A Handbook_NRA - Extended

Moderating 1. The relationship between (IV) and (DV) is moderated Hypos: by (MV). 10 Different 2. The relationship between (IV) and (DV) is positively Structures moderated by (MV). 3. (MV) moderates the relationship between (IV) and (DV). 4. (MV) has a positive moderating influence on the relationship between (IV) and (DV). 5. Increase in (MV) increases the strength of the relationship between (IV) and (DV). 6. The higher the (MV), the stronger the positive effect of (IV) on (DV). 101

Self-assessment: Based on the list of moderating hypos you compiled, which structure appeared the most? Continue…. 7. The positive effect of (IV) on (DV) will be enhanced as (MV) increases. 8. (MV) enhances the effect of (IV) on (DV), such that the effect of (IV) on (DV) will be stronger when the level of (MV) is higher. 9. (MV) moderates the relationship between (IV) and (DV); specifically, the relationship is stronger at higher levels of (MV) than at lower levels. 10. (IV) will have a stronger positive impact on (DV) when (MV) is high than when it is low. Note: To avoid complications, please select moderating hypothesis structures from number 1-8 only! 102

Examples: Moderating Hypos in Action 1. Transformational leadership (MV) has a positive moderating influence on the association between sales strategy (IV) and sales force performance (DV). 2. The negative effect of product innovativeness (IV) on product quality (DV) will be reduced as info integration (MV) increases. 3. Increase in perceived product similarity (MV) decreases the strength of the relationship between perceived manufacturer brand quality (IV) and intention to buy manufacturer brands (DV). 4. Innovative culture (MV) enhances the effect of leadership (IV) on tech innovation (DV), such that the effect of leadership (IV) on tech innovation (DV) will be stronger when the level of innovative culture (MV) is higher. 103

What can we Conclude from Task 2 The PATTERNS… As stated earlier, NRA is particularly well-suited for 1. So many ways to design Moderating Models moderating or interaction • different styles, shapes: simple, unique (endless) effects. Nevertheless, the NRA • models should “look good” (reviewer’s mood) method can still be used to • visually display variable relationships (linkages) develop ‘pure’ Direct Effect models and Mediation models. …. BUT…. Those who plan to 2. Only 10 ways to write Moderating Hypotheses incorporate mediators in their • no need to create a new hypo sentence (same) thesis are advised to search for • if develop your own sentence (likely to de-rail) at least 5 journal articles with • use an existing structure (academic standard) mediating effects; to distinguish the patterns in model and hypo structures. Note: For mediating models, however, the data should be analyzed using PLS-SEM… a brief discussion on this will be provided in Chapter 7. 104

Design Your No matter how good you are at writing MODEL: Getting moderating hypos, you will eventually end up with one of the 10 hypo structures shown earlier. Started Therefore, to avoid unnecessary struggles, you can simply choose a set of hypos from the • Choose a Simple or Average model, but articles you collected, or alternatively, select a with an attractive hypo sentence structure preferred direct and moderating structures that I have outlined here in this book. Similarly, for • You need to understand HOW it’s done first your conceptual model, you can just choose a (but publishable: at least a Research Book) model that you feel is most attractive from the articles you compiled. • Once you have completed the Novel Workshop, then we can work on a more This procedure provides the platform for you Sophisticated model (perhaps for your PhD to “initiate” your research. Specifically, you will Proposal or Journal Publication) know how many IVs, DVs, & MVs to search for in your study. Note: The model & hypo structure that you choose will likely change (slightly altered) when you latter incorporate your variables. Can you see now? The NRA does not 105 start with the Research Problem or Literature Review; it starts with finding a suitable design & structure for your conceptual model and hypothesis, respectively.

To understand how the BEST model and hypo structures are selected, please refer to GradEx Academy’s website and search for workshop Video Recordings: Number 3. The Your best model and hypotheses Blueprint structure represents your BLUEPRINT • hold on tight to your blueprint article(s) • it will be used to develop your model • it will also guide you latter in your write-up, story telling, and arguments Now let's find your research TOPIC and VARIABLES … 106

These questions are 1. Find 7-10 journal articles on a MARKETING specifically outlined for topic from the CUSTOMER’S perspective… the 3-day NRA workshop; or any area related to your research interest that’s why it is advised (management, education, economy, NOT to use managers as psychology, public health, etc.) that involves respondents (to ease data employees, general public, or students as survey respondents… NOT managers. collection). 2. List down all their IVs, Ms, & DVs in a Table Task 3 (30 minutes) 3. Based on the list, identify the IV(s) that you think is the most interesting. Why? ----------------------------------------------------------- Note: At this stage, the most important is the IV; To thoroughly understand this it will determine your research TOPIC Chapter and get an idea of how to address these tasks/questions, please refer to GradEx Academy’s website and search for workshop Video Recordings: Number 3. 107

