We checked prospective differences by the webpages, geographic part, and you may ethnicity having fun with t-examination and you may study regarding difference (ANOVA) to your LIWC classification percentages. On a few other sites, half a dozen of several t-screening were tall in the after the groups: first-person one [t(3998) = ?5.61, p Second Dining table dos having mode, simple deviations, and contrasts anywhere between cultural communities). Contrasts found significant differences between White as well as almost every other cultural organizations inside the four of your half a dozen significant ANOVAs. Thus, i integrated ethnicity due to the fact an effective dummy-coded covariate inside the analyses (0 = Light, 1 = Any cultural groups).
Of your twelve ANOVA tests pertaining to geographic region, just several was basically tall (family and you may self-confident emotion). Because distinctions weren’t commercially meaningful, i didn’t imagine geographic area during the next analyses.
Abilities
Frequency regarding keyword use is evident from inside the detailed statistics (find Table step 1) and you can via term-clouds. The word-cloud approach illustrates the quintessential commonly used words across the whole shot and in each of the age range. The expression-affect system immediately excludes specific terms, in addition to posts (a good, and you may, the) and you may prepositions (to help you, with, on). The rest articles words are scaled in size according to its volume, starting an intuitive portrait really commonplace blogs terms round the this new sample ( Wordle, 2014).
Profile 1 shows the latest 20 most common articles terms and conditions utilized in the complete attempt. As well as be seen, the most frequently employed terms was indeed love (looking when you look at the 67% out of users), particularly (searching when you look at the 62% of pages), appearing (appearing for the 55% off pages), and people (appearing during the 50% hot sexy Enterprise, OK girls out-of profiles). For this reason, widely known conditions was comparable round the age groups.
Profile 2 shows next 29 most frequent stuff terminology in brand new youngest and you may earliest age groups. By removing the original 20 prominent articles terms and conditions along the attempt, we illustrate heterogeneity regarding relationship profiles. Next 30 conditions towards the youngest age group, high percentage conditions integrated rating (36% out of profiles in the youngest generation), wade (33% from profiles in the youngest generation), and you may performs (28% from profiles about youngest age group). However, the newest earliest generation had large rates away from terms including traveling (31% away from users on oldest generation), high (24% off profiles on the oldest age group), and you will relationships (19% out-of profiles regarding oldest generation).
2nd 30 popular terms about youngest and you may oldest ages organizations (once subtracting the latest 20 popular terms off Figure 1).
Theory Testing old Variations in Words for the Dating Pages
To check on hypotheses, this new percentage of terms and conditions on the dating reputation that fit for each and every LIWC classification served given that founded variables when you look at the regressions. We checked age and you can gender since the independent variables in addition to changing to have webpages and you will ethnicity.
Hypothesis step one: Old years is regarding the a higher part of terms in the adopting the groups: first-individual plural pronouns, family unit members, family, wellness, and you may confident feelings.
Conclusions largely offered Theory step one (get a hold of Desk 2). Five of five regressions shown a critical chief impact to have years, in a way that since age the brand new profile copywriter enhanced, new part of words regarding classification improved on the pursuing the categories: first-individual plural, family, health, and you will positive feelings. We discover no extreme age perception towards the ratio out of conditions about family unit members category.
a great Gender: 0 (female) and you can step one (male). b Webpages: The two other sites was in fact dictomously coded once the step 1 and you will 0. c Ethnicity: 0 (White) and step 1 (Ethnic otherwise racial minority).
an effective Gender: 0 (female) and you may 1 (male). b Site: Both other sites have been dictomously coded because the 1 and you will 0. c Ethnicity: 0 (White) and you can step 1 (Cultural otherwise racial minority).