I have 4 groups that identify different research field (animal, plant, microbe and insect). Scientists were asked to indicate in which group they belong and they were allowed to indicate more than one group - i.e. a scientist can be both in "animal" and "microbe". We then ask scientists a set of questions about time spent on research activities. We would like to compare means across groups. Normally, I'll use ANOVA but in this case the four groups are not mutually exclusive. Can I use t-test comparing each group against the rest of the population, i.e. those in the Animal group against all others? Is that appropriate? Are there better methods? I currently use R for my analysis.

  • $\begingroup$ Consider a single one of these proposed comparisons.and note that the people that are in both groups being compared contribute no information to identifying a difference (their data are paired but the variance of their internal differences are all zero). If you properly account for their presence you'll add nothing over leaving them out altogether. $\endgroup$ – Glen_b -Reinstate Monica Mar 11 '17 at 0:53
  • $\begingroup$ @Fedrica Fuci time spent on research activities - percentage on different classified fields or hours etc. ? $\endgroup$ – Subhash C. Davar Mar 11 '17 at 8:48
  • $\begingroup$ Hi @subhashc.davar, thanks for your comment. Time spent on research activities is a continuous variable, measuring the number of hours they spend on research. Respondents were free to enter any number. We also want to compare a Likert scale across groups. $\endgroup$ – Federica Fusi Mar 11 '17 at 21:42
  • $\begingroup$ @Federica. Fusi For the purpose of analysis,You can make five groups - one with mixed i.e. two or more area of specialisation. Would you post your Scale and scoring method. $\endgroup$ – Subhash C. Davar Mar 12 '17 at 3:28
  • $\begingroup$ @subhashc.davar It's a five-point Likert scale created by combining 5 items. The idea of a mixed group is interesting. I didn't think about it $\endgroup$ – Federica Fusi Mar 12 '17 at 19:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.