Example (for Task 3) Example of articles with their IVs, DVs, MVs, & MedVs, respectively… Note: Pay attention to your list of IVs: is there an IV(s) that you particularly like? 108

Task 4 Once we have decided on your IV… (30 minutes) 1. Based on your List, select a suitable DV, and then potential MVs for your research model 2. Next, insert all the variables in your chosen Model and Hypo sentence (blueprint) 3. Examine the variable relationships… • Do they make sense (logical)? • Is it worthwhile studying the effect (so what)? • Is there variety in your hypo directions (mixture)? • Did you manage to find the Cherry (bonus)? To understand this Chapter 4. Repeat process 1-3 until you are satisfied and how these tasks can be 5. Present your proposed model & hypos addressed most effectively, please refer to GradEx Academy’s website and search for workshop Video Recordings: Number 3. 109

Variable Selection Criteria Selecting the variables for your study is vital for three reasons: (1) to get published in a reputable journal, (2) to make your model interesting, and (3) to increase your chance of getting significant results in your data analysis. 1. Must be within the CONTEXT of study Note: If you want to get • logical / in practice / theoretical grounds published in a good journal, 50% 2. Choose variables with MEASUREMENTS only • papers with survey / questionnaire items of your hypotheses should be ‘supported’. Achieving significant 3. Select an INTERESTING variable statistical results starts with proper variable selection… so try your best to fulfil the criteria listed here. • sounds exciting / not too common / rarely tested 4. One moderator that qualifies as a CHERRY • this is optional / bonus (*but necessary for Q1 or Q2 journals) 110

To thoroughly understand this Chapter and get an idea of how to address these tasks, please refer to GradEx Academy’s website and search for workshop Video Recordings: Number 3. Task 5 After presenting Task 4… (1 hour) 1. Improve your model based on the comments - you might have to find more journals if your list of variables is not enough 2. Provide a DEFINITION for all your variables 3. Show your finalized model and hypos -------------------------------------------------------- Note: If you have time, try to find a Theory for your conceptual model. If you didn't find one, we will solve this matter in the LR session. 111

Example (for Task 4 & 5) A “sketched” model produced by one of the participants in a recent NRA workshop… developed in just 2 days! A list of direct and moderating hypotheses introduced by the participant… meets PhD standard & this is journal publication material! 112

113

Chapter 4 Designing Effective Questionnaires Stage 2: Applied Novel Research Section 114

“The Simple always produces the Marvelous” Being simple may be the hardest thing to do for academicians & PhD students. The consequences are heavy if not given careful attention at this stage. So let’s learn how to develop a “simple” but “effective” questionnaire. 115

Highlights of  Scaling Issues (Which type is most suitable?) this Chapter  Characteristics of a Good Questionnaire (Tips)  Avoid Glaring Mistakes (Wordings and structure)  Questionnaire Administration (Pre-test, cover letter)  Finding your Measures (Can you use existing ones?)  Constructing your Questionnaire (Getting it right)  Develop your Online Survey (Collect your data now!)  Other Concerns (Demographics, pilot test, controls?)  Pulling it all Together (Model + Hypo + Def + Quest = 50% of your study!)  Mock Table (Making sure it works) 116

What is the Mid-Point of Research? To know the answer for this question Hint: A well-thought out questionnaire can be the and why the mid-point is important, difference between success please refer to GradEx Academy’s and failure. website and search for workshop 117 Video Recordings: Number 4.

Questionnaire as Instrument Surveys are widely used to gather data in various fields,  Serves as an instrument to acquire info including social science, - most common method for collecting primary psychology, public health, and data (survey) political studies. The survey method is suitable if you want to  Directly linked to your research find out about the perceptions, - measures all variables related to your preferences, opinions, experiences, research and model (connected) and behaviors of a group of people.  Used to test your research hypotheses Note: For the NRA workshop, - to find statistical evidence to support/reject we will focus our attention on your hypos (quantitative) customers as this will be easier and faster for participants to collect data (online survey). For your own thesis, however, you can go ahead with your target sample (e.g., managers, employees, students). The measures in your questionnaire What is a research variable? represents each of your variables (IVs, DVs, Specific properties that you intend Moderators/Mediators) in your model – main objective is to statistically examine the variable to study and can be measured relationships that you have predicted quantitatively. (hypotheses). Note: Never reveal your model to your participants… to avoid bias answers. 118

