Doing Quantitative Research 26E02900, 6 ECTS Cr.

Doing Quantitative Research 26E02900, 6 ECTS Cr. Olli-Pekka Kauppila Daria Kautto Lecture 1 April 18, 2017

Course Objectives This is a skills course that improves your ability to carry out a quantitative research project in management studies Different to statistics courses, our approach is very hands-on; to learn more about statistics, please take e.g. 30C00600 Continuation course in statistics Our primarily focus is on the use of survey methods After this course you should be prepared to design, execute, and report a research project using basic quantitative methods able to understand and evaluate more advanced quantitative methods that are most commonly used in management studies

Course Readings Compulsory pre-readings Ertug, G., Cuypers, I. R. P., Noorderhaven, N. G., Bensaou, B. M. 2013. Trust between international joint venture partners: Effects of home countries. Journal of International Business Studies, 44: 263 282. Renko, M. 2013. Early challenges of nascent social entrepreneurs. Entrepreneurship Theory and Practice, 37(5): 1045 1069. Schilke, O. 2014. On the contingent value of dynamic capabilities for competitive advantage: The nonlinear moderating effect of environmental dynamism. Strategic Management Journal, 35: 179 203. Sliter, M., Kale, A., & Yuan, Z. 2014. Is humor the best medicine? The buffering effect of coping humor on traumatic stressors in firefighters. Journal of Organizational Behavior, 35: 257 272. Recommended readings Field, A. (2009) Discovering Statistics Using SPSS. Sage Publications Ltd: London, UK. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2014 or any earlier edition) Multivariate data analysis. Pearson Education Limited: Essex, UK. Various additional readings for each topic (see, Syllabus)

Lectures (each day 9:15-12:00 + additional times as marked) Date Topic 18.4.2017 Introduction to the course. Research design in management studies. 19.4.2017 Measurement scales: reliability and validity assessment. Exploratory factor analysis. 20.4.2017 Descriptive statistics, variable transformations Returning and presenting Assignment 1* 21.4.2017 Regression analysis 13:15-16:00 Exercise Session: Start working on your research project 24.4.2017 Regression analysis cont d: indirect effects (moderation and mediation) 25.4.2017 Logistic regression: Analysis of categorical data Returning and presenting Assignment 2* 26.4.2017 Structural Equation Modeling - Theory 13:15-16:00 Structural Equation Modeling Practice Assignment 3* as classroom exercise 27.4.2017 Exercise Session: Continue working on your research project 28.4.2017 9:15-16:00 Course Project Day Returning and presenting Assignment 4*

Office Hours and Communication During the Course Office Hours on Tuesdays and Fridays after the class If you have any questions, need for feedback, or you wish to discuss something else, please come to see us during the office hours (do your best to avoid the use of email) Go to the course page on MyCourses for Lecture slides Assignment instructions Data files Other material Any news or updates

Assessment Assignments 70%: Assignment 1 (10 %): article-based measurement scales assignment; Assignment 2 (10 %): article-based analysis methods assignment; Assignment 3 (15%): SEM exercise; Assignment 4 (35 %): case-based coursework project. Final course presentations: 10%. Active participation and knowledge of SPSS analytics software application: 20%.

How we work? During the course, we will use different datasets that you can access from MyCourses We use SPSS software; please make sure that it opens on your computer We will use two screens; one to show the lecture slides and the other to demonstrate the use of SPSS If you can t keep up, experience difficulties, or have questions, please raise your hand and one of us will come to help you Please be active in completing and presenting the assignments Most of the time, you ll be working on your own but some assignments involve working in pairs/group

Today s lecture Research design in management studies. What happens before we collect the data?

