New version 1.6.8: clean up effect structures; one step estimator.

[TomSnijders]

Description

• Reorganization and clarification of effectGroups for combinations
  of two dependent networks and an actor covariate.
• More possibilities with one-step estimator.
• Further, see NEWS.md.

Checklist:

Checks

• The package builds on my OS without issues
• My changes generate no new warnings

Documentation

• I have commented my code inline, particularly in hard-to-understand areas
• If function added/modified, I have added/modified documentation (in .R file)
• I have added a description of all changes to this pull request above and to the NEWS.md file
• I have bumped the version in the DESCRIPTION and man/RSiena-package.md files by the appropriate increment:
  • 'major' (incompatible API changes)
  • 'minor' (add functionality in a backwards compatible manner)
  • 'patch' (backwards compatible bug fixes)

[csteglich]

Passes checks and looks good (I checked only "effects.R" in more detail).