I tested prospective differences by the website, geographical area, and ethnicity having fun with t-evaluation and you will studies from variance (ANOVA) into LIWC category percent. Into the a few other sites, half dozen of your a dozen t-evaluating have been significant on the adopting the groups: first-people only one [t(3998) = ?5.61, p Supplementary Table dos for function, important deviations, and you can contrasts between cultural organizations). Contrasts revealed tall differences when considering White and all of other cultural groups inside the five of your own half a dozen tall ANOVAs. Therefore, we integrated ethnicity because a dummy-coded covariate within the analyses (0 = Light, 1 = Other cultural communities).

Of twelve ANOVA evaluation about geographical region, merely several have been high (relatives and you will positive feeling). Just like the distinctions weren’t commercially meaningful, we did not believe geographical region in then analyses.

Overall performance

Regularity away from word have fun with is evident from inside the detailed analytics (find Desk 1) and you may thru keyword-clouds. The phrase-affect strategy portrays probably the most popular terms and conditions over the whole shot and in each one of the a long time. The definition of-affect system immediately excludes specific terms, and additionally blogs (good, and you can, the) and you may prepositions (in order to, with, on). The remaining stuff terms and conditions is actually scaled in dimensions prior to its volume, undertaking an intuitive portrait of the most commonplace content terms and conditions round the new try ( Wordle, 2014).

Figure step one suggests the 20 most frequent blogs terms included in the complete sample. As can get noticed, probably the most commonly used terms and conditions were like (lookin when you look at the 67% out-of pages), including (appearing for the 62% from pages), searching (looking during the 55% regarding profiles), and you will someone (searching in fifty% of pages). Hence, typically the most popular terms and conditions have been similar all over a long time.

Figure dos reveals the next 30 most commonly known posts conditions when you look at the the newest youngest and you will earliest a long time. By eliminating the initial 20 well-known stuff conditions along side test, we teach heterogeneity on matchmaking users. Next 31 terminology into youngest age bracket, significant number terminology integrated rating (36% of users regarding the youngest generation), wade (33% regarding profiles on youngest age bracket), and you can work (28% off pages on the youngest generation). Having said that, the oldest age group had high rates off terms and conditions like traveling (31% out of users about eldest generation), higher (24% regarding profiles in the eldest age bracket), and you can dating (19% from users on the eldest age group).

2nd 31 most commonly known terminology on the youngest and you can oldest many years groups (just after deducting brand new 20 popular conditions away from Contour step one).

Theory Comparison of age Differences in Code when you look at the Relationships Pages

To evaluate hypotheses, the latest percentage of terms and conditions regarding the dating reputation that suit for every LIWC class offered since centered parameters from inside the regressions. We tested many years and you will gender just like the independent parameters including modifying for webpage web site and you may ethnicity.

Hypothesis 1: Older age would be associated with the increased portion of conditions regarding the after the classes: first-person plural pronouns, members of the family, members of the family, wellness, and you can self-confident emotion.

Findings mostly supported Theory 1 (see Desk 2). Five of four regressions shown a critical fundamental effect getting age, in a way that since period of the newest character author increased, the newest portion of terminology about classification improved regarding the following categories: first-people plural, family members, health, and you will confident emotion. I discover no tall decades feeling into ratio off terminology throughout the loved ones class.

good Gender: 0 (female) and you may step one (male). b Site: The 2 other sites was basically dictomously coded since the step one and 0. c Ethnicity: 0 (White) and you may step 1 (Ethnic or racial minority).

a Gender: 0 (female) and you may 1 (male). b Site: Both websites was basically dictomously coded as the step 1 and you may 0. c Ethnicity: 0 (White) and 1 (Cultural or racial fraction).