I examined potential distinctions from the website, geographic region, and you may ethnicity playing with t-examination and you can study of difference (ANOVA) with the LIWC class percent. Into two other sites, six of several t-examination have been tall on the following the classes: first-people only one [t(3998) = ?5.61, p Secondary Dining table dos to own means, basic deviations, and contrasts anywhere between ethnic organizations). Contrasts shown tall differences between Light as well as almost every other cultural organizations inside four of your own six extreme ANOVAs. Thus, i incorporated ethnicity given that a beneficial dummy-coded covariate when you look at the analyses (0 = White, step 1 = Some other ethnic groups).

Of twelve ANOVA screening regarding geographical part, merely a couple of had been high (relatives and you will positive feeling). Due to the fact distinctions weren’t theoretically meaningful, we did not consider geographical part in the then analyses.

Efficiency

Frequency from keyword fool around with is obvious within the descriptive statistics (discover Desk step one) and you may via term-clouds. The term-cloud approach depicts probably the most widely used words across the whole attempt plus all the age groups. The expression-cloud program immediately excludes particular words, as well as articles (an excellent, and you will, the) and you may prepositions (to help you, that have, on). The remainder blogs words is actually scaled in proportions in accordance with its frequency, performing an intuitive portrait of the most common stuff conditions around the new take to ( Wordle, 2014).

Figure step 1 suggests the latest 20 most typical content conditions included in the entire attempt. As well as get noticed, the quintessential frequently used words was in fact love (appearing during the 67% out-of profiles), particularly (searching into the 62% from pages), lookin (appearing inside 55% of pages), and you can anybody (searching in 50% away from users). Hence, the most used terms and conditions have been similar all over age range.

Shape dos suggests the second 31 popular stuff conditions into the brand new youngest and you can earliest age range. By removing the initial 20 well-known stuff conditions over the test, we instruct heterogeneity regarding the relationship users. In the next 29 terminology for the youngest generation, raised percentage terms and conditions provided rating (36% from users on youngest generation), go (33% off profiles regarding the youngest generation), and you will functions (28% out-of profiles from the youngest age group). Conversely, the brand new oldest generation had higher percent regarding terms and conditions for example traveling (31% away from profiles regarding Klikk pГҐ denne lenken nГҐ the eldest age group), high (24% out of pages in the oldest generation), and you can matchmaking (19% out-of users regarding the eldest age bracket).

Next 31 most typical terms in the youngest and you can eldest decades groups (after deducting this new 20 common terminology away from Shape step 1).

Theory Review of age Variations in Vocabulary from inside the Relationship Profiles

To check on hypotheses, the new portion of terms on the dating reputation that suit for every single LIWC category offered while the centered parameters inside regressions. We checked-out years and you can gender as independent variables as well as changing to own website and you may ethnicity.

Hypothesis 1: Older many years would-be from the increased percentage of terms throughout the after the kinds: first-person plural pronouns, family members, family relations, fitness, and you may self-confident emotion.

Findings mainly offered Hypothesis step 1 (come across Dining table 2). Four of your own five regressions shown a critical main effect to own age, in a fashion that as age the latest character blogger improved, the brand new part of terms about classification increased about following categories: first-people plural, friends, wellness, and self-confident feelings. I found no extreme many years impression to the proportion regarding conditions regarding the relatives category.

a good Gender: 0 (female) and step 1 (male). b Website: The 2 other sites have been dictomously coded once the step one and you will 0. c Ethnicity: 0 (White) and step 1 (Ethnic otherwise racial fraction).

an excellent Gender: 0 (female) and 1 (male). b Webpages: The 2 other sites was basically dictomously coded as step one and you may 0. c Ethnicity: 0 (White) and you can step 1 (Ethnic otherwise racial fraction).