I looked at possible variations of the webpages, geographical area, and ethnicity having fun with t-examination and study out of difference (ANOVA) towards the LIWC group percent. On a few other sites, six of your a dozen t-screening have been extreme on the after the classes: first-people only 1 [t(3998) = ?5.61, p Supplementary Dining table 2 to have means, fundamental deviations, and you will contrasts anywhere between cultural teams). Contrasts found extreme differences when considering Light as well as almost every other ethnic teams inside the four of your six significant ANOVAs. Ergo, we incorporated ethnicity Reno, TX women dating while the an excellent dummy-coded covariate when you look at the analyses (0 = Light, 1 = Every other cultural teams).
Of the twelve ANOVA tests about geographic part, just one or two was indeed extreme (family unit members and you can positive feeling). Given that differences just weren’t officially meaningful, i failed to thought geographic area during the then analyses.
Show
Volume out of keyword use goes without saying inside the descriptive analytics (come across Table step 1) and you may through term-clouds. The word-affect strategy depicts one particular widely used terminology over the entire decide to try as well as in each of the age groups. The expression-affect system immediately excludes particular terminology, plus posts (good, and, the) and you can prepositions (to, having, on). The remaining articles conditions was scaled in proportions according to their volume, carrying out an intuitive portrait quite common stuff terminology across the the fresh attempt ( Wordle, 2014).
Shape step one suggests new 20 most common content terms found in the complete try. As can be seen, one particular commonly used conditions was indeed like (looking inside the 67% from profiles), for example (appearing for the 62% regarding users), searching (searching in the 55% off pages), and you will some one (searching into the fifty% out-of profiles). Thus, the preferred words were equivalent across age groups.
Profile 2 reveals the following 29 most common content terms within the the brand new youngest and oldest a long time. By eliminating the initial 20 common blogs conditions along the attempt, we instruct heterogeneity in the matchmaking profiles. Next 31 terminology towards youngest generation, high level percentage conditions integrated get (36% of pages regarding the youngest age bracket), wade (33% off users in the youngest age group), and you will really works (28% away from profiles in the youngest generation). Alternatively, the latest eldest age group got highest rates off terms and conditions for example take a trip (31% away from pages regarding earliest age group), higher (24% from pages from the eldest generation), and dating (19% of pages regarding oldest generation).
Next 29 popular conditions on youngest and you can oldest years communities (immediately after deducting brand new 20 most typical conditions regarding Contour step 1).
Hypothesis Testing of age Variations in Vocabulary within the Relationship Pages
To check hypotheses, this new percentage of terminology on the matchmaking character that suit for every single LIWC group offered once the mainly based parameters for the regressions. I examined decades and you can gender given that independent variables in addition to changing for site and you will ethnicity.
Theory 1: Older ages is associated with the a top part of words regarding the after the kinds: first-individual plural pronouns, household members, loved ones, wellness, and you will confident feelings.
Results largely served Hypothesis step one (select Table 2). Five of the five regressions found a significant fundamental effect to possess ages, in a fashion that because age of this new reputation journalist improved, new percentage of terminology on category improved in the adopting the categories: first-individual plural, friends, fitness, and you can self-confident feelings. I discovered no tall many years impression for the proportion out of terms and conditions about members of the family classification.
good Gender: 0 (female) and you can step 1 (male). b Site: The two websites was indeed dictomously coded while the step one and you can 0. c Ethnicity: 0 (White) and you will 1 (Ethnic or racial minority).
a great Gender: 0 (female) and step 1 (male). b Web site: Both websites were dictomously coded as step one and you can 0. c Ethnicity: 0 (White) and you can step one (Ethnic or racial fraction).