Learning objectives for today s lecture 1. Gain a deeper understanding of what research question is and how is guides the research process 2. Improve skills to develop testable hypotheses to address the research question(s) 3. Develop ability to evaluate different data sources and choose an appropriate sample to test your hypotheses 4. Deepen knowledge of different forms of measures in management studies 5. Understand various aspects of different data collection methods 6. Understand and appreciate the ethics of quantitative research

Topics of today s lecture 1. Gain a deeper understanding of what research question is and how is guides the research process Theory part - i.e. what is it that you are set out to investigate 2. Improve skills to develop testable hypotheses to address the research question(s) 3. Develop ability to evaluate different data sources and choose an appropriate sample for different research projects 4. Deepen knowledge of different forms of measures in Empirical part - i.e. how would you test your predictions empirically management studies 5. Understand various aspects of different data collection methods 6. Understand and appreciate the ethics of quantitative research Ethical part - i.e. what are the rules and standards when studying the topic

Part 1: Research question and hypotheses

Research process Research problem / question What might be the answer? Hypotheses & theoretical model Where and how to collect data? How and when to measure the variables? Evaluation the quality of data and measures Testing the hypotheses Reporting the findings Answering the research question

Research process Research problem / question What might be the answer? Hypotheses & theoretical model Where and how to collect data? How and when to measure the variables? Evaluation the quality of data and measures Testing the hypotheses Reporting the findings Answering the research question

Research question Explicates the purpose of your research However, it is not always stated explicitly Good research question: Addresses an important topic - i.e. is interesting Current knowledge does not have an answer to it, OR Current knowledge provides an incorrect or insufficient answer Answering the question is likely to change the way we think about the topic Clear, focused, empirically addressable PhD Students: Note that you need to provide a theoretical contribution! i.e. it is not sufficient to demonstrate how a pre-existing theory works in an empirical setting

Examples of research questions What factors determine the likelihood of an organization's settlement on a new, contested institutional practice? Does the team goal orientation relate to team creativity, and if so, then how? Given that individuals have an overwhelming desire to gain status and they receive significant advantages when they have it, how do they react when they lose it?

Classroom exercise I Based on your reading of the article introduction, please answer the following three questions: 1. What is the research question in this article? 2. How does the author justify the importance of addressing this particular question? 3. To what theories the author claim that his findings contribute and how?

Forming the hypotheses The primary purpose of a theory section is to ground the hypotheses The role of the hypotheses is twofold: 1. Identify and organize different issues that we need to consider in order to answer the research question 2. Provide predictions of how different elements in the model are linked together The argumentation preceding each hypothesis Positions your hypothesis in related research Leverages logic and/or existing theoretical understanding to predict the relationship between (or among) specific variables

Pitfalls to avoid Stating the obvious Lack of coherence Fragmented theorizing Stretching theories too far Empirically untestable hypotheses

Grounding hypotheses to address the research question Research question: Given that individuals have an overwhelming desire to gain status and they receive significant advantages when they have it, how do they react when they lose it? Guidance from theory (hmm how could it be?) Argumentation Hypothesis 1. The negative effect of status loss on performance quality for high-status individuals is stronger than it is for low-status individuals. Hypothesis 2. Self-affirmation moderates the effect of initial status position on performance quality after status loss: high- (but not low-) status individuals perform better when they are given an opportunity for selfaffirmation.

Part 2: Data collection, data sources, and measures

Research process Research problem / question What might be the answer? Hypotheses & theoretical model Where and how to collect data? How and when to measure the variables? Evaluation the quality of data and measures Testing the hypotheses Reporting the findings Answering the research question

Objective data sources Independent of human perception The word objective does not necessary mean that the measure is valid or reliable Examples of objective measures used in management studies Number and types of the firm s alliance partners Firm performance data from the annual report Blood pressure, cortisol activity, cognitive ability Employee performance in terms of sales volume etc. Age, organizational tenure, functional affiliation, salary

Subjective data collection sources Based on human perceptions (i.e. things that we ask individuals to assess) Self-reported What are the issues that individuals can most reliably report themselves? E.g. Emotions and preferences that others are not aware of Other-reported When is it better that someone else is evaluating the focal individual? E.g. Task performance that individuals would be likely to report in a biased manner

