The type of data that is collected determines the statistical tests that are used. This can be summarised as follows:
1. Looking for differences between groups. These are average differences between two separate groups, between two related groups and among three or more groups. The tests are independent-samples t-test, matched-pairs t-test and analysis of variance (ANOVA) with the Post-hoc follow-up tests.
2. Looking for relationships. This can be average relationship between two variables, relationships among clusters of variables or predicting one variable from one or more variables. The associated tests are Pearson's r, Factor analysis, and Regression analysis.
3. Looking at ratings. This includes looking for differences in ratings between two independent groups, differences in ratings between two related groups, differences in ratings among three or more groups and relationships between pairs of ratings.
The respective tests are Mann- Whitney U test, Wilcoxon Signed-Rank test, Kruskal- Wallis test, and Spearman's Rho.
I would suggest adding a point 0: describe the context of the study, the sample you have (in particular, how far it can be considered as a random sample), and the available variables, including the unit of measurement, and then formulate the research question.
For point 1, I don't think that a formal test of the assumptions is impossible without having decided on the statistical methods. I would rather suggest an exploratory analysis with informative, simple plots (not pie charts or superposed charts).