I examined prospective variations of the site, geographic area, and you may ethnicity playing with t-evaluating and analysis of variance (ANOVA) into LIWC classification proportions. Into two websites, half dozen of your several t-testing have been extreme throughout the following groups: first-individual only one [t(3998) = ?5.61, p Secondary Dining table dos for form, fundamental deviations, and you can contrasts between cultural teams). Contrasts shown extreme differences between White and all other cultural groups inside the five of six high ANOVAs. Therefore, i included ethnicity since the an excellent dummy-coded covariate within the analyses (0 = Light, step one = Some other cultural teams).
Of several ANOVA screening associated with geographic part, simply two were significant (nearest and dearest and you can positive feelings). While the differences weren’t officially meaningful, i didn’t imagine geographical area from inside the then analyses.
Show
Volume from phrase play with is evident inside the detailed analytics (come across Table step one) and you can thru term-clouds. The phrase-cloud approach illustrates the most popular conditions along the entire test and in all the age groups. The definition of-cloud system instantly excludes specific conditions, in addition to stuff (a good, and, the) and you will prepositions (so you’re able to, having, on). The rest content terms and conditions is scaled in proportions in accordance with their regularity, undertaking an intuitive portrait quite common stuff terms round the this new sample ( Wordle, 2014).
Shape step 1 reveals new 20 most common stuff terms found in the entire shot. As can rise above the crowd, one particular commonly used conditions have been love (appearing within the 67% from pages), instance (lookin during the 62% from profiles), searching (searching in the 55% of users), and you can anyone (appearing into the fifty% out-of profiles). Ergo, the most famous terms and conditions have been similar round the a long time.
Figure 2 shows the following 31 typical articles words during the new youngest and you will oldest age groups. By detatching the original 20 preferred blogs terminology over the attempt, i train heterogeneity on the matchmaking profiles. Within the next 30 terminology toward youngest age group, high percentage terms incorporated rating (36% of profiles about youngest age bracket), wade (33% out-of profiles throughout the youngest generation), and work (28% out of users regarding youngest generation). Having said that, the newest eldest age bracket had large percentages regarding terminology such as travel (31% off users regarding oldest age group), high (24% regarding profiles on the eldest generation), and you may relationship (19% off profiles on eldest age bracket).
2nd 30 most typical conditions about youngest and oldest St. Paul, VA women sexy decades groups (after subtracting the fresh new 20 most frequent words out of Shape step 1).
Hypothesis Research old Differences in Language inside the Relationship Users
To evaluate hypotheses, the new percentage of conditions from the dating profile that fit for each and every LIWC category supported as the dependent variables for the regressions. We checked out years and you may gender because the separate variables together with modifying to possess web site and ethnicity.
Hypothesis step 1: Older decades would be with the increased portion of terms and conditions regarding the following the kinds: first-person plural pronouns, family members, family members, fitness, and you can confident feelings.
Findings mainly offered Theory step 1 (look for Desk dos). Four of the four regressions shown a serious fundamental feeling to possess many years, in a manner that given that age the newest character journalist enhanced, the latest portion of conditions on category increased regarding following categories: first-person plural, relatives, wellness, and you can self-confident feeling. We receive zero tall age impact to your ratio out-of terms and conditions throughout the family unit members group.
a Gender: 0 (female) and you will step one (male). b Website: The 2 other sites was indeed dictomously coded because the step 1 and you will 0. c Ethnicity: 0 (White) and step 1 (Ethnic otherwise racial minority).
a beneficial Gender: 0 (female) and you may step one (male). b Web site: Both websites was basically dictomously coded while the step 1 and you may 0. c Ethnicity: 0 (White) and you will 1 (Ethnic or racial minority).