We examined potential variations by the webpages, geographic part, and ethnicity having fun with t-testing and you may research regarding variance (ANOVA) to your LIWC classification percentages. With the a few websites, six of your own twelve t-assessment was high from the after the groups: first-individual only one [t(3998) = ?5.61, p Supplementary Table 2 to have mode, basic deviations, and you may contrasts between ethnic communities). Contrasts revealed extreme differences when considering Light as well as other ethnic groups within the four of your half dozen significant ANOVAs. Hence, we provided ethnicity given that an effective dummy-coded covariate into the analyses (0 = Light, step 1 = Another cultural communities).
Of one’s a dozen ANOVA testing related to geographic part, simply a couple were tall (nearest and dearest and self-confident feelings). Since distinctions weren’t officially meaningful, i didn’t consider geographical part inside then analyses.
Overall performance
Volume regarding word explore is evident when you look at the detailed analytics (look for Table step 1) and through keyword-clouds. The term-cloud approach depicts the absolute most commonly used terms and conditions across the whole take to plus in each one of the a long time. The expression-affect system instantly excludes certain terminology, and stuff (a great, and you can, the) and prepositions (so you’re able to, having, on). The rest content terminology try scaled in size prior to their frequency, doing an intuitive portrait of the most commonplace content conditions round the the fresh new shot ( Wordle, 2014).
Profile 1 shows the brand new 20 most common content words found in the entire attempt. As can get noticed, the essential commonly used terminology have been like (searching inside 67% off users), such as for example (searching into the 62% off pages), lookin (searching during the 55% away from profiles), and you may people (appearing into the 50% from pages). For this reason, widely known words was in fact equivalent around the age range.
Contour dos shows the following 31 common posts words into the the brand new youngest and you will earliest a long time. By detatching the original 20 preferred posts terms along side take to, we instruct heterogeneity in the relationships pages. In the next 30 terminology to your youngest age bracket, high level percentage terms incorporated rating (36% regarding users in the youngest age bracket), go (33% regarding profiles from the youngest age group), and you will work (28% of profiles regarding the youngest age group). However, the newest earliest generation had high percentages out of terms and conditions including take a trip (31% of users about earliest age bracket), high (24% out of profiles on eldest age group), and relationship (19% from pages regarding oldest generation).
2nd 29 most commonly known conditions from the youngest and you can oldest years organizations (after deducting this new 20 most frequent words regarding Shape step 1).
Theory Evaluation of age Differences in Code inside Dating Pages
To check on hypotheses, this new part of terms regarding relationship character that fit for each and every LIWC classification served due to the fact created parameters from inside the regressions. I checked age and you can intercourse as the separate variables and adjusting getting site and you will ethnicity.
Theory step one: Elderly years would-be on the increased part of terminology on the pursuing the classes: first-people plural pronouns, friends, family members, fitness, and you may positive emotion.
Findings mostly supported Theory 1 (come across Table 2). Four of one’s five regressions found a critical main feeling to have ages, in a way that as the age the newest character creator enhanced, the brand new portion of conditions in the class increased on following categories: first-individual plural, family unit members, fitness, and confident feelings. I discovered zero extreme ages impression towards ratio out of words regarding the relatives class.
a gender: 0 (female) and you may 1 (male). b Site: The 2 other sites had been dictomously coded given that 1 and 0. c Ethnicity: 0 (White) and step 1 (Ethnic otherwise racial fraction).
a sex: 0 (female) and step one (male). b Website: The 2 other sites was basically dictomously coded since the 1 and you will 0. c Ethnicity: 0 (White) and you may step one (Ethnic or racial minority).