Research design question

Hi guys

I would like to explore the relationship between part-time work and job satisfaction and intention to quit in two different industries

I have a data set for two different industries with the following variables:

demographic variables:
age (11 levels)
sex (female and male)
education (5 levels)

professional demographics:
employment (part-time, full-time)
size of the company (7 levels)
function (6 different type of works, e.g. marketing, customer support, etc.)
collar (blue collar, white collar)
management level (3 levels)
scope (number of countries the company is operating) (3 levels)
tenure (years in company) (7 levels)

Job satisfaction (5 levels)
Intention to quit (6 levels)

What would be a good research design to explore the relationship?

Some clarifying questions:

  • Is your data cross-sectional (one observation per individual), or is it also longitudinal (a sequence of observations over time for each individual)?
  • Do you want to look at job satisfaction and intention to quit as two separate outcomes, or is there one outcome of primary interest?
  • Are the levels for job satisfaction and intention to quit scalar such that the difference between levels is consistent, or are those measures really more categorical than scalar?
  • What's your sample size? I.e., how many observations do you have? (And if the data are longitudinal as well as cross-sectional, how many individuals are there?)

the data is cross-sectional
the data is not longitudinal
I have data from 19 European countries for the two industries
4125 observations

I would like to compare job satisfaction and intention to quit of part-time workers in the finance industry and health-care industry.
Job satisfaction and intention to quit is of interest to me.
I guess job satisfaction affects the intention to quit.

Job satisfaction:
0 = very dissatisfied
25 = Dissatisfied
50 = Neutral
75 = Satisfied
100 = Very satisfied

Intention to quit:
1 = “I plan to leave as soon as an opportunity arises with another organization”
2 = “I plan to leave for another job within the next year”
3 = “I plan to take another job as result of possible redundancy / layoff in the next year”
4 = “I plan to stay for at least another year”
5 = “I plan to stay for the foreseeable future”

Well, there are a lot of options, but here are a few thoughts.

  • Your "intention to quit" measure is kind of scalar, but definitely not interval, so you probably should treat it as categorical rather than numeric. You could summarize it into a binary (e.g., 1-3 vs. 4-5) or use something like multinomial logistic regression.

  • You have individuals nested in countries crossed with two industries. I would estimate a hierarchical (i.e., mixed-effects) model with random intercepts for countries and only revert to a fully pooled analysis if those results implied that there weren't significant differences across countries. I would also start with separate models for the two industries and only pool them if you got very similar results across the two. Otherwise, you could either include industry as a covariate or try a mixed-effects model with crossed effects between countries and industries.

  • I would rescale that measure of job satisfaction to -2 to +2. That will make interpretation of the results easier.

Thanks for your answer.
I guess I my research question is almost to difficult for an undergraduate student.
I am not really familiar with a mixed-effects model.

Should I first compare part-time and full-time employees with regard to job satisfaction and intention to quit with a chi-square test (or a t-test with a 5% significance level)?

In a second step I could run a regression with control variables (gender, etc.).

Well, I guess I have to learn more about statistics...

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.