Phil, M.Com, PGCERT, Fellow of the HEA Associate Professor of International Business and Strategy Centre for International Business, University of Leeds (CIBUL) Survey Data Leeds University Business School Outline for week 5-6 Recap Survey Data Basics Principles of Data Collection through Questionnaire Ethical Considerations Goodness of Questionnaire Data – Reliability – Validity Reporting and Referencing Data Data Analysis Leeds University Business School Recap Leeds University Business School Recap Leeds University Business School Survey Data Primary data is often collected using questionnaire. Other dominant forms of collecting primary data includes: – Interviewing – Observation – Focus Group However, most of other approaches are qualitative. Questionnaire is a pre-formulated, written set of questions of which the respondent records his answers. Leeds University Business School Pros and cons of survey data collection through questionnaire Reasonably straightforward data collection Potential for bias in responses (are people responding honestly ) Standardised data Low response rates create nonresponse bias Efficient for collecting a lot of data Ambiguity in questions can be difficult to detect (and deal with) ex post Cost-efficient (especially with on-line surveys) Difficult to know if respondents have taken the questionnaire seriously Allows respondents to maintain anonymity Note: anonymity ≠ confidentiality Leeds University Business School Principles of Survey Data Collection through questionnaire General Principles – Administering questionnaire – Appearance of questionnaire – Length of questionnaire – Introduction and instructions to respondents Leeds University Business School Principles of Survey Data Collection through questionnaire Principles of Wording – Content and purpose of questions – Wording and language – Type and forms of questions – Sequencing – Classification data or Personal information – Respondent validation Principles of Measurement – Categorization – Coding – Scale and scaling – Reliability and validity Leeds University Business School Administering Questionnaire Personal administration Mailing – Post – Email Online – Qualtrics – Survey Monkey Note: LUBS students are entitled to a Qualtrics account. You can self-enrol for your Qualtrics account here (remember to use your University of Leeds email address): – https://leedsubs.eu.qualtrics.com/login Leeds University Business School General Principles Classification Data or Personal information – Age, Gender, Education, Income, Profession, Marital Status – Use options, e.g. age/income groups, rather than asking specific detail Used for controlling heterogeneity and representativeness of the sample – necessary items only – avoid names for anonymity – contact details, if further contact is necessary such as second round – tracking, via IP address Leeds University Business School General Principles Questionnaire should look attractive and neat Use introduction to provide details about the motivation and reason of the survey. Provide contact details of the surveyor Refer to the ethical approval obtained Ensure confidentiality and non-commercial usage of the information collected Give instructions and guidance necessary Finally thanks in advance Leeds University Business School Ethical considerations Ethical concerns may emerge at all stages of research. The main issues to consider are: – The rights of privacy of individuals – Voluntary nature of participation – and the rights of individuals to withdraw partially or completely from the process – Consent and possible deception of participants – Maintenance of the confidentiality of data provided by individuals or identifiable participants and their anonymity – Reactions of participants to the ways in which researchers seek to collect data – Effects on participants of the way in which data is analyzed and reported – Behaviour and objectivity of the researcher Source: Saunders, Lewis and Thornhill (2003, p. 131) Leeds University Business School Question Wording Use simple language and words Avoid double negatives Avoid double-barrelled questions Avoid ambiguity Avoid questions leading to biased responses Socially desirable language/questions Avoid recall dependent questions Avoid long questions – ~ 20 words – one full line in print Leeds University Business School Question Sequence Easy Difficult General Difficult Leeds University Business School Principles of Measurement Data should be collected with appropriate scales so that it is useful for testing hypotheses. – Likert Scale: 3 point / 5 point / 7 point (strongly agree, agree, do not know, disagree, strongly disagree) Donot use event points. – Ratio Scales: numeric scale with zero (indicating nil value) as origin – Interval scales: numeric scale that classifies and orders a measurement, e.g. temperature, age group etc. Zero doesn’t mean nil value. Goodness of data is assessed through tests of validity and reliability. Leeds University Business School Survey construct: example Construct is a conceptual term describing a phenomenon of interest FOR EXAMPLE: Exporting risk perception: is the degree to which a manager is positively predisposed to approach foreign markets There can be multiple ways to measure one construct; the goal is to find the best one among them. Acedo, F. J. & Jones, M. V. 2007. Speed of internationalization and entrepreneurial cognition: Insights and a comparison between international new ventures, exporters and domestic firms. Journal of World Business, 42(3), 236-252. The item of a single scale is measured by using 1 to 7 Likert scale Item 1: Selling products in foreign markets implies high risk. Item 2: Exports are an important opportunity for my firm. Item 3: International activity is a positive thing in my business. Item 4: My firm has a high probability of success in foreign markets Leeds University Business School Validity Validity establishes how well an indicator seems to be a reasonable measure of its underlying construct. – Convergent validity (CV) refers to the closeness with which an item relates to (or converges on) the construct that it is supposed to measure. Correlation should be 0.6 or higher. – Discriminant validity (DV) refers to the degree to which a measure does not measure (or discriminates from) other constructs that it is not supposed to measure. Correlation should be 0.3 or lower. Bivariate correlation analysis between items can be used to test CV and DV Leeds University Business School Reliability Reliability indicates consistency (not accuracy!) between different items of the same construct – Cronbach’s alpha, a reliability measure, calculated using the following formula: where K is the number of items in the measure, is the variance (square of standard deviation) of the observed total scores, and is the observed variance for item i. α should be higher than 0.7 Leeds University Business School Seminar Activity Prepare a short survey to collect firm level data on entry modes and location choices made by firms planning to enter into the Europe: – You will need an introductory statement – A set of questions to control for firm level heterogeneity – A set of questions on location choices – A set of questions on entry modes Use Qualtrics account to design the survey.
