Descriptive Analytics and you can Graphics away from Commonly used Terms

We looked at possible distinctions of the webpages, geographic region, and you can ethnicity using t-assessment and data out-of variance (ANOVA) to the LIWC class rates. On a few websites, six of 12 t-tests was in fact extreme in the following the kinds: first-individual just one [t(3998) = ?5.61, p Second Dining table 2 to have mode, standard deviations, and you will contrasts ranging from cultural communities). Contrasts revealed extreme differences when considering Light and all sorts of most other ethnic groups into the five of the half dozen significant ANOVAs. Hence, we included ethnicity since a good dummy-coded covariate when you look at the analyses (0 = Light, step 1 = Some other ethnic groups).

Of one’s a dozen ANOVA screening about geographical area, merely several was basically extreme (members of the family and positive feelings). Since distinctions just weren’t technically meaningful, i don’t think geographical part in the further analyses.


Frequency out-of keyword use is obvious in detailed analytics (discover Desk step one) and you can via keyword-clouds. The definition of-cloud method portrays probably the most widely used conditions over the whole take to and in each of the age groups. The word-cloud system immediately excludes certain terms, plus articles (a good, and, the) and prepositions (so you can, that have, on). The rest blogs words is actually scaled in dimensions in line with its frequency, carrying out an intuitive portrait of the very common stuff terms round the this new decide to try ( Wordle, 2014).

Contour 1 suggests the newest 20 common posts terms and conditions found in the whole try. As can rise above the crowd, more frequently used words was indeed like (looking from inside the 67% from pages), particularly (looking when you look at the 62% away from pages), looking (looking inside the 55% out-of pages), and you can anybody (searching into the fifty% regarding profiles). Hence, the best words have been equivalent across a long time.

Profile dos reveals the second 31 most common posts terminology inside the new youngest and you will eldest age range. By eliminating the first 20 prominent blogs conditions over the attempt, i teach heterogeneity in the relationship profiles. In the next 31 terms and conditions to your youngest generation, significant number terminology incorporated score (36% out-of pages regarding the youngest age bracket), go (33% regarding pages regarding the youngest age bracket), and you can work (28% of users on the youngest age group). In contrast, this new oldest age bracket got higher percentages regarding words instance take a trip (31% from profiles from the earliest generation), great (24% away from pages about oldest age group), and you may matchmaking (19% out-of users about oldest generation).

Second 31 common terminology regarding youngest and oldest years organizations (immediately following subtracting the 20 most commonly known terms off Contour step one).

Theory Analysis old Differences in Vocabulary within the Relationships Profiles

To check on hypotheses, the new portion of words from the matchmaking character that fit each LIWC category supported while the depending details for the regressions. I tested decades and you will sex as the independent variables also adjusting to possess site and ethnicity.

Hypothesis step one: More mature ages might be from the a high percentage of words throughout the following the kinds: first-individual plural pronouns, household members, family relations, fitness, and positive emotion.

Conclusions mostly offered Theory step 1 (get a hold of Desk dos). Five of your own five regressions shown a life threatening fundamental effect having decades, in a manner that as the chronilogical age of the character publisher enhanced, the portion of conditions regarding the class improved regarding the after the categories: first-individual plural, loved ones, health, and you will self-confident feelings. We found no tall years impression on the proportion from terms and conditions in the family relations classification.

a gender: 0 (female) and you may step 1 (male). b Site: Both websites have been dictomously coded once the 1 and you can 0. c Ethnicity: 0 (White) and you can 1 (Cultural or racial fraction).

an intercourse: 0 (female) and you may 1 (male). b Webpages: The 2 websites was in fact dictomously coded as the step 1 and you can 0. c Ethnicity: 0 (White) and you will 1 (Cultural or racial minority).