We looked at possible differences of the site, geographic area, and you can ethnicity using t-evaluation and you may investigation away from difference (ANOVA) towards LIWC classification percent. For the a couple websites, half a dozen of 12 t-tests had been high in the pursuing the groups: first-person only 1 [t(3998) = ?5.61, p Second Table dos to possess means, fundamental deviations, and you will contrasts between ethnic communities). Contrasts found extreme differences when considering Light as well as most other ethnic teams for the five of your half a dozen extreme ANOVAs. Thus, we included ethnicity as a beneficial dummy-coded covariate within the analyses (0 = Light, step 1 = Various other cultural groups).
Of your a dozen ANOVA examination related to geographical part, only a few had been extreme (loved ones and self-confident feeling). Because the distinctions just weren’t technically important, i failed to imagine geographic part for the after that analyses.
Results
Regularity of word explore is evident inside the descriptive statistics (select Table step 1) and via word-clouds. The definition of-affect techniques depicts the most popular terms along side whole take to plus in each of the age groups. The definition of-affect program instantly excludes particular terms and conditions, and additionally articles (a beneficial, and, the) and you can prepositions (to help you, with, on). The remainder blogs terms was scaled sizes prior to the frequency, doing an intuitive portrait of the most extremely commonplace content words round the the test ( Wordle, 2014).
Shape 1 reveals new 20 most commonly known articles words utilized in the whole take to. As can rise above the crowd, one particular frequently used terms and conditions was in fact love (looking inside the 67% off users), such as (appearing during the 62% of users), looking (looking inside 55% out-of users), and you will people (appearing into the fifty% regarding pages). Therefore, the best terms and conditions have been similar round the age groups.
Shape dos shows next 29 most common articles terms inside the the youngest and you can oldest age groups. By removing the initial 20 popular articles terms over the attempt, we show heterogeneity on the relationship users. In the next 29 terminology toward youngest generation, high level percentage terminology included get (36% away from pages on youngest age group), go (33% out-of users from the youngest age bracket), and you may functions (28% out-of users regarding youngest generation). Alternatively, the latest earliest generation had large proportions of terms such as take a trip (31% away from users regarding the earliest generation), higher (24% out of users throughout the earliest age group), and relationship (19% off pages regarding oldest age bracket).
Next 29 typical terms throughout the youngest and you will oldest decades communities (after deducting the latest 20 common terminology away from Shape step one).
Theory Comparison old Variations in Code from inside the Matchmaking Pages
To check on hypotheses, the percentage of terms in the relationship reputation that suit for every LIWC category offered once the established details in the regressions. I examined ages and you can sex while the independent details together with adjusting to have webpages and you may ethnicity.
Hypothesis step one: Earlier ages could be of this a higher part of conditions regarding the pursuing the classes: first-person plural pronouns, family relations, members of the family, wellness, and confident emotion.
Findings mainly offered Hypothesis step 1 (get a hold of Dining table 2). Five of your five regressions shown a serious main feeling for ages, in a way that given that chronilogical age of the brand new reputation journalist enhanced, brand new portion of conditions throughout the class enhanced regarding the pursuing the categories: first-people plural, family relations, wellness, and self-confident feelings. We located zero tall years effect to your ratio of words throughout the family unit members classification.
a sex: 0 (female) and 1 (male). b Webpages: The 2 other sites have been dictomously coded since step 1 and you will 0. c Ethnicity: 0 (White) and you can step 1 (Cultural otherwise racial minority).
a sex: 0 (female) and you can 1 (male). b Web site: The two websites was dictomously coded given that step one and you may 0. c Ethnicity: 0 (White) and you will step 1 (Cultural otherwise racial fraction).