Measurement Scales Measurement scales are used to capture data… questionnaires are 1) Nominal (categorical choices) 2) Ordinal (rank-order responses) derived from these 4 types of measurements that quantify variables for statistical analysis. 3) Interval (numerically equal distance) 4) Ratio (absolute response) At PhD level, you need to understand beyond the mere differences between the 4 scales: Which scale is normally used to  Can nominal scales be used for regression analysis? test directional hypotheses?  Are ordinal scales common in academic research?  What about ratio scales, are these commonly used? To know the answer for  Which scale is used to test direct & moderating effects? these questions, please refer  Can interval scales be tested with ratio & nominal scales?  What is dichotomous scale? How should we code this? to GradEx Academy’s website and search for  Answers to these questions are vital for your research, workshop Video Recordings: particularly laying out your objectives, creating your model, and analyzing your data (avoiding major flaws). Number 4. 119

What type of scale is this? 120

Examples of Nominal Scale Nominal scale is a categorical scale where variables are simply “labeled” or identified/named. • Gender: Male, Female • Marital Status: Bachelor, Married, Other. • Education: Diploma, Degree, Masters, PhD • Employment: Public, Private, Self-employed, Homemaker, Student, Other. • Ethnicity: Malay, Chinese, Indian, Other. --------------------------------------------------------------- What about Age & Income…. Nominal or Ratio? 121

What type of scale is this? 122

Examples of Ordinal Scale Ordinal scale has all its variables in a ranking order. Miss Malaysia Results? • First, Second, Third Rank the Order of Preference? • Mister Potato, Chipster, Pringles, Cottage Fries, Lays Potato Chips 123

What type of scale is this? The most common format used in a survey is the “agree-disagree” scale (Likert scale)… interval scale. 124

Is this Interval Scale? 125

Examples of Interval Scale Interval scales offer a numerical “equal distance” between responses. 126

Interval …. 127

Words used for Interval Scales 128

Words used for Interval Scales 129

What type of scale is this? 130

Ratio scales bare absolute Age response… request for the Weight of a Package of Nuts actual number or figure. Sales Figures Monthly Income Examples Years of Education of Ratio Number of times visiting a restaurant Scale Number of children 131

“Filling a questionnaire is a complex process. The job of a questionnaire designer is to make it easy for respondents to respond.” The most Difficult Skill is to be Simple… This may be one of the main reasons why many Academics have been trained to be researches have problems meticulous and highly intellectual when it with their statistical results… mostly NOT comes to writing papers and doing significant (p > .05). research. There is one exemption though – 132 when designing the questionnaire…

Characteristics of a Good Questionnaire • Asks relevant questions Questionnaire design involves a - questions that will give answer to your complicated process because it requires research questions/objectives/hypos attention to many details at once. These • Should be short include finding, selecting, adjusting, - to avoid fatigue (5-7 pages) organizing, writing, and testing the • Instruction and wording should be simple measures, among others. - clear and easy to understand Note: I did not mention “developing” the measures… did you realize this? • Presented in appropriate order - begin with easy, non-challenging questions first You should avoid questions that are not relevant or not directly related to your • Use fixed-response questions study’s objective. - do not use/minimize open-ended questions All instructions and question items should be easy to understand, narrowly focused or precise, and with enough substance so that respondents can provide accurate answers. 133

Characteristics of a Good Questionnaire • Make sure the question items matches your Note: The questionnaire items construct and definition - you can choose the most “fitting” set of items should be consistent with your • Should look professional variable definition. Hence, you can - good appearance, easy on the eye… choose a set of measurements – • Possible follow-up perhaps from 2 or 3 different - with reward, charity, or summary of key findings sources – that best suits your - to increase response rates study’s context. • Use a cover letter A cover letter shows your professionalism and demonstrates attention to detail. 134

