Diglossia involves the use of at least two distinct variations of the same language or dialect, one considered "high" and one "low," within a single society. Which variation is used depends on the social situation. For example, people probably speak differently in a job interview than they do when unwinding with friends. Diglossia research can record and analyze the variations of language used in different situations and by different populations within the same society. Collect data through participant observation or through silent, "fly on the wall" observation.
A sociophonetic research project could focus on the way sounds are produced in a certain dialect. A past research project at the University of Washington explored the way native English speakers from the Pacific Northwest pronounce vowels. Select a sample of speakers from a certain geographic region and record them as they read a list of pre-determined words. Then record a casual conversation. Computer software, such as Praat, can be used to analyze the recordings. Compare your results to previous geographic studies.
Previous sociolinguistics research, such as that included in Nancy Bonvillain's 2007 book, "Language, Culture, and Communication," has explored language-based gender differences. According to Bonvillain, women tend to use more honorifics, or phrases that signify respect, and hedges, or phrases that soften speech such as "maybe" or "I was wondering if." A gender-based research project could analyze rates of hedging or honorific use by men and women. Collect data through conversations or through observation.
People from the same geographic region, the same ethic or cultural group, and even the same family can exhibit wide variations in language use. A study at Ohio State University explores language changes through three generations of the same Columbus, Ohio, family. A chronological research project could collect language data from different age groups in either a spoken or written format, such as in letters or emails, and compare variations.