I looked at possible differences because of the webpages, geographic part, and you can ethnicity playing with t-examination and you may studies away from variance (ANOVA) on the LIWC classification percentages. On the one or two websites, half a dozen of twelve t-evaluation were extreme in the adopting the kinds: first-people only 1 [t(3998) = ?5.61, p Second Desk 2 to own form, important deviations, and contrasts anywhere between ethnic groups). dil mil username Contrasts found high differences between Light and all of most other ethnic organizations during the five of your half dozen high ANOVAs. Ergo, we provided ethnicity because the a good dummy-coded covariate from inside the analyses (0 = White, 1 = All other ethnic organizations).
Of 12 ANOVA assessment pertaining to geographical area, just several have been significant (family and you may positive emotion). Since distinctions just weren’t officially significant, we failed to envision geographic area in further analyses.
Results
Volume of word have fun with goes without saying in the detailed statistics (come across Desk 1) and through keyword-clouds. The word-cloud strategy portrays probably the most commonly used terms over the entire test plus all the age range. The expression-cloud system immediately excludes specific terminology, along with posts (a, and you will, the) and you can prepositions (to help you, with, on). The rest articles terms try scaled sizes relative to the volume, carrying out an intuitive portrait really commonplace content terms and conditions round the the newest sample ( Wordle, 2014).
Figure 1 shows new 20 popular blogs words found in the whole attempt. As can rise above the crowd, the essential frequently employed conditions was like (searching from inside the 67% of users), like (lookin inside the 62% of users), appearing (searching into the 55% regarding pages), and some body (lookin into the fifty% regarding pages). Ergo, the most famous terms and conditions was basically similar across the age groups.
Shape 2 shows the following 31 most commonly known content conditions from inside the brand new youngest and you will oldest age groups. By removing the original 20 popular posts terms and conditions along side test, we illustrate heterogeneity throughout the relationship profiles. Within the next 29 conditions into the youngest age group, high level percentage terms and conditions provided get (36% away from pages from the youngest generation), wade (33% regarding users in the youngest age bracket), and you can functions (28% out of users regarding youngest age group). Alternatively, new eldest age group got highest percentages regarding terms such as travel (31% from pages in the oldest age bracket), high (24% of users about earliest generation), and you can matchmaking (19% out of pages from the eldest generation).
Next 30 most common terms throughout the youngest and eldest years organizations (shortly after subtracting the brand new 20 most common terms out-of Contour step one).
Hypothesis Analysis old Variations in Language for the Matchmaking Users
To check on hypotheses, this new part of terms and conditions regarding matchmaking character that fit for each LIWC class supported while the founded parameters within the regressions. I checked-out age and you may gender since separate parameters including adjusting getting web site and you may ethnicity.
Hypothesis 1: More mature ages could well be for the a higher part of words in the following the kinds: first-people plural pronouns, loved ones, nearest and dearest, health, and confident feelings.
Results mostly supported Hypothesis 1 (come across Dining table dos). Four of your four regressions revealed a serious chief feeling to possess many years, in a manner that due to the fact age of the new profile journalist enhanced, the brand new percentage of terms and conditions throughout the category enhanced in the adopting the categories: first-person plural, members of the family, fitness, and you can confident feeling. We discover zero tall decades effect into the proportion out of terminology regarding family relations class.
an effective Gender: 0 (female) and you can step 1 (male). b Website: The 2 websites had been dictomously coded since the step 1 and you will 0. c Ethnicity: 0 (White) and you can step 1 (Cultural otherwise racial minority).
an excellent Gender: 0 (female) and step 1 (male). b Site: The two websites have been dictomously coded while the step 1 and you will 0. c Ethnicity: 0 (White) and you can step 1 (Ethnic or racial fraction).