Choosing a sample I.e. drawing a research sample from population Must be appropriate for the research question(s) E.g. if you are studying work-related phenomena, it is appropriate to sample individuals in work contexts Various sampling methods exists (there is a good summary in Wikipedia) Probability sampling: each member of the population has a (equal) chance to be sampled; e.g. when you study Finnish SMEs and randomly draw a sample from the entire population Non-probability sampling: a sample is drawn non-randomly; e.g. employees of a certain firm (convenience sample). When using non-probabilistic methods, you must carefully consider the generalizability of your findings

Sample size The higher the number of observations (e.g. number of respondents), the higher the likelihood that you will find statistically significant relationships In most research designs in management studies, the required sample size varies between 100 and 300

Validity A measure is valid when it accurately represents what it is supposed to measure E.g. which of the following is more valid measure of employee satisfaction: (I) employee s self-report of his or her job satisfaction (subjective), or (II) the number of days the employee is absent from work (objective)? To assure validity, make sure that the operationalization of the measure is entirely consistent with the theoretical definition of the construct that you are measuring

Reliability A measure is reliable when it produces consistent results E.g. Which of the following items you think could be used to form a reliable measure for job satisfaction? a) This organization has a great deal of personal meaning for me b) Most days I am enthusiastic about my job c) I want to learn as much as possible from my job d) I feel fairly well satisfied with my present job e) In my job, I prefer tasks that really challenge me so I can learn new things f) I consider my job rather unpleasant g) I find real enjoyment in my work h) I desire to completely master my job i) I do not feel a strong sense of "belonging" to my organization

Two types of variables Study variables Dependent variables (outcomes that we are trying to explain) Independent variables (what we argue explains the outcome) Control variables (rival explanations for the outcome) Control variables are any variables that may offer alternative explanations of the outcome All the most likely explanations of the dependent variable should be controlled for (especially if they are likely explain the outcome for the same reason than the independent variable) It is important that your independent variables explain a proportion of variance in dependent variable that is above and beyond the variance that is explained by control variables E.g. What control variables you would include in investigating the effect of organizational commitment on individual creativity?

Operationalization of measures Variables are always numerical, and in many cases, we must transform qualities into numerical form Nominal scale: dummy variables (two categories); e.g. 1 = condition (e.g. female) is met and 0 = condition is not met Ordinal scale: variables with values presented in rank order (e.g. high-low); e.g. 7 is the highest value - 1 is the lowest value Ratio scale: scales with an interpretative zero point; e.g. firms net sales in euros In management studies, we are often interested in things that are not directly observable Capabilities, orientations, emotions, attitudes, behaviors, etc. Thus, we use multi-item scales to better establish reliability

Likert scales Usually, 5-point or 7-point scales Midpoint of the scale (usually 3 or 4) is defined as neutral (i.e. equally characterized by both endpoints of the scale) Each item (i.e. survey question) captures a key aspect of the variable Uncaptured variable Variable captured with 6 items

Job satisfaction on a questionnaire

Classroom exercise II In your research project, you are interested in studying how team leaders inspiration leadership influences team performance Inspirational leadership = leadership style that focuses on communicating a compelling vision for the team, expressing confidence in team members, and energizing the team Based on your extensive literature review, you have learned that there are no pre-existing measures capturing inspirational leadership Thus, you decide to develop a new scale Come up with measure items that you think would capture inspirational leadership (validly and reliably) What is your data source (i.e. who do you ask) and why?

Using cross-sectional data and data with time lags When your hypotheses suggest that there is some type of change or causality between the variables, it is important that there is a time lag between independent and dependent variables Cross-sectional data shows that there is correlation, but it does not suffice to establish causality inferences E.g. collect data on inspirational leadership at Time 1 and data on team performance on Time 2 (to determine the length of time lag between Time 1 and 2, you need to estimate the effect time) Doctoral students: the use of cross-sectional data is one of the most common reasons for rejection at top journals

Common method variance/bias When common method bias is present, the measurement method is the reason for significant relationships between the study variables Is usually caused by a single individual being the source of both independent and dependent variables The most relevant bias related to research design To control for common method variance Do not use cross-sectional designs Whenever possible, obtain data from different sources (also, mix subjective with objective data) Protect respondents anonymity Avoid ambiguous questions

