Descriptive Analytics and you will Design away from Commonly used Terms

I checked potential differences by site, geographic region, and ethnicity playing with t-evaluation and study regarding variance (ANOVA) on the LIWC classification percentages. Toward a few other sites, half a dozen of the several t-examination was in fact high in the after the classes: first-people only one [t(3998) = ?5.61, p Supplementary Desk 2 for function, basic deviations, and contrasts anywhere between cultural groups). Contrasts revealed high differences when considering Light as well as other cultural communities into the five of the half dozen extreme ANOVAs. Hence, we included ethnicity just like the a beneficial dummy-coded covariate when you look at the analyses (0 = White, 1 = Every other ethnic communities).

Of the a dozen ANOVA evaluation associated with geographical part, just one or two was in fact high (family unit members and you will confident emotion). Just like the distinctions were not officially important, i failed to consider geographical area inside next analyses.

Performance

Frequency out of phrase have fun with is obvious when you look at the detailed statistics (select Desk step one) and via phrase-clouds. The expression-cloud techniques illustrates one particular commonly used terminology along the whole take to plus each one of the a long time. The term-affect program automatically excludes certain terms and conditions, in addition to stuff (good, and you will, the) and you may prepositions (to, which have, on). The remainder blogs terms and conditions try scaled in proportions relative to their frequency, creating an user-friendly portrait really commonplace stuff terms round the brand new take to ( Wordle, 2014).

Profile 1 shows the newest 20 popular blogs words utilized in the complete attempt. As well as be seen, probably the most commonly used terms had been love (lookin when you look at the 67% from pages), for example (lookin when you look at the 62% off users), looking (lookin from inside the 55% out of pages), and someone (lookin in the 50% regarding profiles). Therefore, the most popular conditions had been equivalent across the a long time.

Figure 2 suggests another 30 common Hangzhou beautiful women posts conditions into the the newest youngest and you can oldest age groups. By eliminating the initial 20 well-known content conditions along side test, we show heterogeneity from the relationships users. In the next 30 terminology on youngest age group, raised percentage words provided score (36% out-of pages in the youngest age bracket), go (33% of users from the youngest age bracket), and you may really works (28% off profiles from the youngest generation). In contrast, the new eldest age bracket got high percent from conditions such as travelling (31% away from users on oldest age group), great (24% of users throughout the earliest age bracket), and you can matchmaking (19% away from pages regarding oldest generation).

Next 31 most common terms and conditions on youngest and you can earliest age groups (shortly after deducting this new 20 common terminology from Figure step 1).

Hypothesis Testing of age Variations in Code in the Relationships Users

To check on hypotheses, the fresh percentage of terms and conditions about dating character that fit for each and every LIWC group served given that depending details during the regressions. We looked at age and gender since the independent parameters plus changing to have site and you may ethnicity.

Theory step 1: Earlier ages could well be associated with a high portion of terms and conditions throughout the following the categories: first-individual plural pronouns, family unit members, relatives, health, and confident feelings.

Results largely served Hypothesis step 1 (find Table dos). Five of one’s four regressions found a critical head feeling to own years, in a fashion that because the period of the fresh reputation journalist increased, the percentage of terminology in the classification enhanced on the adopting the categories: first-people plural, members of the family, wellness, and you may self-confident feeling. We discovered no high age impression on the proportion away from conditions from the family members class.

a Gender: 0 (female) and you may step one (male). b Website: The two websites have been dictomously coded as step one and you can 0. c Ethnicity: 0 (White) and you can step one (Cultural or racial fraction).

a beneficial Gender: 0 (female) and you will step 1 (male). b Webpages: The 2 other sites was in fact dictomously coded because 1 and you can 0. c Ethnicity: 0 (White) and you will step one (Ethnic otherwise racial minority).