Detailed information about the course

[ Back ]
Title

Scale Development and Validation Procedures in Information Systems

Dates

Avril-June 2021

Responsable de l'activité

Kévin HUGUENIN

Organizer(s)

Dr. Jean-Charles Pillet

Speakers

One of the following external speakers (depending on availabilities): Prof. Kai Larsen, University of Boulder Prof. Rajeev Sharma, University of Waikato (New Zealand)

Description

Construct development is cornerstone to the expansion of knowledge in survey-based research and is common ground in disciplines as varied as marketing, organizational behavior, strategy, entrepreneurship or information systems management. From a set of disparate practices, construct development and the related validation procedures have grown into a cohesive framework that has been iteratively enriched over the past 50 years (Churchill 1979; DeVellis 2003; Diamantopoulos and Winklhofer 2001; Edwards and Greenberg 2003; Hinkin 2005; Lewis et al. 2005; MacKenzie et al. 2011; Straub 1989). For PhD students, appropriating the constantly expending toolkit pertaining to the measurement of social constructs is particularly challenging. The purpose of this doctoral seminar is to facilitate this appropriation process while exposing participants to some key debates and issues in the area. Voir le syllabus joint en PJ pour une description détaillée des activités, estimation de charge de travail, et modalités évaluation. Charge de travail estimée: 75 h (3 ECTS)

Program
  • Session 1) April 21 09-12:00
  • Session 2) April 28 18:30-21:30 (R. Sharma, from NZ)
  • Session 3) May 12 14-17:00
  • Session 4) May 26 14-17:00
  • Session 5) June 02 09-12:00
  • Session 6) June 09 14-17:00

Motivation

Construct development and validation is cornerstone to the expansion of knowledge in positivist, quantitative research, in disciplines as varied as marketing, organizational behavior, strategy, entrepreneurship or information systems management. From a set of disparate practices, construct development and the related validation procedures have grown into a cohesive framework that has been iteratively articulated and enriched over the past 50 years. For PhD students, appropriating the constantly expanding toolkit pertaining to the measurement of social constructs is particularly challenging. The purpose of this doctoral seminar is to facilitate this appropriation process while exposing participants to some key debates and issues in the area.

Learning goals and approach

At the end of the course, the participants should be able to:

  • engage in a rigorous conceptualization leading to a new construct;
  • design high quality questionnaires that minimize response effects and yield results;
  • present preliminary evidence of validity and reliability of their measurement scales.

 

In order to achieve these goals, the seminar builds on the following pillars:

1) A combination of the most established and of the most promising methodologies;

2) An applied approach to the issues introduced theoretically (e.g., implementation exercise);

3) A space for discussing disciplinary expectations and sharing best practices.

 

 

Seminar Structure

To achieve these goals, this seminar is structured around six sessions of three hours each. These sessions follow the linear sequences of the scale development process, highlighting key issues that could arise at each stage.

 

 

Session

 
 

Topic

 
 

Introduction

 
 

1

 
 

Construct Development Process and Key Concepts (JC. Pillet)

· Latent variables, construct dimensionality, construct proliferation, nomological network, construct "mixology", new versus adapted scales, types of validity.

Activity: first exposure to a published scale development paper

 
 

Step 1. Define the Construct

 
 

2

 
 

Construct Conceptualization and Definition (R. Sharma)

· What is a theoretical contribution, characteristics of good definitions, establishing of conceptual distinctiveness.

Activity: review of proposed conceptual definitions

 
 

Step 2. Operationalize the Construct

 
 

3

 
 

Development of Measurement Items (JC. Pillet)

· Inductive vs deductive generation, cognitive processing of items, undesirable item characteristics, formative vs reflective measurement scales.

Activity: review of operationalizations

 
 

4

 
 

Pretesting and Assessing Measurement Scales (JC. Pillet)

· content validity assessment and evidence, face validity, pretests.

Activity: review of instrument pretest strategies

 
 

Step 3. Establishing Construct Validity

 
 

5

 
 

Survey Design and Administration (JC. Pillet)

· Dirty data, order effects, common method effects and mitigation strategies, mode of administration, web questionnaires, item distribution strategy, crowdsourced samples.

Activity: review of online questionnaires

 
 

6

 
 

Examining Measurement Properties (JC. Pillet)

· Dataset screening and cleaning, exploratory factor analysis, scale purification, evidence of validity and reliability.

Activity: analyze dataset using software

 

Evaluation

Research project: the resulting paper should have a substantive focus or a methodological focus and requires collecting empirical data through a survey questionnaire. This assignment should help participants solidify their research pipeline. It has two parts:

· Research proposal - 3 pages, 20% of final grade, due 2 weeks after the first session

· Term paper – 10 pages, 80% of final grade, due 4 weeks after the last session

 

Evaluation details will be shared with the participants on due time.

Seminar Spirit

This class is designed in a seminar format. This heart of any Ph.D. seminar is discussion and analysis of assigned readings. To do that, you must read all assigned readings before class, think about these issues throughout the semester, debate these issues with your peers, and synthesize these issues mentally to develop yourself as a researcher. Please kindly note that if you do not come to class fully prepared, you will be totally lost and will not learn much from the class.

 

Estimated Workload

· In-class hours: 18h

· Readings: 20-30h

· Assignments:

o Research proposal: 10-15h

o Research project: 40-60h

· Estimated workload: 88-123 hours

 

Prerequisites

Participants should have basic knowledge of statistics and of the underlying assumptions of ordinary least square regressions in addition to having access to a statistical analysis software to perform exploratory factor analyses (e.g., SPSS, Stata, R, MPlus, AMOS, SmartPLS, etc.).

 

Participants are expected to run a survey and collect a number of responses that will allow them to evaluate the measurement properties of their instrument. Thus, securing budget for these activities early in the process is essential. Crowdsourced samples cost approximately CHF 7.50 per respondent (e.g., www.prolific.co). In addition, participants should anticipate the need to collect Ethics approvals from their home university if this is required.

 

Data collection is typically performed using online survey tools. While free software is sufficient for the purpose of this class, it may not offer the same depth as priced solution (e.g., Qualtrics) and may not be compliant with your institution's data collection policy. Please enquire about options at your home university.

 

Location

UNIL (ou Zoom)

Evaluation

Research project: the resulting paper should have a substantive focus or a methodological focus and requires collecting empirical data through a survey questionnaire. This assignment should help participants solidify their research pipeline. It has two parts:

· Research proposal - 3 pages, 20% of final grade, due 2 weeks after the first session

· Term paper – 10 pages, 80% of final grade, due 4 weeks after the last session

Information

jean-charles.pillet(at)unil(dot)ch

Registration

Flease fill this form: https://forms.gle/tyCgNTgstTRy3RYo8

Places

10

Deadline for registration 16.04.2021
Contact

jean-charles.pillet(at)unil(dot)ch

short-url short URL

short-url URL onepage