The SEM Survey Blueprint: design a questionnaire your model can actually use

You only get one shot at your SEM questionnaire.

Once the responses are in, nothing can be fixed. You cannot add a missing item, undo a leading question, or change a scale you switched halfway through. By the time the problem shows up in SmartPLS, it is already 210 responses too late. Most candidates discover this the hard way, during analysis, when the model refuses to converge.

This blueprint is the deck I use to teach questionnaire design in the SEM master class. It turns each latent variable in your model into questions it can actually measure.

What you get

1. The 5 non-negotiable rules of a SEM-ready questionnaire

2. A worked example: effort expectancy measured with real, validated items

3. The "borrow and cite" method reviewers trust, with sources

4. The 10× rule, so you know your real minimum sample size before you launch

5. A pre-launch checklist: screening, pilot, coding, and the mistakes that quietly kill a dataset

Who this is for

Doctoral candidates and researchers running a survey-based SEM or PLS-SEM study in management, marketing, information systems, education, or the social sciences. If you are about to send a questionnaire to respondents, this is for you.

What you will be able to do

• Write 4 to 5 clean items per latent variable, on one consistent scale

• Reuse validated scales correctly, and cite them in your methods chapter

• Calculate the minimum sample size your model actually requires

• Pilot and lock your instrument before a single real response comes in

• Hand SmartPLS data that is coded, labelled, and ready to estimate

Get the blueprint

Drop your email below. The full deck lands in your inbox in under 2 minutes.