Example of Cover Letter 135

Tips: Questionnaire Development To ensure the validity and reliability of your  Use established measures measurements, you have to carefully consider - do not develop your own measures… each question (items) in your survey. That’s - adapt/adopt questions from previous studies why you have the option to “adapt” your measurements to your study. - to obtain high Cronbach’s Alpha score Above all, use ONLY established or  Choose measures from the same author who previously validated measurements from leading international journals. defined the variable - to ensure accuracy/consistency between variable definition and measurement Note: PhD & journal paper  Choose measures with 5-7 items - to avoid item overlap and shorten the questionnaire examiners will scrutinize your - make sure from a good journal measurements for any defaults… it’s the most easiest and quickest  Pre-test with 2 Professors and Proofread! - free from content and grammatical error way to find your weakness, so take extra care when designing your questionnaire! If problems arise in your pre-test, it is likely that similar problems will occur in your actual data collection. Problems and solutions should be acknowledged early. 136

“It can be difficult to get the phrasing of the question right, so it means the same thing to all respondents.” 137

Avoid any words that Wordings to Avoid… may be unclear to a respondent… e.g., • Avoid sophisticated/high academic language acronyms and • Avoid ambiguity (might be necessary to provide an unfamiliar words. example) “How would you rate our • Avoid double-barreled questions (… and …) product, specifically in • Avoid negative-wording questions unless necessary terms of its durability and (e.g., switching intention, compulsive buying behavior, reliability?” customer indecisiveness, green product skepticism) Note: Some respondents • Avoid leading questions A well-written question may feel that you offer a • Avoid long sentences highly durable product but • Avoid redundant questions allows respondents to give has reliability issues. truthful answers without being “pulled” to one side or the other. The questions should also not leave respondents feeling confused about which option to select. 138

“It can be tempting to put lots of explanation into the question – but sometimes this just confuses things even further.” 139

Common Mistakes What might be wrong with this question? Need to cover all options 140

Common Mistakes What might be wrong with this question? Options should be consistent. 141

Common Mistakes What might be wrong with this question? Options should be balanced. 142

Common Mistakes What might be wrong with this question? Separate the question if it can be broken down into parts… 143

Common Mistakes What might be wrong with this question? A non-leading question would be… 144

Rushing to distribute your Questionnaire questionnaire without pre-testing Administration can result in unforeseen problems with data collection and analysis. Pre-test the questionnaire Survey pre-test helps you avoid • to detect weakness and improve those issues and raise the quality clarity of your data. Content of the cover letter • purpose, importance, estimated time range, assure confidentiality Improving response rates • shorter quest, personalize the questionnaire, contact the respondent beforehand, follow-up 145

Types of Questionnaire Self- • Postal completed • Electronic Interviewer- • Face-to-face led • Telephone 146

The PROS Speed - Scale - Quality? & CONS: Delivery • online, mail, phone surveys: far reaching & Modes can reach a broad range of people • person administered: time-consuming, closer locations, but will get quality feedback Respondent’s Character? • person, phone calls: feel shy to answer certain questions • but can provide clarifications if respondents ask Personal Questions? • online survey: suitable for private issues (e.g., Smoking, Compulsive buying, Switching intention) Here are some vital aspects for you to consider when choosing a preferred method. Note: For this workshop, however, the Online Survey method is advised – to collect data within a reasonable timeframe. 147

Mid-point or Neutral opinions The issue of whether or not to offer a midpoint has been disputed for decades... Using mid-points in a questionnaire… Strongly   Strongly Disagree Agree • Historically controversial: it was thought that respondents chose a neutral option to avoid thinking about the question • Now shown that including a mid-point increases accuracy of responses (reliability & validity test) Note: Studies have also shown that respondents use the mid- point category more often as a “true neutral meaning” than a “don’t know” response. 148

Questionnaire: Other Concerns • Where can you find the measure(s)? To know the answer for • Where should you place demographic questions these questions, please (age, income, education, marital status)? refer to GradEx • Can you place the DV measures in front? Academy’s website and • Difference between Pre-test & Pilot test? search for workshop Video • Can you combine variables (as one construct) in the Recordings: Number 4. data analysis? • What are control variables? Should we have them in our study? • What is common method bias & how to overcome them (practical approach)? • How does questionnaire relate to CFA, Correlation Matrix, and Regression Analysis? 149

Task 6 1. Find existing measurements for each variable in (30 minutes) your model (refer to Task 4, should be at hand) 2. Indicate the SOURCE (journal) for all your variables 3. Develop a draft of your questionnaire - with instructions, sections, construct, items, and scale (use my Questionnaire Template) 4. Assess your questionnaire (any problems?) 5. Present your Questionnaire and Cover Letter To thoroughly understand this Chapter and get an idea of how to proficiently complete these tasks, please refer to GradEx Academy’s website and search for workshop Video Recordings: Number 4. 150