Part 3: Ethics in Management Studies

Ethics in research and publishing Based on the Code of Ethics of the Academy of Management (Academy of Management Journal 2011, Vol. 54, No. 6, 1299 1306.) All researchers have an obligation to be familiar with general ethical principles not knowing is not an acceptable excuse

Some essential ethical guidelines (not a complete list!) Always explicitly identify, credit, and reference the author of any data or material taken from written work, whether that work is published, unpublished, or electronically available i.e. no plagiarism of any kind Keep your promises to the research subjects Never cause any harm to your research subjects Never fabricate data or falsify results Do not omit data or findings for the sake of presenting better results Once the results are published, they are published. Do not try to publish the same (or closely related) results again

ASSIGNMENTS!

Assignment 1: Research design Based on four articles assigned for compulsory pre-reading (see the course syllabus or MyCourses), analyze the research design of each article following the four steps below: 1. Describe a purpose and a research question of the study; 2. Explain the rationales behind choosing a data collection source(s) and process (research design); 3. Identify and assess the sampling procedure; 4. Inspect the choice and quality of measurement constructs applied in the study. Conduct the analysis separately for each research article. Deliverables: 4 PowerPoint slides (1 slide per article) stored on your memory stick to be presented in the class on 20.4.

Assignment 2: Analysis methods Based on four articles assigned for compulsory pre-reading (see the course syllabus or MyCourses), analyze statistical methods applied for testing research hypotheses in each article following the six steps below: 1. Draw a model of the hypothesized relationships; 2. Inspect a correlation pattern of the variables; 3. Explain a choice of statistical methods and composition of independent variables applied for testing the hypotheses; 4. Assess a fit of estimated statistical models; 5. Interpret the results of conducted analyses; 6. Evaluate the generalizability of the findings. Conduct the analysis separately for each research article. Deliverables: 4 PowerPoint slides (1 slide per article) stored on your memory stick to be presented in the class on 25.4.

Assignment 3: Classroom SPSS exercise Group assignment You will complete the assignment during the Structural Equation Modeling -exercise session on 26.4. More detailed instructions for the assignment and the deliverables will be given in during the exercise session

Assignment 4: Coursework project The purpose of the coursework project is to improve your ability to carry out a research project using quantitative methods We employ an article case and your role is to complete the empirical parts that are missing from the paper I.e. your task is to complete the sections of the article that focus on Method and Measures, Analyses, and Results To carry out your assignment, you will receive the following 1. An unfinished article manuscript 2. A separate document that gives you some background information about data and data collection 3. A dataset 4. A separate document defining items in the dataset

What is involved? Key elements of the project: 1. Explain your sample and procedures 2. Select measures that are appropriate for the research question and hypotheses testing; don t forget the control variables 3. Prepare data and measures for analyses 4. Evaluate the psychometric properties (e.g. reliability and validity) of the measures 5. Analyze the data using appropriate analytical methods 6. Report the findings Also note: 1. You may modify the headings, structure, and all other aspects of the paper starting from the Methods section (from p. 8 onwards), but you may not change Introduction and Hypotheses sections

Assessment criteria The coursework project will be assessed based on how well and thoroughly you execute all key elements of the project (listed on the previous slide) Given that this case exercise is about preparing an academic publication, we also evaluate the extent to which the form and quality of your work corresponds to research published in top-level management studies journals (e.g. ETP, JIBS, JOB, SMJ) Overall, the coursework project will make 45 % of your course grade Written report (i.e. article): 35 % Class presentation of your work: 10 %

Returning your work You will present your coursework project during the class on April 28 Please prepare some slides or other illustrative material In your presentation, focus on two primary issues: What choices and analyses you did? and Why did you make these choices / analyses rather than some others? Also, you are required to hand in the finished paper on or before May 15

What did we learn today? 1. We understand the role of research questions in quantitative management research 2. We know how to develop testable hypotheses 3. We can assess different data sources and choose an appropriate sample to test your hypotheses 4. We have gained a basic understanding of different forms of measures in management studies 5. We understand and appreciate the ethics of management research 6. We know what is expected of us in this course